Side-chain hydrophobicity scale derived from

transmembrane protein folding into

lipid bilayers
C. Preston Moon and Karen G. Fleming1
Thomas C. Jenkins Department of Biophysics, Johns Hopkins University, 3400 North Charles Street, Baltimore, MD 21218

Edited* by Christopher Miller, Brandeis University, Waltham, MA, and approved April 29, 2011 (received for review March 11, 2011)

The transfer free energies of the twenty natural amino acid Results and Discussion
side chains from water to phospholipid bilayers make a major con- We used the outer membrane phospholipase A (OmpLA) as a
tribution to the assembly and function of membrane proteins. transmembrane scaffold on which to introduce amino acid side
Measurements of those transfer free energies will facilitate the chains of our choice at various membrane depths. We selected
identification of membrane protein sequences and aid in the under- OmpLA because it: (a) spontaneously folds and inserts into lipid
standing of how proteins interact with membranes during key bio- membranes from a solubilized unfolded state (14), (b) has a
logical events. We report the first water-to-bilayer transfer free known three-dimensional structure (Fig. 1A) (15), and (c) has
energy scale (i.e., a “hydrophobicity scale”) for the twenty natural enzymatic activity that can be monitored to confirm native-like
amino acid side chains measured in the context of a native trans- folding (16, 17). To measure side-chain hydrophobicity in a mem-
membrane protein and a phospholipid bilayer. Our measurements brane, we selected the alanine at position 210 in OmpLA’s
reveal parity for apolar side-chain contributions between soluble sequence as a host site for substitution to each of the other 19
and membrane proteins and further demonstrate that an arginine amino acids as guests. Position 210 is a lipid-facing exterior re-
side-chain placed near the middle of a lipid bilayer is accommo- sidue whose α-carbon is located only 0.2 Å from the midplane
dated with much less energetic cost than predicted by molecular between OmpLA’s membrane/water interfacial regions (Fig. 1A).
dynamics simulations. Our membranes were large unilamellar vesicles (LUVs) of 1,2-
dilauroyl-sn-glycero-3-phosphocholine (DLPC).
To determine the energetics of membrane partitioning of
T he transfer of amino acid side chains from water into phos-
pholipid lipid bilayers is a fundamental energetic contribution
to the total thermodynamic stability of transmembrane proteins
the guest amino acids, we adapted soluble protein chemical dena-
turation methods to measure the stability of folded proteins using
tryptophan fluorescence spectroscopy (18, 19). This approach
(1, 2). A hydrophobicity scale that ranks the water-to-bilayer has only previously been attempted with two other transmem-
transfer energies of the twenty natural amino acid side chains brane proteins folded into lipid vesicles (20–23) and it is only valid
could allow the identification of genes that code for membrane if folded and unfolded protein populations come to reversible
proteins (3–6) and aid the understanding of how proteins interact equilibrium at all concentrations of denaturant (18). We verified
with membranes during key biological events, such as cell signal- reversible folding of OmpLA at pH 3.8 (Fig. S1A). We do not yet
ing, drug binding, and the gating of ion-channels. However, understand why acidic pH promotes reversible folding of OmpLA
despite the abundance of available hydrophobicity scales, there or whether this is a common feature for other membrane proteins.
have been no experimental measurements of water-to-bilayer We confirmed the observed folded state of OmpLA was in a
transfer free energies of amino acid side chains that derive from native-like transmembrane conformation by verifying that it had
the partitioning of a native transmembrane protein from water enzymatic activity, which requires an acyl-chain of a substrate in
to the interior of a phospholipid bilayer. Existing estimations of the bilayer to bind to a hydrophobic pocket on the transmem-
water-to-bilayer partitioning of side chains were obtained from brane surface of OmpLA (16, 17). Further, we verified the folded
systems comprising small compounds to represent proteins state had its tryptophans embedded in the apolar lipid environ-
and/or organic solvents to mimic bilayers. The only experimental ment and was protected from protease digestion by being inserted
hydrophobicity scale derived from proteins and bilayers (the across the membrane (Fig. S2). We determined the unfolded
“translocon scale”) described the translocon-to-bilayer transition state of OmpLA was not embedded in the membrane and was
(7, 8). Mediated membrane insertion by the translocon complex instead solvated by water by verifying that it was enzymatically
does not explain the stability of membrane proteins in lipid inactive, that its tryptophans were in a polar environment
bilayers relative to a water-solvated state. whether or not LUVs were present, and it was completely diges-
Here we achieve thermodynamic measurements of amino acid tible by a protease (Fig. S2).
side-chain transfer into lipid bilayers in the context of a natively At equilibrium, OmpLA also adopted a thermodynamic inter-
folded transmembrane protein spanning a phospholipid bilayer. mediate state at a moderate range of denaturant concentrations,
Our measurements reveal the hydrophobic effect contributes as observed from the fluorescence emission of its tryptophans
equivalently to the stability of both membrane and soluble (Fig. 1 B–E and Fig. S1 B and C). Therefore, we used a three-
proteins (9). Our experiments further demonstrate that a mem- state linear-extrapolation model to extract a total standard-state
brane protein can accommodate an arginine side chain placed
near the apolar middle of a lipid bilayer with much less cost in Author contributions: C.P.M. designed research; C.P.M. performed research; C.P.M. and
energy than has been previously predicted by molecular dynamics K.G.F. analyzed data; and C.P.M. and K.G.F. wrote the paper.
simulations (10–11). Moreover, we found the membrane parti- The authors declare no conflict of interest.
tioning of two arginine residues to be thermodynamically coop- *This Direct Submission article had a prearranged editor.
erative. Therefore, our measurements should inform the study of See Commentary on page 10027.
voltage-sensing ion channels, which concertedly move multiple 1
To whom correspondence should be addressed. E-mail: karen.fleming@jhu.edu.
arginines into the hydrophobic interior of membranes during This article contains supporting information online at www.pnas.org/lookup/suppl/
their gating (12, 13). doi:10.1073/pnas.1103979108/-/DCSupplemental.

