Journal Pre-proof

Lysozyme thermal stability in the presence of cyclodextrins at

different pH values

A. Jessica Díaz-Salazar, Daniel Ondo

PII: S0301-4622(25)00081-X
DOI: [Link]
Reference: BIOCHE 107469

To appear in: Biophysical Chemistry

Received date: 27 March 2025

Revised date: 16 May 2025
Accepted date: 16 May 2025

Please cite this article as: A.J. Díaz-Salazar and D. Ondo, Lysozyme thermal stability
in the presence of cyclodextrins at different pH values, Biophysical Chemistry (2024),
[Link]

This is a PDF file of an article that has undergone enhancements after acceptance, such
as the addition of a cover page and metadata, and formatting for readability, but it is
not yet the definitive version of record. This version will undergo additional copyediting,
typesetting and review before it is published in its final form, but we are providing this
version to give early visibility of the article. Please note that, during the production
process, errors may be discovered which could affect the content, and all legal disclaimers
that apply to the journal pertain.

© 2025 Published by Elsevier B.V.

 Journal Pre-proof

Lysozyme Thermal Stability in the Presence of Cyclodextrins

at Different pH Values

A. Jessica Díaz-Salazara, Daniel Ondob,*

a
Laboratorio de Biofisicoquímica, Departamento de Fisicoquímica, Facultad de Química,

of
Universidad Nacional Autónoma de México (UNAM), Mexico City 04510, Mexico
b
Department of Physical Chemistry, University of Chemistry and Technology, Technická 5, 166 28

ro
Prague 6, Czech Republic

-p
re
Abstract
lP

In the present study, the primary action of native cyclodextrins (CDs) on lysozyme protein as

binding ligand and secondary as aggregation inhibitor were probed. Thermally induced
na

denaturation using differential scanning calorimetry (DSC) was measured in the presence of native

 -,  - and  -CDs. The denaturation process in CD absence was reversible to 60–80% at pH  6

ur

with maximum Tm at pH = 4. Denaturation in the presence of native  -CD at pH from 2 to 10, at

Jo

the least stable and partially reversible conditions in presence of  -CD and  -CDs at single pH 2

only, was measured. The protein thermal stability decreases in the presence of CDs, with the most

evident for  -CD, followed by  -CD and almost no effect for  -CD. The reversibility in the

presence of  -CD was similar to that in its absence. The best protection performance against

*
Correspondence to: Dr. Daniel Ondo, Department of Physical Chemistry, University of Chemistry and Technology,
Technická 5, 166 28 Prague 6, Czech Republic
Email address: [Link]@[Link] (Daniel Ondo)
 Journal Pre-proof

heat-induced denaturation was found at pH 2 for  -CD. The heat capacity data for  -CD at

acidic pH were fitted by the protein-ligand binding model in the whole temperature and ligand

concentration ranges studied. The decrease in thermal stability for  -CD at all pH,  - and 

-CD at pH 2 were fitted linearly as a function of ligand concentration. The CD-to-lysozyme

binding parameters obtained in this work and from the literature for other CDs are briefly

discussed using the concept of cyclodextrin cavity size, charge distribution, solvent accessible

surface area and amino acid hydrophobicity.

of
ro
Keywords:

-p
cyclodextrin, differential scanning calorimetry, inclusion complex lysozyme binding model
re
lP

1. Introduction

Cyclodextrins (CDs) are cyclic oligosaccharides; the most common types are  -,  - and  -CD,
na

which contain six, seven, and eight  -1-4-linked D-glucopyranosyl units, respectively. The
ur

conformation of glucopyranose units leads to cyclodextrins having a truncated cone-like structure

Jo

with a hydrophobic inner cavity and a highly hydrophilic exterior [1, 2]. This geometry allows

CDs (host) to trap hydrophobic molecules (guest) of varied structures to form inclusion complexes

[3–5]. These complexes can be formed by noncovalent interactions such as hydrogen bonds,

hydrophobic interactions, van der Waals forces, and dipole-dipole interactions, depending on the

properties of the guest, the chosen cyclodextrin, and the solvent [2, 6, 7].

Proteins are one of the major constituents of the living cell membrane and are responsible

for transport, cell adhesion, and reception. CDs can extract proteins from cell membranes [4],

suppress their aggregation [8], inhibit fibrillation and rupture of already mature amyloids [9],
 Journal Pre-proof

discriminate between homologous proteins [10] and were proposed for the membrane protein

reconstitution process [11] by stripping the surfactant.