10174–10177 ∣ PNAS ∣ June 21, 2011 ∣ vol. 108 ∣ no. 25 www.pnas.org/cgi/doi/10.1073/pnas.1103979108

 SEE COMMENTARY
Fig. 2. Whole-protein hydrophobicity scale determined from the OmpLA
system. The difference in the Gibbs free energy of unfolding (ΔΔG∘w;l ) of
each amino acid variant at position 210 is compared to the wild-type OmpLA.
Error bars represent standard errors of the mean from individual titration
experiments. The color scheme is the same as in Fig. 1 B–E.

surface areas (ASAs) of those side chains and their values in

our whole-protein scale (Fig. S3). Importantly, the ASAs were
independent of the particular structural context of OmpLA be-
cause they were derived from exposure of the guest side chains in
an extended model tripeptide. This independence supports the
conclusion that our whole-protein scale is applicable to trans-
Fig. 1. Outer membrane phospholipase A (OmpLA) as a host-guest system
for measuring the membrane insertion energies of lipid-exposed amino acid
membrane proteins of diverse architecture. Indeed, a traditional
residues. (A) Backbone of OmpLA rendered using PyMol (DeLano Scientific) hydropathy analysis (3) using our whole-protein scale with the
and Protein Data Bank (PDB) ID code 1qd5 with the α-carbon of the alanine at Membrane Protein Explorer (MPEx) (5) application correctly
sequence position 210 shown as a black sphere. Solid horizontal lines repre- predicted the location of all seven transmembrane α-helices of
sent the boundaries of OmpLA’s transmembrane region. The dashed horizon- bovine rhodopsin (Fig. S4).
tal line represents the midplane of that region. (B–E) Example denaturation The correlation of the whole-protein scale and ASAs revealed
data for A210X sequence variants, each with a fit line from a three-state lin- that each Å2 of surface area of a hydrophobic side chain removed
ear-extrapolation model. Data points represent tryptophan fluorescence
from water and buried into a membrane added 23 cal mol−1 to
emission measurements, normalized such that the emission from folded pro-
tein is 1 and the emission from unfolded protein is 0. Data for the wild-type
the stability of OmpLA. This value is a quantitative description
OmpLA is shown in black in each panel. (B) Data for X ¼ F, I, L, M, P, V, W, and of the hydrophobic effect and constitutes an atomic solvation
Y are shown in orange. (C) Data for X ¼ C and G are shown in gray. (D) Data parameter (ASP) for transferring nonpolar side chains from
for X ¼ N, Q, S, and T are shown in teal. (E) Data for X ¼ D, E, H, K, and R are water into a lipid bilayer. Our ASP matches the ASP previously

COMPUTATIONAL BIOLOGY
shown in blue. found for transferring nonpolar side chains from water into the

BIOPHYSICS AND
nonpolar solvent octanol (24) as well as that for burying nonpolar
Gibbs free energy of unfolding (ΔG∘w;l ) for each of the OmpLA se- side chains into the hydrophobic interior of a water-soluble pro-
quence variants (Fig. 1 B–E) in the presence of just water (w) and tein (9). Therefore, the hydrophobic effect stabilizes both mem-
lipid bilayers (l); i.e., with no denaturant present. Because Gibbs brane and soluble proteins to an equivalent degree. We also used
free energy is a state-function, the appearance of an intermediate our ASP to calculate an insertion energy for the alanine side
in our denaturant titrations does not affect our total ΔG∘w;l values. chain by multiplying it by alanine’s nonpolar ASA (69 Å2 ). By
The differences in the ΔG∘w;l for the side-chain variants at po- shifting the whole-protein scale by alanine’s insertion energy,
sition 210 and the ΔG∘w;l of the wild-type OmpLA (ΔΔG∘w;l ) com- we can express a scale for all twenty side chains (Table S1) that
prise a unique whole-protein hydrophobicity scale (Fig. 2 and does not use any particular one of them as a reference. It should
Table S1). Because the contributions of other residues of OmpLA be noted that this side-chain scale does not include a contribution
are subtracted, the ΔΔG∘w;l values in the whole-protein scale re- of the peptide bond.
flect the energies required to partition just the guest amino acid Using our thermodynamically rigorous host-guest system, we
side chains into the membrane relative to the alanine side chain at were well positioned to investigate the energetic cost of water-
position 210 in the wild-type sequence. Therefore, these values to-membrane partitioning of arginine. This cost has been a matter
should be applicable to membrane proteins of all architectures of controversy because it concerns the ability of the arginine-rich
and will have utility in calibrating force fields or protocols for S4 helix of KvAP and related ion channels to sense membrane
membrane protein calculations. depolarization and mediate channel gating (10, 11, 13, 25, 26).
Exclusion from water appeared to be the main contribution to Earlier molecular dynamics simulations (10, 11) calculated the
the membrane partition energies for the hydrophobic side chains total cost of moving a single arginine to the middle of a mem-
of Ala, Leu, Ile, Met, Phe, Tyr, and Val. We found a strong linear brane to be 14–17 kcal mol−1 . If the side chain snorkeled toward
correlation (R ¼ 0.96) between the nonpolar solvent accessible the water/bilayer interface, the calculated cost would still be

Moon and Fleming PNAS ∣ June 21, 2011 ∣ vol. 108 ∣ no. 25 ∣ 10175
6–7 kcal mol−1 (8). The energetics of water-to-membrane parti-
tioning of arginine would also depend on possible rearrangement
of other parts of the protein and/or a deformation of the mem-
brane to reduce its hydrophobic thickness (10, 11, 27, 28). How-
ever, no such energetics have been measured with a whole protein
until the present study. The previous translocon scale had a mod-
est value for arginine (7), but it did not represent the water-to-
bilayer transition (8). Our measurement for arginine here does
represent a water-to-bilayer transition and it also is quite modest:
2.1
0.1 kcal mol−1 (Table S1). We speculate that at least part of
the discrepancy between our experimental value and the previous
theoretical values may be due to differences in the hydrophobic
thicknesses of the lipid bilayers used in the different systems and
may be resolved by further calculations that are matched to our
experimental conditions.
To address the effects of bilayer depth on side-chain partition-
ing, we studied side chains at different positions on the OmpLA
scaffold. In addition to position 210, we engineered both arginine
and leucine at five additional locations in the OmpLA sequence
(positions 120, 164, 212, 214, and 223) and compared the result-
ing variants to alanine variants at the same positions. Arginine
Fig. 4. A double-mutant cycle reveals that two arginines cooperate energe-
partition energies were most unfavorable at the middle of the tically to partition into a membrane. The OmpLA system is shown similarly
membrane (Fig. 3 and Table S2) and displayed a shape with a as in Fig. 1 with the atoms of the side chains varied in the cycle shown as
depth-dependence that recapitulated the trends observed in ear- spheres. The arginine side-chain orientations shown are for illustration of size
lier molecular dynamics simulations (8, 10, 11, 27, 28) and in an and are not intended to depict any known side-chain conformations relative
earlier depth analysis of the translocon scale (6, 7). Leucine parti- to the lipid bilayer in our experiments.
tion energies had an opposite response—they were more favor-
able closer to the middle of the membrane. burden to pass across the apolar interior of the membrane during
We also examined a double-arginine variant (A210R, G212R) gating of the KvAP channel.
to address the issue of tandem membrane insertion of multiple The power of our whole-protein hydrophobicity scale is that it
arginine side chains that could occur with KvAP channel gating. unambiguously reflects the thermodynamics of water-to-bilayer
Fig. 4 and Table S2 show that the ΔΔG∘w;l of the double-arginine partitioning of side chains in the context of a native transmem-
variant is 1.6 kcal mol−1 less than the sum of the ΔΔG∘w;l for each brane protein spanning a phospholipid bilayer. The earlier trans-
locon scale (7) did not reflect a system verifiably at equilibrium,
single-arginine variant (A210R and G212R). Therefore, there is
and its values are relevant to a different event: translocon-to-
cooperativity between the two arginines that reduces their overall
bilayer partitioning (8). The Wimley White water-to-octanol scale
energetic cost to be in the membrane. The source of this coop-
(24) did derive from equilibrium thermodynamic measurements,
erativity could be the ability of the two arginines to share access to
however, it represented the protein by short peptides and it
a deformation or water penetration in the adjacent lipid bilayer. approximated the lipid membrane environment with octanol.
Therefore, our results suggest that the multiple arginines of Our observations reveal that both of those previous scales would
the S4 helix could function together to suppress their energetic undervalue the energetics of most amino acids if they were used
to represent water-to-bilayer side-chain partitioning (Fig. S5).
Two exceptions are aspartic acid and glutamic acid, whose en-
ergetics are underrepresented in our whole-protein scale because
the pH of our experiments was 3.8, which is close to the pKa
values for model compounds of Asp and Glu side chains (29).
It is reasonable to expect that a significant population of the
guest Asp and Glu side chains in our system were protonated and
thereby more easily partitioned into the membrane than they
would have at normal physiological pH. Therefore, our results
for Asp and Glu may help explain why certain membrane active
peptides, such as bacterial toxins (30), can partition across mem-
branes only at low pH. This phenomenon of acidic membrane
partitioning should be considered in the study of how protein–
lipid interactions affect the many types of tumors that feature
acidic extracellular pH (31).
Our measurements here provide water-to-bilayer transfer
free energies of amino acid side chains determined from the
spontaneous insertion of a whole transmembrane protein into
membranes. Therefore, our whole-protein free energy measure-
ments are highly applicable to models of spontaneous changes
Fig. 3. Energetics of side-chain partitioning varies by depth in the mem- in protein–lipid interactions, such as models of channel gating.
brane. The OmpLA host-guest system is shown similarly as in Fig. 1 with
To completely describe such events, however, more than just
the α-carbons of sequence positions 120, 164, 210, 212, 214, and 223 shown
as black spheres. The membrane depth of those five α-carbons versus the
equilibrium energies should be known. Channel gating is a dy-
ΔΔG∘w;l of leucine and arginine variants (compared to alanine variants) is namic and kinetic event, so the activation energy barriers to amino
shown aligned with the OmpLA image. Normal distributions fit to the leucine acid side-chain insertion into membranes should be determined.
and arginine data are also shown. Error bars represent standard errors of the Our system of using a scaffold transmembrane protein for guest
mean from individual titration experiments. amino acids should also allow us to measure those kinetic barriers.

10176 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1103979108 Moon and Fleming

 SEE COMMENTARY
Methods ments, we collected 350 readings of emission intensity at 330 nm with
Protein Folding denaturant Titrations. The wild-type OmpLA and its sequence 0.3 s of signal averaging for each reading. For samples used to show full tryp-
variants were engineered, expressed, and purified as previously described tophan spectral properties (Figs. S1A and S2B), we averaged four emission
(17). Prior to each experiment, we made fresh unfolded protein stocks dis- scans from 280 to 400 nm at 1 nm resolution. We found the wavelength posi-
solved in 8 M guanidine HCl (UltraPure powder, Invitrogen). The background tion of maximum intensity (λmax ) by fitting spectra to a log-normal function
buffer for all fluorescence experiments was 100 mM citrate (Sigma) and 2 mM (32, 33).
EDTA (Sigma), pH 3.8. For the variant A210C, we included 1 mM dithiothreitol
(DTT, from Sigma) in the buffers. We handled lipids (Avanti Polar Lipids) and Free Energy Measurements. We globally fit all sets of titration data for all pro-
prepared LUVs as previously described (14), but with 21 passes through two teins to a three-state linear-extrapolation model (34) using Igor Pro v6.12
stacked polycarbonate filters having a 0.1 μM pore size. (WaveMetrics) to find a common measure of the two m-values in the model
We set up protein folding and unfolding reactions in three steps. The first (see SI Text). To find total free energies of unfolding for the proteins, we
step was a dilution of the unfolded protein stocks to a final guanidine con- summed the two ΔG∘w;l values for each set of titration data included in
centration of 2.5 M and a final protein concentration of 6.0 μM with 1.4 mM the fit. Our measured free energy of unfolding for the wild-type OmpLA
3-N,N-Dimethylmyristyl-ammonio [propanesulfonate (SB3-14, from Sigma)]. (32.4
0.1 kcal mol−1 ) is about two- to threefold higher than previously re-
The SB3-14 prevented protein aggregation before the LUVs were added. ported values for OmpA (35) and PagP (23), which are smaller proteins. Thus,
The second step was a threefold dilution of the protein/detergent mixture the energy measurements for all three proteins may be consistent consider-
into LUVs of DLPC to attain a 2;000∶1 lipid-protein ratio and to take the ing the different sizes of the proteins and also the variation in the lipid sys-
SB3-14 below its critical micelle concentration. For titrations for free energy
tems used to make the measurements.
measurements (Fig. 1), this second step kept the protein unfolded at 5.0 M
guanidine HCl. For reversible folding verification (Fig. S1A), we prepared two
Membrane Depth Measurements. We determined where the membrane mid-
reactions at the second step: one allowing the protein to fold at 2.5 M gua-
plane should intersect OmpLA by using its atomic coordinates in PDB ID code
nidine HCl and the other keeping the protein unfolded at 5.0 M guanidine
1qd5 (Fig. 1A). First, we defined a boundary plane for each side of OmpLA’s
HCl. We incubated all samples from the second step at 37 °C with gentle
hydrophobic surface (extracellular and intracellular) by averaging the set of
rotation for 5 h. The third step was a fivefold dilution into buffer/guanidine
HCl mixtures to attain different final concentrations of guanidine HCl. The planes that are formed by each triad of polar atoms of OmpLA’s two aromatic
final protein concentration after the third step was 400 nM and the final lipid “girdles” of residues (i.e., the OH atoms of tyrosines and the NE1 atoms of
concentration was 800 μM. We incubated the titration samples from the third tryptophans) (36). The distance between the two boundary planes (24.4 Å)
step at 37 °C with gentle rotation for 40–50 h before fluorescence experi- was similar to a previously calculated value for OmpLA’s hydrophobic thick-
ments. For free energy measurements, we carried out three titrations of ness (23.2
1.5 Å) (37). We placed the membrane midplane halfway be-
the wild-type OmpLA and two titrations for each of the sequence variants. tween the boundary planes. We assigned the membrane depths of OmpLA’s
Each titration was an independent experiment with its own protein, lipid, residues as the distance between their α-carbons to the midplane along a
denaturant, and buffer stocks. normal to the midplane. The depths for the variant positions in this work
were: −4.0 Å for position 120, −10.1 Å for position 164, 0.2 Å for position
Tryptophan Fluorescence Emission Spectroscopy. Tryptophan fluorescence 210, 5.3 Å for position 212, 8.8 Å for position 214, and 7.1 Å for position 223.
emission was monitored on an ISS PC1 photon counting spectrofluorometer.
The excitation wavelength was 295 nm and the path length was 1 cm. We ACKNOWLEDGMENTS. This work was supported by grants from the National
used an excitation polarizer at 90° with 2.4 mm slits and an emission polarizer Science Foundation (MCB0423807, MCB0919868) and the National Institutes
at 0° with 2.0 mm slits (32, 33). For samples used for free energy measure- of Health (R01 GM079440, T32 GM008403).