Cell membranes, pharmaceutical formulations, and human fluids comprise complex

reaction matrices in which the equilibria between individual reactants must be determined and

dissected into contributions that combine into overall effects. For the purpose of predicting ligand

binding in complex mixtures, it is desirable to have a rough model linking the experimental

observables to globular protein structural markers in simple systems. As such, in addition to NMR

of
spectroscopy [8, 10, 12, 13] and circular dichroism [14], differential scanning calorimetry (DSC)

ro
proved to be a method of choice for cyclodextrin-protein interaction [3, 15, 16] and any

-p
ligand-protein thermal studies [17–20]. Applying the ligand binding model [3, 21, 22] to primary
re
acquired data allowed CDs to identify as ligands, which preferentially stabilize the unfolded

protein state and are more effective denaturants than osmolytes [3]. Since the lysozyme is a protein
lP

with a well-defined structure, biological functions, and relative stability that generally does not
na

aggregate in dilute aqueous solutions [23], it is an ideal candidate for ligand-biomacromolecules

interaction studies [3, 14–16, 24, 25].

ur

In this work, we probed the thermal stability of lysozyme from chicken egg white (LYZ) in
Jo

the presence of all three native cyclodextrins. Changes in the thermal stability of LYZ at different

pH values and as a function of cyclodextrin concentration were determined using differential

scanning calorimetry (DSC). To obtain the number of protein binding sites (residues) involved in

the cyclodextrin-protein inclusion complex formation, the heat capacity data and change of

melting temperature are analyzed by two variants of the thermodynamic binding model. To the

best of our knowledge, the presently developed binding model as function of ligand concentration

and temperature has not yet been applied on cyclodextrin-to-protein interaction. The obtained
 Journal Pre-proof

cyclodextrin-to-protein binding parameters are discussed in the frame of the residue

hydrophobicity, solvent-accessible surface area, individual cyclodextrin-to-residue and sequence

structural concept. The potential of cyclodextrins to inhibit protein aggregation is also

investigated.

2. Materials and methods

of
2.1 Chemicals

ro
Chicken egg white lysozyme (LYZ, L6876 Lot nr. SLBH9534V),  -cyclodextrin (  -CD, ord.

-p
nr. C4642, water content 10% w/w determined by Karl Fisher titration), and  -cyclodextrin ( 

-CD, ord. nr. C4767, water content 11% w/w) were purchased from Sigma-Aldrich, USA. 
re
-Cyclodextrin (  -CD, Lot nr. F8045-6, water content 9.7% w/w) was obtained from Cerestar,
lP

USA. Sodium phosphate monobasic monohydrate and sodium phosphate dibasic heptahydrate
na

were purchased from J.T. Baker, USA. Deionized water with resistivity 18.2 M  cm was obtained

from a Millipore system (Simplicity, Millipore, France).

ur
Jo

2.2. Sample preparation

Phosphate solutions were prepared at a concentration of 25 mM at pH = 2 , 3, 4, 6, 7, 8, 9, and 10,

and carefully adjusted to the desired value using a calibrated pH-meter (Oakton pH 700, Eutech

Instruments, Singapore). The 90 mM  -CD, 16 mM  -CD and 100 mM  -CD stock solutions

were prepared by weighing the powder into a 10 mL volumetric flask, filled by phosphate solution

until mark and dissolved. Before further dilution, the stock solutions were sonicated for 15 min at

50 C in an ultrasonic bath (Cole-Parmer, 08892-1, Illinois, USA). The  -CD was diluted by the
 Journal Pre-proof

phosphate solution at each pH to concentrations: 5, 10, 15, 20, 30, 40, 50, 60, 70, and 80 mM. The

 -CD was diluted by the phosphate solution at pH = 2 to concentrations: 1, 3, 5, 7, 10, 12, and

13.8 mM. The  -CD was diluted by the phosphate solution at pH = 2 to concentrations: 5, 10,

20, 30, 40, 50, 60, 70, 80, and 90 mM. The LYZ protein lyophilized powder was dissolved to 4

mg/mL in phosphate solution at the desired pH. The protein stock solution was further dialysed

(Spectra/Por 3 dialysis membrane, standard RC tubing, MWCO 3.5 kDa) against phosphate

solution of the same pH at 100 fold excess, in the fridge at a temperature of 4 C for 24 h. The

of
phosphate solution was exchanged for a fresh one after 4 h. The final exact protein concentration

ro
was determined by absorbance measurements at 280 nm with a UV-visible spectrophotometer