1. Tanford C (1978) The hydrophobic effect and the organization of living matter. Science 19. Street TO, Courtemanche N, Barrick D (2008) Protein folding and stability using dena-
200:1012–1018. turants. Methods Cell Biol 84:295–325.
2. White SH, Wimley WC (1999) Membrane protein folding and stability: Physical 20. Hong H, Tamm LK (2004) Elastic coupling of integral membrane protein stability to
principles. Annu Rev Biophys Biomol Struct 28:319–365. lipid bilayer forces. Proc Natl Acad Sci USA 101:4065–4070.
3. Kyte J, Doolittle RF (1982) A simple method for displaying the hydropathic character of 21. Pocanschi CL, Patel GJ, Marsh D, Kleinschmidt JH (2006) Curvature elasticity and refold-
a protein. J Mol Biol 157:105–132. ing of OmpA in large unilamellar vesicles. Biophys J 91:L75–L77.
4. Bernsel A, et al. (2008) Prediction of membrane-protein topology from first principles. 22. Sanchez KM, Gable JE, Schlamadinger DE, Kim JE (2008) Effects of tryptophan micro-
Proc Natl Acad Sci USA 105:7177–7181. environment, soluble domain, and vesicle size on the thermodynamics of membrane
5. Snider C, Jayasinghe S, Hristova K, White SH (2009) MPEx: A tool for exploring protein folding: lessons from the transmembrane protein OmpA. Biochemistry
membrane proteins. Protein Sci 18:2624–2628. 47:12844–12852.
6. Hessa T, et al. (2007) Molecular code for transmembrane-helix recognition by the 23. Huysmans GHM, Baldwin SA, Brockwell DJ, Radford SE (2010) The transition state for
Sec61 translocon. Nature 450:1026–1030. folding of an outer membrane protein. Proc Natl Acad Sci USA 107:4099–4104.
7. Hessa T, et al. (2005) Recognition of transmembrane helices by the endoplasmic 24. Wimley WC, Creamer TP, White SH (1996) Solvation energies of amino acid side chains

COMPUTATIONAL BIOLOGY
reticulum translocon. Nature 433:377–381. and backbone in a family of host-guest pentapeptides. Biochemistry 35:5109–5124.
8. Schow EV, et al. (2011) Arginine in membranes: The connection between molecular 25. Hessa T, White SH, von Heijne G (2005) Membrane insertion of a potassium-channel

BIOPHYSICS AND
dynamics simulations and translocon-mediated insertion experiments. J Membr Biol voltage sensor. Science 307:1427.
283:35–48. 26. Roux B (2007) Lonely arginine seeks friendly environment. J Gen Physiol 130:233–236.
9. Chothia C (1974) Hydrophobic bonding and accessible surface area in proteins. Nature 27. Freites JA, Tobias DJ, von Heijne G, White SH (2005) Interface connections of a trans-
248:338–339. membrane voltage sensor. Proc Natl Acad Sci USA 102:15059–15064.
10. Dorairaj S, Allen TW (2007) On the thermodynamic stability of a charged arginine side 28. Bond PJ, Sansom MSP (2007) Bilayer deformation by the Kv channel voltage sensor
chain in a transmembrane helix. Proc Natl Acad Sci USA 104:4943–4948. domain revealed by self-assembly simulations. Proc Natl Acad Sci USA 104:2631–2636.
11. MacCallum JL, Bennett WF, Tieleman DP (2007) Partitioning of amino acid side chains 29. Nozaki Y, Tanford C (1967) Examination of titration behavior. Methods Enzymol
into lipid bilayers: results from computer simulations and comparison to experiment. 11:715–734.
J Gen Physiol 129:371–377. 30. O’Keefe DO, Cabiaux V, Choe S, Eisenberg D, Collier J (1992) pH-dependent insertion
12. Jiang Y, et al. (2003) X-ray structure of a voltage-dependent Kþ channel. Nature of proteins into membranes: B-chain mutation of diphtheria toxin that inhibits
423:33–41. membrane translocation, Glu-349 → Lys. Proc Natl Acad Sci USA 89:6202–6206.
13. Ruta V, Chen J, MacKinnon R (2005) Calibrated measurement of gating-charge argi- 31. Andreev OA, et al. (2007) Mechanism and uses of a membrane peptide that targets
nine displacement in the KvAP voltage-dependent Kþ channel. Cell 123:463–475. tumors and other acidic tissues in vivo. Proc Natl Acad Sci USA 104:7893–7898.
14. Burgess NK, Dao TP, Stanley AM, Fleming KG (2008) β-Barrel proteins that reside in the 32. Moon CP, Fleming KG (2011) Using tryptophan fluorescence to measure the stability of
Escherichia coli outer membrane in vivo demonstrate varied folding behavior in vitro. membrane proteins folded in liposomes. Methods Enzymol 492:189–211.
J Biol Chem 283:26748–26758. 33. Ladokhin AS, Jayasinghe S, White SH (2000) How to measure and analyze tryptophan
15. Snijder HJ, et al. (1999) Structural evidence for dimerization-regulated activation of fluorescence in membranes properly, and why bother? Anal Biochem 285:235–245.
an integral membrane phospholipase. Nature 401:717–721. 34. Latypov RF, Cheng H, Roder NA, Zhang J, Roder H (2006) Structural characterization of
16. Dekker N, Tommassen J, Lustig A, Rosenbusch JP, Verheij HM (1997) Dimerization an equilibrium unfolding intermediate in Cytochrome c. J Mol Biol 357:1009–1025.
regulates the enzymatic activity of Escherichia coli outer membrane phospholipase 35. Hong H, Park S, Flores-Jiménez RH, Rinehart D, Tamm LK (2007) Role of aromatic side
A. J Biol Chem 272:3179–3184. chains in the folding and thermodynamic stability of integral membrane proteins.
17. Stanley AM, Fleming KG (2007) The role of a hydrogen bonding network in the J Am Chem Soc 129:8320–8327.
transmembrane β-barrel OMPLA. J Mol Biol 370:912–924. 36. Lee AG (2003) Lipid-protein interactions in biological membranes: A structural
18. Santoro MM, Bolen DW (1992) Test of the linear extrapolation of unfolding free perspective. Biochim Biophys Acta 1612:1–40.
energy changes over an extended denaturant concentration range. Biochemistry 37. Lomzie AL, Pogozheva ID, Lomzie MA, Mosberg HI (2006) Positioning of proteins in
31:4901–4907. membranes: A computational approach. Protein Sci 15:1318–1333.