-p
(CARY 50 Bio, UV0909M003, Varian, Australia) using the extinction coefficient averaged value
re
38300 M 1 cm 1 [26].
lP

2.3. Differential scanning calorimetry

na

The protein thermal stability was probed by using a VP-capillary differential scanning calorimeter
ur

(VP-cap DSC, Microcal, Malvern) equipped with autosampler and tantalum cells of reaction
Jo

volume of 137  L. 450  L of prepared sample solution was consumed due to the tubing dead

volume connecting the cell dispenser and the sample cell. The heat capacity measurements were

performed at 90 C/h heating rate in the temperature range from 10 to 100 C; only the pH 6

experiments were performed from 10 to 90 C. The thermal stability of LYZ in the absence of CD

was probed at pH = 2 , 3, 4, 6, 7, 8, 9, and 10. Protein thermal stability in the presence of  -CD

was probed at 5, 10, 15, 20, 30, 40, 50, 60, 70, and 80 mM  -CD concentrations, but at pH 6 at

10, 20, 40, 60 and 80 mM CD. The experiments with  - and  -CDs were performed only at pH
 Journal Pre-proof

2, the CD concentrations used were as follows: 1, 3, 5, 7, 10, 12 and 13.8 mM for  -CD [27]; and

5, 10, 20, 30, 40, 50, 60, 70, 80 and 90 mM for  -CD. Each heat capacity measurement

C p = f (T , ct ,CD ) at a selected pH consisted of two independent runs: (1) to obtain the protein

thermal stability, the sample cell was filled with the protein solution, (2) to obtain an instrumental

baseline, the sample cell was filled with the phosphate solution, keeping in both (1) and (2) the

reference cell filled with the identical phosphate solution. The partial molar heat capacity C p

of
values of protein denaturation for further modeling were obtained by subtracting trace (2) from the

ro
trace (1). The reversibility is defined as percent ratio of the total area of the second heating cycle to

area of the first heating cycle, in most cases the first cycle ends at 100 C if not stated other.
-p
re
2.4. Thermodynamic binding model
lP

Full mathematical derivation of the thermodynamic binding model [3, 21, 22, 28] and detailed
na

discussion of assumptions (i)-(iv) is provided in Section S.1 of SI. The aqueous solution consist of

two reactants ligand cyclodextrin (CD) and protein (P), where (i) protein is present in two states
ur

native (folded, N), unfolded (U), with possible number of ligand binding sites to native state nF
Jo

and unfolded state nU , respectively, and their difference n(= nU  nF ) ; CD ligand binds (ii) to

both protein N and U states [29–31]; amino acids forming protein sequence bind weakly with CDs

(iii) and are considered as identical binding sites in both N and U form (iv); the data are

standardized to protein native state (see Section S.1.2). Since in most of the experiments

performed, the ligand (CD) is in greater excess as protein ( ct ,CD ct ,P ), in the calculation the free

ligand concentration [CD] can be numerically approximated by total ligand concentration ct ,CD (

[CD]  ct ,CD , see Section S.1.3).

 Journal Pre-proof

Below we list only the working equations to fit the C p = f (T , ct ,CD ) at selected pH. The

enthalpy difference is calculated by Eq. (S.36)

K d (1  Kb [CD])n  Kb [CD] 
H HN =  H  b H n
n  d  = f (T ) (1)
1  K d (1  Kb [CD])  1  Kb [CD] 

and for the experimentally determined heat capacity C p Eq. (S.37) reads

( H  H N )i  ( H  H N )i 1
C p ,i = C p ,i  C p ,N = , (2)
Ti  Ti 1

of
where C p ,i is at averaged temperature (Ti  Ti 1 ) / 2 , ( H  H N )i and ( H  H N )i 1 are the

ro
H  H N values calculated numerically at Ti and Ti 1 by Eq. (1), respectively. In the above

-p
expressions, the temperature dependence Kd (T ) and d H (T ) of pure protein denaturation in
re
the forms of integrated van’t Hoff (S.27) and Kirchoff (S.28) eqs., respectively, was considered.
lP

The temperature dependence of Kb (T ) and b H (T ) for CD binding to protein is described by

na

Eqs. (S.29) and (S.30), respectively. To fit the heat capacity curves C p = f (T , ct ,CD ) at selected

pH value to Eqs. (1) and (2), 6 adjustable parameters were considered: tm ,0 ,  d H 0 and  d C p as
ur

parameters of Eqs. (S.27) and (S.28), Kb,0 , b H0 as parameters of Eqs. (S.29) and (S.30), and
Jo

n . The experimentally obtained C p were fitted to Eqs. (1) and (2) by considering C p ,N below

Tm,0 as linear function of T [32].

At low ligand concentrations and moderate binding, the T -dependent binding model

defined by Eqs. (1) reduces to measurement of the change of thermal stability [3, 21]

Tm (= Tm  Tm ,0 ) = tm (= tm  tm ,0 )  nK b RTm2,0 [CD] /  d H 0 , (3)

where Tm and Tm,0 are the T values at which Kd =1 in the presence and absence of the ligand
 Journal Pre-proof

at that pH, respectively, the  d H 0 is the enthalpy of protein denaturation in absence of the ligand

at Tm,0 . The product nKb is considered as a single adjustable parameter from measured

Tm = f ([CD]) and is defined at Tm,0 .

2.5. Molecular modeling

The solvent accessible surface area (SASA) for folded LYZ was calculated using the Gromacs

of
package [33] for the entry of the lysozyme crystal structure 1IEE taken from PDB [34]. The SASA

ro
for unfolded form was calculated for pulled structure, distrupting the disulfide bridges. The SASA

-p
was calculated for 1.4 Å diameter solvent water molecule. The protein charge as function of pH for

folded and unfolded state was calculated using the propKa version 3.5.1 [35, 36]. The individual
re
residue p Ka values obtained by propKa are listed in Table S.2.
lP
na

3. Results and discussion

ur

3.1. Lysozyme in the absence of cyclodextrins

Jo

The thermal stability of LYZ in the absence of CD was probed at pH 2, 3, 4, 6, 7, 8, 9, and 10.

Exemplary heat capacity curves for measurements at pH 2, 3, and 4 are depicted in Figure 1A.

Increasing the pH from 2 to 4 causes increased thermal protein stability by at least 20 C. The tm ,0

(temperature at which Kd =1 , fraction of native = fraction of denatured) is evaluated at

max C p = f (T ) as a function of pH is shown in Figure 1B. The highest thermal stability of the

LYZ protein in the phosphate solution was observed at pH 4. At higher pH values, the thermal

stability decreased by approximately 10 C from the maximal value.

 Journal Pre-proof

Figure 1B shows the reversibility of LYZ protein heat induced denaturation as a function

of pH examined by the second heating cycle of the same cell content. The reversibility is relatively

high (60% to 85%, first heating to 100 C). However, the maxima occur at a higher pH value, equal

to 6. If first heating cycle only to 80 C the reversibility at pH 2 is 88%. At even more basic pH, the

heat induced denaturation was found irreversible as found for other similar experimental

conditions [37, 38].

The measured heat capacity of LYZ denaturation as a function of pH provides a great

of
opportunity to test our binding model described in Section 2.4 above and in more detail in SI. In

ro
this test, the proton was considered as the ligand and its concentration was calculated from actual

pH value ( c
H -p
= 10  pH ). The quality of the fit by Eqs. (1) and (2) and the parameter numerical
re
values, are shown in Figure 1A and listed in Table 1, respectively. Totally 370 heat capacity C p
lP

data points, from three data sets at pH 2, 3, and 4 were fitted simultaneously with tm ,0 ,  d H 0 ,

 d C p , n , Kb,0 , and b H0 as adjustable parameters, describing C p = f (t , c  ) all the

na

H H

temperature ranges studied. The standard deviation of the fit s calculated by Eq. (S.32) is 0.6
ur

kJ/mol/ C. It is seen, that the experimental data are well represented by the model, implicating that
Jo

in the absence of CD the lysozyme thermal stability only depends on the protons available in the

media that bind to the protein. Due to the low ligand proton concentration at pH 6, the heat

capacity curve C p = f (t ) calculated by the binding model with parameter values in Table 1 does

not fit experimental points. Hence, it was not considered as a possible fourth dataset into

simultaneous fitting. For comparison, Table 1 lists the values for parameters for LYZ denaturation

in glycine buffer [22].