Moon and Fleming PNAS ∣ June 21, 2011 ∣ vol. 108 ∣ no. 25 ∣ 10177
Supporting Information
Moon and Fleming 10.1073/pnas.1103979108
SI Methods instead of working with mixed micelles. Being a serine hydrolase,
Analysis of Denaturant Titrations. Fig. S1B shows an example of a OmpLA is not active at the acidic pH of our normal reversible
two-state folding/unfolding model (Eq. 1) that relates to ΔG∘w;l to folding experiments, so we performed all activity measurements
the variation of observed fluorescence emission intensities (Y obs ) at pH 8.0. At that pH, the behavior of the folded state of OmpLA
to the concentrations of guanidine HCl ([D]) fit to a titration of in 1.0 M guanidine HCl is the same as it is at pH 3.8, as judged by
the wild-type outer membrane phospholipase A (OmpLA). This tryptophan fluorescence emission and SDS-PAGE. The buffer for
model does not describe our data well. our samples at pH 8.0 was 100 mM glycine-glycine, 2 mM EDTA.
We started the activity assay with protein samples from the
Y obs ð½D
Þ ¼ ððSunf  ½D
þ Y unf;w;l Þ third step described above that were incubated for 24 h at 37 °C
in 1.0 M guanidine HCl. We diluted these samples with a buffer/
þ ððSfold  ½D
guanidine mixture such that the final protein concentration
during measurements would be 48 nM and the final guanidine
þ Y fold;w;l Þ ðexpð−ðΔG∘w;l þ m ½D
Þ∕RTÞÞÞÞ∕ð1
concentration would remain at 1.0 M. We dried an aliquot of
þ ðexpð−ðΔG∘w;l þ m ½D
Þ∕RTÞÞÞ; [S1] the substrate 2-hexadecanoyl-1-ethylphosphorylcholine (HEPC,
from Cayman) briefly under nitrogen and then hydrated it to
where Y unf;w;l and Y fold;w;l are fluorescence emission intensities in 25 mg mL−1 in buffer. We added enough hydrated HEPC to the
the absence of denaturant for the unfolded and folded conforma- protein samples such that the final HEPC concentration would
tions, respectively; Sunf and Sfold are the slopes of linear baselines be 2.5 μM, which is below its reported critical micelle concentra-
in the unfolded and folded regions of the data, respectively; the tion of 3.5–4.5 μM (4). The folded protein/HEPC mixtures were
m-value is a constant that describes how steeply the protein’s free then incubated for at least 12 h at room temperature (22–24 °C)
energy depends on [D]; R is the gas constant; and T is the tem- to allow full incorporation of the substrate into the lipid bilayers.
perature in Kelvin. We also prepared blank samples: one containing protein un-
Instead, we used a three-state linear extrapolation model folded in 5 M guanidine HCl in the presence of LUVs and the
(Eq. 2) (1) to fit our titration data. other containing protein unfolded in 5 M guanidine HCl with no
LUVs. To begin activity measurements, we added the secondary
substrate 5,5′-dithiobis(2-nitro-benzoic acid) (DTNB, from Cay-
Y obs ð½D
Þ ¼ ðððSfold  ½D
þ Y fold;w;l Þ man) such that it would be at a final concentration of 0.8 mM.
þ ððexpð−ðDG∘1;w;l þ m1  ½D
Þ∕RTÞÞ ðSint  ½D
þ Y int;w;l ÞÞ For each measurement, we blanked a Beckman Coulter DU 730
spectrophotometer on the mixture following addition of DTNB.
þ ððexpð−ðΔG∘1;w;l þ m1  ½D
Þ∕RTÞÞ ðexpð−ðΔG∘2;w;l We then monitored absorbance at 412 nm over time. After 2 min
þ m2  ½D
Þ∕RTÞÞ ðSunf  ½D
þ Y unf;w;l ÞÞÞÞ∕ð1 of baseline collection, we added CaCl2 to reach a final calcium
concentration of 20 mM. Calcium initiates OmpLA’s activity on
þ ðexpð−ðΔG∘1;w;l þ m1  ½D
Þ∕RTÞÞ HEPC because it mediates dimerization of the protein (5, 6).
The hydrolysis of HEPC releases a product that then cleaves
þ ððexpð−ðΔG∘1;w;l þ m1  ½D
Þ∕RTÞÞ ðexpð−ðΔG∘2;w;l
DTNB producing a yellow moiety. Because DTT reacts with the
þ m2  ½D
Þ∕RTÞÞÞÞ; [S2] DTNB, we did not carry out activity measurements on the A210C
variant.
where Y int;w;l is the fluorescence emission intensity in the absence
of denaturant for the intermediate conformation; Sint is the slope Protease Protection Assay. Because the protease experiments were
of the linear baseline in the intermediate region of the data; analyzed by SDS-PAGE, we made changes to the normal folding
ΔG∘2;w;l and ΔG∘2;w;l are the free energies of the first and second protocol described above. Unfolded protein stocks were made in
structural transitions, respectively; and the m1 and m2 -values 10 M urea. Urea solutions were prepared from ultra pure grade
describe how steeply ΔG∘1;w;l and ΔG∘2;w;l , respectively, depend powder (Amresco) and then preincubated with AG 501-X8
on [D]. resin (BioRad) for at least 1 h prior to addition of buffer ions.
An example of a three-state fit to a titration of the wild-type The buffer was 100 mM glycine-glycine (Sigma), 10 mM taurine
OmpLA is shown in Fig. S1C. Because the intermediate baseline (Sigma), and 2 mM EDTA, pH 8.0. Glycine-glycine and taurine
region is not well resolved in the data from some of the variants, were chosen for their ability to suppress formation of and/or
the two m-values from individual fits of those datasets were scavenge cyanate ions in the urea (7). The folding reactions had
poorly determined. Therefore, we made the assumption that two steps. The first step was a dilution of unfolded protein into
our sequence substitutions would not appreciably alter the two LUVs to attain a urea concentration of 4.5 M, a protein concen-
m-values of OmpLA (2) and we globally fit all titration data from tration of 9.0 μM, and a lipid-protein ratio of 1;000∶1. The second
every sequence variant with Eq. 2 to find common measures of folding step was a further threefold dilution to yield a final urea
the two m-values shared by all variants. From the global fit, the concentration of 1.5 M and a final protein concentration of
m-value of the first transition was determined to be 3.0 μM, which is high enough to easily visualize as bands in gels.
2.03 kcal mol−1 M−1 and the m-value of the second transition To achieve high folding efficiency and prevent aggregation at this
was determined to be 7.18 kcal mol−1 M−1 . All other parameters high protein concentration (8), we used LUVs of 1,2-didecanoyl-
in Eq. 2 were determined locally for each dataset. We used Igor sn-glycero-3-phosphocholine (DDPC). Following 24 h of incuba-
Pro v6.12 (www.wavemetrics.com) for all model fitting routines. tion at 37 °C with gentle rotation, two 100 μL aliquots of each
folded protein sample were placed in fresh tubes at room tem-
Enzymatic Activity Assay. We measured activity in a similar way to a perature. We added 1.24 μL of trypsin (1 mg mL−1 in 2 mM
previously reported method (3), except that we aimed to preserve CaCl2 and 1 mM HCl) into one of the two aliquots for each sam-
the bilayer structure of our large unilamellar vesicles (LUVs) ple. The other aliquot was left undigested. All aliquots were then