The present parameter values for protein denaturation in phosphate are slightly lower than
 Journal Pre-proof

for denaturation in glycine buffer. The proton in the phosphate medium binds to LYZ with higher

affinity and with more negative enthalpy than in a glycine buffer. Figure 1C compares the total

charge of folded and unfolded protein as a function of pH for theoretically calculated and

experimentally determined by propKa [35, 36] and potentiometric titration method [32] (in

absence of buffer), respectively. The fitted n agrees well with the theoretically calculated (
H

 3 ) and experimentally determined (  4.8 ) value. In summary, the model describes the binding

of the ligand to the protein in thermodynamically consistent matter in a wider temperature range

of
(70 C) and ligand concentrations (2 orders of magnitude).

ro
-p
Figure 1: Effect of pH on LYZ thermal stability in the absence of CD. LYZ heat capacity C p as a
re
function of temperature t (A) at several pH (2, 3, and 4). Experimental C p values (symbols) and
lP

fitting by ligand (proton) biding model (solid lines) Eqs. (1) and (2) with binding parameters listed

in Table 1. Effect of pH on LYZ (B) tm ,0 (left axis, points) and reversibility of the thermal
na

denaturation process for LYZ as a function of pH (right axis, bars). Change of LYZ protein charge
ur

(C) calculated by propKa 3.0 for PDB entry 1IEE (solid lines, left axis, pKa values for individual
Jo

residue are listed in Table S.2 in SI), and experimentally determined proton uptake at 25 C and

70 C [32] (dashed lines, right axis).

Table 1: Parameters tm ,0 ,  d H 0 ,  d C p , n , Kb,0 and b H0 of Eqs. (1), (S.27), (S.28), (S.29)

and (S.30) for binding of proton to LYZ determined in this work and in comparison with literature

[22].
 Journal Pre-proof

buffer tm ,0 d H 0 d C p n Kb,0 b H0 ref

H

phosphate 77.7 538.1 2.49  5.1 2217 -21 this work

glycine 78.8 568.5 5.43 5 1288 -9.2 [22]

tm ,0 in C, H in kJ/mol, C p in kJ/mol/ C, K b in M 1 .

3.2. Lysozyme in the presence of cyclodextrins

of
The thermal stability of LYZ in the presence of  -CD was probed at pH 2, 3, 4, 6, 7, 8, 9, and 10,

ro
compared to  - and  -CD at a single pH 2 value. Exemplary heat capacity curves at two selected

 -CD concentrations at each pH for measurements at pH 2, 3, 4, and 6 are shown in Figure 2. On

-p
first glance, it is seen that CD addition decreases the protein thermal stability and decreases total
re
heat for denaturation. The most notable stability decrease is up to 4.3 C for LYZ in 80 mM 
lP

-CD at pH 2.

The measured heat capacity of LYZ denaturation as a function of  -CD concentration

na

and temperature t at pH 2, 3, 4, and 6 were fitted by the ligand binding model described in
ur

Section 2.4 above and in more detail in Section S.1. In this model, CD was considered the single
Jo

ligand, and its free concentration [CD] was approximated by total CD concentration ct ,CD , known

from solution preparation and after dialysis (for details of this approximation see Section S.1.3 in

SI). C p datasets for all  -CD-LYZ mixtures with datasets without CD were used for the fitting

at a single pH. The datasets were fitted simultaneously with tm ,0 ,  d H 0 ,  d C p , n , Kb,0 , and

b H0 as adjustable parameters, describing C p = f (t ,[CD]) in whole studied temperature

ranges. Totally 400 (pH 6) and > 900 (other pH) heat capacity C p data points were used in the

fitting.
 Journal Pre-proof

The quality of the fit by Eqs. (1) and (2), and the parameter numerical values, are shown in

Figure 2 and listed in Table 2, respectively. Table 2 also lists the values of nKb (tm,0 ) , the value of

the nKb product calculated at tm ,0 from T -dependence of Kb,0 by Eq. (S.29). Due to clarity,

the calculated curves in Figure 2 are shown only for three selected C p datasets (  300 data

points). As seen, the experimental data are well represented by the model. The standard deviation

of the fit s calculated by Eq. (S.32) ranged from 1.0 to 1.4 kJ/mol/ C, the highest differences

of
between the calculated and the experimental data being around the tm point, and at t > tm before

ro
the tail of the calorimetric peak. The highest tm ,0 ,  d H 0 , and Kb,0 values were obtained at pH 4;

-p
on the contrary, at the same pH, the lowest  d C p . The most steep temperature dependence of Kb
re
was found at pH 6 (most exothermic b H0 ). The lowest n was obtained at pH 3, followed by
lP

pH 6, and the highest at pH 2.

na

Figure 2: LYZ denaturation heat capacity C p as function of temperature t at three selected 

ur

-CD concentrations at pH 2 (A), pH 3 (B), pH 4 (C) and pH 6 (D). Experimental C p data

Jo

(symbols) and fitted values (solid lines) using the binding model by Eqs (1) and (2). Fitted

parameter values are listed in Table 2, plotted every second C p point due clarity.

Table 2: Parameters tm ,0 ,  d H 0 ,  d C p , n , Kb,0 and b H0 of Eqs. (1), (S.27), (S.28), (S.29)

and (S.30) for LYZ denaturation in presence of  -CD, and for CD binding as function of pH. The

nKb (tm,0 ) is nKb product calculated at tm ,0 .

pH tm ,0 d H 0 d C p n Kb,0 b H0 nKb (tm,0 )

 Journal Pre-proof

2 54.7 384.2 3.6 8.3 6.6 -23.1 24

3 72.6 482.9 2.9 1.2 139 -26.7 38

4 77.1 530.5 2.6 2.6 228 -58.3 18

6 74.5 522.7 3.5 4.0 188 -63.4 19

tm ,0 in C, H in kJ/mol, C p in kJ/mol/ C, K b in M 1 .

The change in thermal stability tm as a function of [CD] for  -CD at all studied pH,

of
and for  - and  -CDs at pH 2 were fitted by simplified linear Eq. (3) and are plotted in Figure

ro
3A and 3B, respectively. The nKb (tm,0 ) numerical values for all studied pH and native CDs in

-p
this work are listed in Table 3 and Figure 3A and B. The steepness increases in the order  -CD <
re
 -CD (pH 6–8) <  -CD (pH 2) <  -CD. The nKb values at pH 2, 3, 4, and 6 obtained by
lP

linear fit tm = f ([CD]) are close to the values obtained by binding model fit C p = f (t ,[CD]) .

The biggest discrepancy being for pH 3, which is probably due to slightly outlier tm points at
na

ct , CD  20 mM (see Figure 3A). Figure 3C depicts the change in thermal stability tm as a
ur

function of [CD] due addition of native or modified CDs [14–16, 39]. For some of them [15, 16],
Jo

the evaluation of nKb parameter for binding of CDs with LYZ was enabled and are listed in

Table 4 for fast comparison. From Figure 3C and nKb values in Table 4 it is seen, that the most

pronounced effect on tm was found for  -CDs (native, Ac-, RaMe-), followed by HP-  -CD

and native  -CD and almost no effect for  -CD. The molecular interpretation of cyclodextrin

binding will be discussed in following Section.

Figure 3: LYZ thermal stability changes tm as a function of cyclodextrin concentration and
 Journal Pre-proof

type. Data from this work (A) for  -CD at different pH values and (B) for  -,  - and  -CD at

pH 2 (25 mM phosphate solution). Solid lines in (A) and (B) are linear fits by Eq. (3) with

numerical nKb parameter values listed in Table 3. Data from literature (C) Tavornvipas et al.

(legend code 06TavHir, phosphate buffer, pH 4, I = 0.05 , [15]), Yamamoto et al. (legend code

06YamFuk, phosphate buffer, pH 7, [14]), Kamiyama et al. (legend code 12KamSat, no buffer,

[16]), Arsiccio et al. (legend code 23ArsSar, citrate buffer, pH 3, [39]), numerical values of nKb

of
are listed in Table 4.

ro
Table 3: Binding parameter nKb of Eq. (3) for CDs to lysozyme and protein reversibility at the

-p
highest CD concentration examined due thermal denaturation at varied pH.
re
CD pH nKb (M 1 ) Reversibility / %
lP

 -CD 2 23.7  0.5 60

3 21.4  0.8 76
na

4 16.3  0.4 83
ur

6 14.3  0.6 85

7 14.0  0.4 n.d.

Jo

8 13.8  0.5 n.d.

9 18.7  0.6 n.d.

10 19.0  0.3 n.d.

 -CD 2 46.0  1.6 95 a

 -CD 2 3.6  0.5 56

a
First heating cycle ends at 80 C.
 Journal Pre-proof

Table 4: Binding parameter nKb of Eq. (3) for native and modified CDs to lysozyme (data from

literature).