Moon and Fleming www.pnas.org/cgi/doi/10.1073/pnas.1103979108 1 of 6

incubated for an additional 12 h at room temperature prior to relative to the transfer of segments from water into lipid bilayers.
SDS-PAGE. We quenched the samples by mixing them at a The ΔCONH value was set to its default of 0 kcal mol−1 . The hy-
4∶1 ratio with 5× gel loading buffer (5) and then split each sample dropathy window was 19 residues. All aspartic acid, glutamic acid,
into two portions. For each sample, we immediately placed one and histidine residues were considered to be protonated because
portion on ice and then boiled the other portion for 7 min. Other- our scale was measured at pH 3.8. The known transmembrane
wise, the quenched samples were never frozen and were never segments of rhodopsin were identified using the Orientations
warmed above 4 °C until after electrophoresis. Gels were stained of Proteins in Membranes database (http://opm.phar.umich.
with GelCode Blue (Pierce) and were then imaged with an Epson edu/) (10).
4490 flatbed scanner, using its transparency lamp.
Accessible Surface Area Calculations. Representations of Glycine-
Hydropathy Analysis. We used the Membrane Protein Explorer X-Glycine tripeptides (X ¼ A, F, L, I, Y, V, and M) were con-
(MPEx) application (http://blanco.biomol.uci.edu/mpex) (9) to structed in PyMol (DeLano Scientific). The ASA for each result-
measure hydropathy values and to predict transmembrane seg- ing structure was calculated using the program calc-surface (11)
ments of bovine rhodopsin (3cap.pdb). Hydropathy values were using the default probe size of 1.4 Å.

1. Latypov RF, Cheng H, Roder NA, Zhang J, Roder H (2006) Structural characterization of 7. Lin MF, Williams C, Murray MV, Conn G, Ropp PA (2004) Ion chromatographic quanti-
an equilibrium unfolding intermediate in Cytochrome c. J Mol Biol 357:1009–1025. fication of cyanate in urea solutions: Estimation of the efficiency of cyanate scavengers
2. Huysmans GHM, Baldwin SA, Brockwell DJ, Radford SE (2010) The transition state for for use in recombinant protein manufacturing. J Chromatogr B 803:353–362.
folding of an outer membrane protein. Proc Natl Acad Sci USA 107:4099–4104.
8. Burgess NK, Dao TP, Stanley AM, Fleming KG (2008) β-Barrel proteins that reside in the
3. Dekker N, Tommassen J, Lustigi A, Rosenbusch JP, Verheij HM (1997) Dimerization
Escherichia coli outer membrane in vivo demonstrate varied folding behavior in vitro.
regulates the enzymatic activity of Escherichia coli outer membrane phospholipase
J Biol Chem 283:26748–26758.
A. J Biol Chem 272:3179–3184.
4. Aarsman AJ, van den Bosch H (1979) A comparison of acyl-oxyester and acyl-thioester 9. Snider C, Jayasinghe S, Hristova K, White SH (2009) MPEx: A tool for exploring
substrates for some lipolytic enzymes. Biochim Biophys Acta 572:519–530. membrane proteins. Protein Sci 18:2624–2628.
5. Stanley AM, Fleming KG (2007) The role of a hydrogen bonding network in the trans- 10. Lomzie AL, Pogozheva ID, Lomzie MA, Mosberg HI (2006) Positioning of proteins in
membrane β-barrel OMPLA. J Mol Biol 370:912–924. membranes: A computational approach. Protein Sci 15:1318–1333.
6. Ubarretxena-Belandia I, Boots JWP, Verheij HM, Dekker N (1998) Role of the cofactor 11. Gerstein M, Lynden-Bell RM (1993) What is the natural boundary of a protein in
calcium in the activation of outer membrane phospholipase A. Biochemistry solution? J Mol Biol 230:641–650.
37:16011–16018.