CD nKb (M 1 ) reference

 -CD 18.9 [3] a

 -CD 14.1 [15] b

Ac-  -CD 65.2 [16] c

RaMe-  -CD 25.8

of
[16] c

HP-  -CD

ro
19.5 [16] c

HP-  -CD 28.2 [15] b

 -CD 2.5 -p [15] b

re
a
40 mM glycine buffer, pH 3.
lP

b
Phosphate buffer, pH 4, I = 0.05 , evaluated by Eq. (3) in this work.
c
Pure water, no buffer, evaluated by linear fit by Eq. (3) in this work.
na
ur

Table 3 also lists the reversibility of the heat induced denaturation of the LYZ protein as a
Jo

function of the pH examined by the second heating cycle of the same cell content. The reversibility

is higher as 50%, at pH 2 the best aggregation inhibition due to increased temperature is for native

 -CD, at other pH the  -CD is best performing as an inhibitor. The observed reversibility is

similar to that found for other CDs and at similar conditions as in present work [15, 16, 25, 37]. At

neutral pH the heat induced transitions in Me-, HP-, and Ac-  -CD were almost reversible over

90% [16]. The native CDs at 10 mM concentration inhibited aggregation to 70–85%, at higher CD

content the aggregation was inhibited even to 20 and 50% (pH 6.5), further the modified CDs

inhibited the chemically induced aggregation in the order of  -CDs <  -CDs <  -CDs [15].
 Journal Pre-proof

Similar to our results, the LYZ enzymatic activity was recovered to only 10–50% after heat shock

in presence of CDs [25].

3.3. Molecular interpretation of cyclodextrin binding

In this work employed lysozyme from chicken egg white (LYZ), is a protein with a well-defined

structure and biological functions, well soluble in water [23, 40], consisting of 129 amino acid

of
residues, molecular weight 14.3 kDa with an isoelectric point pI 11.05 (folded) [35, 36]. Its

arrangement into four  -helices (between residue 4–15, 24–37, 88–100, 108–115), three 

ro
-sheets (between residue 43–45, 51–53, 58–59) and stabilization by four disulfide bridges

-p
(between residue 6–127, 30–115, 64–80, 76–94) is depicted in Figure 4A and B with color coding
re
for secondary structures for PDB entry 1IEE. In Figure 4 are further illustrated methionines at
lP

positions M12 and M105, the M105 interacts with 3 tryptophans (position 28, 108, 111), tyrosine

(position 23) and valine (residue 99). The sequence position of 45 hydrophobic side chain residues
na

(A, V, I, L, M, F, W, P) is depicted in Figure 4C and also in Figure S.9. During the course of heat
ur

induced denaturation process the secondary and tertiary LYZ structure becomes distorted which

may serve as binding sites for possible CD complexation. In absence of CDs the four disulfide
Jo

bridges initially formed through whole protein structure break and four new between consecutive

pairs of cysteines are formed [41] due chemical and heat denaturation. Further, at higher LYZ

concentration and after several heating cycles, the irreversible formation of more stable 

-amyloids occurs [23]. Because these structural changes are difficult to correlate directly and

dissect to any structural property, the binding parameters will be discussed only tentatively to

cavity size, CD affinity to amino acid residues and peptides, and the solvent-accessible surface

area (SASA).
 Journal Pre-proof

The binding data in Tables 2–4 and tm as a function of CD concentration and type

shown in Figure 3 indicate, that the experimental conditions used (pH, buffer type, ionic strength)

play a crucial role in obtaining CD-to-protein binding parameters. Peptides and proteins are

composed of a sequence of amino acids with individual pKa values close to each other (Table S.2).

The closer the pKa values and the higher the number of chargeable residues, the more complex the

distribution of charge over the species [42]. The species distribution is similar for polyprotic acids,

bases, and amphoteric compounds, but not for the overall binding constant K B as a function of

of
pH. The closer the pKa values, the fewer plateaus are visible, leading to a steep sigmoidal curve

ro
[42]. The situation becomes more complicated if weak CD binding with charged binding sites in

-p
the protein sequence occurs, and the assumption (iv) is invalid (see Sections 2.4 and S.1.1). To
re
elucidate this, we show an exemplary calculation for short artificial model peptide in Section
lP

S.2.3. Not only does the K B = f (pH) curve diminish its sigmoidal shape and the K B

determination is cumbersome, but even the parameters become obtainable they exhibit significant
na

uncertainties of up to 50% [42]. Further, an extensive study of the effect of buffer type and of ionic
ur

strength on binding parameters for CD complexes of the negative host with positive guest showed

that increasing the ionic strength increases the binding constant, whereas the buffer effect
Jo

decreases it [43]. According to their classification, the buffers used in previous LYZ-CD studies

belong to the category with solely ionic strength effect [3, 14, 15] and moderate competitive

binding with CDs [39].