Fig. S1. OmpLA folds reversibly at pH 3.8 in a three-state structural transition across a titration of guanidine HCl. (A) Wavelength position of maximum
fluorescence intensity is shown for samples of the wild-type OmpLA at pH 3.8 in different final concentrations of guanidine HCl. Samples were excited
by light at 295 nm. Filled blue symbols represent a set of “folding” reactions where samples of protein initially unfolded in 5 M guanidine HCl with LUVs
of 1,2-dilauroyl-sn-glycero-3-phosphocholine (DLPC) were diluted to lower final concentrations of guanidine HCl. Open red symbols represent a set of
“unfolding” reactions where samples of protein initially folded in 1 M guanidine HCl with LUVs of DLPC were diluted into higher final concentrations of
guanidine HCl. (B) Intrinsic fluorescence emission intensity at 330 nm for a similar titration as in Panel A, but with more data points. Solid line represents
a two-state reversible equilibrium fit to the data (Eq. 1). (C) Same data points as in panel B, but the solid line represents a three-state reversible equilibrium
fit to the data (Eq. 2).

Moon and Fleming www.pnas.org/cgi/doi/10.1073/pnas.1103979108 2 of 6

Fig. S2. OmpLA and its sequence variants fold and function in lipid bilayers. (A) Enzymatic activity is shown by a change in absorbance at 412 nm over time for
samples of every OmpLA variant in this study (except for A210C). The traces for the folded variants are colored with the same scheme as in Fig. 1. The variants
were first folded into LUVs of DLPC with 1 M guanidine HCl at pH 8.0 and then mixed with the substrate HEPC and a secondary substrate DTNB. After 2 min of
background data collection, calcium was added to initiate OmpLA’s cleavage of the HEPC substrate. The mixtures turned yellow upon the cleavage product
reacting further with the secondary substrate DTNB. Also shown is a lack of activity for two samples of unfolded wild-type OmpLA in 5 M guanidine HCl, one
with LUVs of DLPC (black trace) and the other without LUVs (gray trace). (B) Intrinsic tryptophan fluorescence spectra of a representative set of sequence
variants (wild-type, A210L, A210M, A210Q, A210R, and A210S) folded into LUVs of DLPC with 1M guanidine HCl at pH 3.8. The traces for these folded variants
are colored with the same scheme as in Fig. 1. Also shown are fluorescence spectra for two samples of unfolded wild-type OmpLA in 5 M guanidine HCl, one
with LUVs of DLPC (black trace) and the other without LUVs (gray trace). (C) Protection of folded OmpLA in liposomes from Trypsin digestion. Wild-type OmpLA
(Upper) and the A210R variant (Lower) were initially folded in 1.5 M urea at pH 8.0 prior to the addition of Trypsin. Bands denoted “F” contain folded protein
and bands denoted “U” contain unfolded protein. Trypsin digestion products appear 1–2 kDa smaller than undigested samples.

Moon and Fleming www.pnas.org/cgi/doi/10.1073/pnas.1103979108 3 of 6

Fig. S3. Membrane partition energies of hydrophobic residues strongly correlate with the amount of nonpolar surface area buried in the membrane. The
difference in nonpolar accessible surface area (ASA) between alanine and the hydrophobic residues F, L, I, Y, V, and M is plotted on the horizontal axis. The ASAs
were calculated using a Gly-X-Gly peptide and a rolling probe with a radius of 1.4 Å. The partition energies from our whole-protein hydrophobicity scale for the
same residues are plotted on the vertical axis. Error bars are standard errors of the mean. The solid line represents a linear fit to the data points, having a slope
of 0.023 kcal mol−1 Å−2 and intercepting the vertical axis at 0.164 kcal mol−1 .

Fig. S4. Hydropathy analysis using the whole-protein scale successfully predicts the transmembrane segments of bovine rhodopsin. Hydropathy plot for
bovine rhodopsin (3cap.pdb) was prepared using MPEx (9) with the ΔCONH and window values set to their defaults of 0 kcal mol−1 and 19, respectively.
All histidine, aspartic acid, and glutamic acid residues were considered to be protonated. Hydropathy values are for transfer from water into bilayer.

Moon and Fleming www.pnas.org/cgi/doi/10.1073/pnas.1103979108 4 of 6

Fig. S5. Comparison of two existing hydrophobicity scales to our whole-protein hydrophobicity scale. The existing scales are plotted on the vertical axis. Our
side chain transfer free energy scale from Table S1 (ΔGsc wbi ) is plotted on the horizontal axis. Data points are colored by the same scheme as in Fig. 1. Solid lines
are linear fits through the data points shown with filled symbols. Correlation coefficients for the fits are shown in the lower right of each panel. Data points
shown with open symbols were left out of the linear fits. Data points for aspartic acid and glutamic acid were left out of all the linear fits because their
protonation states are likely sensitive to the particular pH of our experiments (3.8) and that pH is not common to the other scales. (A) Translocon-to-bilayer
transfer scale (1). The data point for helix-breaker proline was left out of the linear fit because the translocon scale used α-helical transmembrane segments.
The slope of the line is 0.45 and the line intercepts the vertical axis at 1.00 kcal mol−1 , and R ¼ 0.92. (B) Wimley White side chain water-to-octanol transfer scale
(2). Two different data points are shown for aspartic acid and for glutamic acid, one each for the deprotonated state from the WW scale determined at pH 9.0
and one each for the protonated state from the WW scale determined at pH 1.0. The slope of the line is 0.50, the line intercepts the vertical axis at
−0.64 kcal mol−1 , and R ¼ 0.89.
1 Hessa T et al. (2005) Recognition of transmembrane helices by the endoplasmic reticulum translocon. Nature 433:377–381.
2 Wimley WC, Creamer TP, White SH (1996) Solvation energies of amino acid side chains and backbone in a family of host-guest pentapeptides. Biochemistry 35:5109–5124.

Table S1. Free energies of unfolding in water and lipids for OmpLA sequence variants at position 210
from the two equilibrium structural transitions during guanidine HCl titrations
First transition Second transition Total ΔΔG∘w;l ΔGscwbi
Variant ΔG∘w;l (kcal mol−1 )* ΔG∘w;l (kcal mol−1 )† ΔG∘w;l (kcal mol−1 ) (kcal mol−1 )ΔG∘w;l ‡ (kcal mol−1 )§
WT (A) 6.49 ± 0.12 25.97 ± 0.04 32.45 ± 0.11 −1.57¶
A210C 6.31 ± 0.04 25.66 ± 0.13 31.97 ± 0.09 0.49 ± 0.15 −1.08
A210D 5.02 ± 0.12 24.49 ± 0.02 29.51 ± 0.14 2.95 ± 0.18 1.38
A210E 5.71 ± 0.02 25.11 ± 0.01 30.82 ± 0.03 1.64 ± 0.12 0.07
A210F 5.94 ± 0.35 28.71 ± 0.10 34.65 ± 0.24 −2.20 ± 0.27 −3.77
A210G 6.02 ± 0.16 24.72 ± 0.13 30.74 ± 0.03 1.72 ± 0.12 0.15
A210H 3.86 ± 0.28 23.83 ± 0.05 27.69 ± 0.24 4.76 ± 0.26 3.19
A210I 5.37 ± 0.35 28.64 ± 0.02 34.01 ± 0.33 −1.56 ± 0.35 −3.12
A210K 3.15 ± 0.50 23.92 ± 0.00 27.07 ± 0.51 5.39 ± 0.52 3.82
A210L 6.09 ± 0.09 28.18 ± 0.16 34.27 ± 0.07 −1.81 ± 0.13 −3.32
A210M 6.45 ± 0.01 26.76 ± 0.14 33.21 ± 0.15 −0.76 ± 0.19 −2.33
A210N 4.95 ± 0.19 24.03 ± 0.07 28.98 ± 0.26 3.47 ± 0.28 1.91
A210P 6.95 ± 0.01 27.02 ± 0.06 33.97 ± 0.05 −1.52 ± 0.12 −3.09
A210Q 5.08 ± 0.07 24.37 ± 0.06 29.45 ± 0.02 3.01 ± 0.11 1.44
A210R 4.53 ± 0.03 24.21 ± 0.09 28.74 ± 0.07 3.71 ± 0.13 2.14
A210S 6.10 ± 0.07 24.52 ± 0.13 30.62 ± 0.19 1.83 ± 0.22 0.26
A210T 6.18 ± 0.26 24.49 ± 0.06 30.67 ± 0.32 1.78 ± 0.34 0.21
A210V 5.61 ± 0.32 27.62 ± 0.06 33.23 ± 0.26 −0.78 ± 0.28 −2.34
A210W 5.49 ± 0.09 27.35 ± 0.09 32.84 ± 0.17 −0.38 ± 0.21 −1.95
A210Y 6.71 ± 0.10 26.83 ± 0.07 33.55 ± 0.03 −1.09 ± 0.12 −2.66
Standard errors of the mean are shown from independent titrations (n ¼ 2).
*An equilibrium m-value of 2.03 kcal mol−1 M−1 for the first structural transition was determined by a global fit to all
guanidine titrations for all sequence variants used in this study.
†
An equilibrium m-value of 7.18 kcal mol−1 M−1 for the second structural transition was determined by a global fit to all
guanidine titrations for all sequence variants used in this study.
‡
Change in stability with respect to the wild type.
§
Water-to-bilayer (wbi) transfer free energies for the amino acid side chains (sc) determined by subtracting the transfer
free energy of alanine from the ΔΔG∘w;l of the side chain variants.
¶
Transfer free energy of alanine comes from its nonpolar ASA in a model tripeptide (69.1 Å2 ) multiplied by the slope of
the line in Fig. S3 (0.023 kcal mol−1 Å−2 ).

Moon and Fleming www.pnas.org/cgi/doi/10.1073/pnas.1103979108 5 of 6

 Table S2. Free energies of unfolding in water and lipids for OmpLA sequence variants at
different membrane depths and for a double arginine variant from the two equilibrium
structural transitions during guanidine HCl titrations
First transition Second transition Total ΔΔG∘w;l
Position/Variant ΔG∘w;l (kcal mol−1 )* ΔG∘w;l (kcal mol−1 )† ΔG∘w;l (kcal mol−1 ) (kcal mol−1 )
120
L120A 5.20 ± 0.11 25.10 ± 0.08 30.03 ± 0.03
WT 6.49 ± 0.12 25.97 ± 0.04 32.45 ± 0.11 −2.16 ± 0.09‡
L120R 3.34 ± 0.16 24.60 ± 0.05 27.95 ± 0.11 2.35 ± 0.12‡
164
WT 6.49 ± 0.12 25.97 ± 0.04 32.45 ± 0.11
A164L 6.84 ± 0.09 26.82 ± 0.10 33.66 ± 0.19 −1.21 ± 0.21§
A164R 6.00 ± 0.27 25.65 ± 0.02 31.66 ± 0.26 0.80 ± 0.27§
210
WT 6.49 ± 0.12 25.97 ± 0.04 32.45 ± 0.11
A210L 6.09 ± 0.09 28.18 ± 0.16 34.27 ± 0.07 −1.81 ± 0.10§
A210R 4.53 ± 0.03 24.21 ± 0.09 28.74 ± 0.07 3.71 ± 0.11§
212
G212A 5.36 ± 0.34 27.70 ± 0.20 33.06 ± 0.14 −0.60 ± 0.16§
G212L 5.57± 0.18 30.11 ± 0.02 35.68 ± 0.16 −2.62 ± 0.21¶
G212R 5.45 ± 0.01 24.55 ± 0.07 30.00 ± 0.06 3.06 ± 0.16¶
214
Y214A 5.68 ± 0.00 24.40 ± 0.01 30.09 ± 0.01 2.37 ± 0.08§
Y214L 6.01 ± 0.05 25.29 ± 0.02 31.30 ± 0.03 −1.21 ± 0.04∥
Y214R 5.39 ± 0.03 24.08 ± 0.01 29.47 ± 0.02 0.61 ± 0.02∥
223
WT 6.49 ± 0.12 25.97 ± 0.04 32.45 ± 0.11
A223L 6.33 ± 0.08 27.96 ± 0.01 34.29 ± 0.07 −1.83 ± 0.11§
A223R 5.50 ± 0.11 24.88 ± 0.11 30.38 ± 0.22 2.07 ± 0.23§
Double ARG
G212A 5.36 ± 0.34 27.70 ± 0.20 33.06 ± 0.14
A210R, 212R 4.47 ± 0.07 23.46 ± 0.06 27.93 ± 0.01 5.13 ± 0.14¶
Standard errors of the mean are shown from independent titrations (n ¼ 2).
*An equilibrium m-value of 2.03 kcal mol−1 M−1 for the first structural transition was determined by a global
fit to all guanidine titrations for all sequence variants used in this study.
†
An equilibrium m-value of 7.18 kcal mol−1 M−1 for the second structural transition was determined by a
global fit to all guanidine titrations for all sequence variants used in this study.
‡
Change in stability with respect to the L120A variant.
§
Change in stability with respect to the wild type.
¶
Change in stability with respect to the G212A variant.
∥
Change in stability with respect to the Y214A variant.

Moon and Fleming www.pnas.org/cgi/doi/10.1073/pnas.1103979108 6 of 6