The size of CD cavity increases in the order  - <  - <  -CD [2] so the binding affinity

to amino acid residue and to protein is expected to decrease in similar order. However, the shift of

tm found in present work and in literature decreases in following order:  -CDs (native, Ac-,

RaMe-), followed by HP-  -CD and native  -CD, and almost no binding effect with  -CD.
 Journal Pre-proof

This indicates, that the medium sized  -CD cavity is best suited for binding with voluminous

hydrophobic side chains of AA (W, Y, F) as  -CD with alkyl chains (M, L, I, P). The obtained

binding parameters listed in Tables 2–4 and the observed trends follow the binding affinity for

individual AAs [30, 44–46] and peptides [29, 46, 47]: Kb,  Kb, < Kb,  < 500 , exothermic

b H in order of tenths of kJ/mol, binding of neutral guest is weaker as negatively charged. It

should be noted that some exceptions to the trends can be found in the literature: the hydrophilic

YIGSR pentapeptide (charge 2+ at pH 7) binds  -CD stronger than hydrophobic YGGFL

of
(charge 1+) [29]. Binding of the both benzene rings separated by glutamic acid with  -CD [47]

ro
was confirmed by NMR, however the entropically unfavorable bending of the rigid protein
-p
backbone is unlikely due to complexation. Extrapolating the binding parameter Kb for 
re
-CD/dipeptide interaction [47] to present tm = 72.5 C, results in n  0.1 for binding with any
lP

protein. The individual binding of phenylalanines separated by at least two aminoacids (S and P)
na

with  -CD is more likely, the penultimate being deeper penetrated into cavity as the middle

positioned [48]. For binding of tryptophan with  - and  -CDs, the preferred inclusion complex
ur

geometry comprises also the binding of benzene ring, from wider CD rim [48].
Jo

The total calculated SASA of the folded LYZ is 65.7 nm 2 , from which 34.94 nm 2

belongs to exposed nonpolar, 17.90 nm 2 to exposed polar, and 12.86 nm 2 to exposed charged

residues. Due to unfolding, the total SASA increases to 225 nm 2 , the buried AAs become exposed

to the solvent and available for complexation with CDs (see Section S.2.2). The highest SASA

increase was calculated for hydrophobic tryptophans (W28, W62, W63, W108, W111, W123),

phenylalanines (F3, F34, F38), and methionines (M12, M105). Note, that the M105 interaction

with W28, W108, W111, Y23 and V99 diminishes due unfolding, which causes further SASA
 Journal Pre-proof

increase for Y23 and V99. The SASA of asparagines N19, N113, and glycine G126 decreases due

unfolding. Of the voluminous hydrophobic residues W and F, the five W62, W63, W108, F3, and

F38 are speculated for possible  -CD binding, since others are involved in stabilization of 

helices and thus form seeds preferentially for  -amyloids [9, 23].

Figure 4: Lysozyme from chicken egg white (PDB 1IEE) with identified secondary structure in

ribbon representation. Methionine’s -CH 3 -S-CH 3 (orange): M12, M105 ; M105 interacts with

of
W28, W108, W111, Y23 and V99; 4 disulfide bridges (yellow): C6-C127, C30-C115, C64-C80,

ro
C76-C94; 4  -helices (magenta),  -sheets (yellow), 3 10 helix (blue), 45 nonpolar residue

-p
(white) and their sequence position, polar (green), positive (blue), negative (red).
re
lP

4. Conclusion
na

Employing differential scanning calorimetry, the lysozyme protein thermal stability in the absence

and presence of native cyclodextrins at varied pH values was obtained. The main conclusions
ur

reached are: (1) The protein maximum thermal stability in absence of CD occurs at pH 4, the
Jo

stability decreases upon cyclodextrin addition; (2) the Tm decreases linearly as a function of

cyclodextrin concentration, the most steep for  -CD, followed by  -CD and with almost no

effect for  -CD; (3) the thermal denaturation is partially reversible at acidic pH values, but at

basic, no matters on presence of cyclodextrins; and (4) the two ligand binding model approaches

and varied structural concepts used, allowed to dissect individual n and Kb contributions to

nKb and identify 3 to 10 most likely binding aromatic and hydrophobic residues, respectively.

The developed binding model can be further applied to analyze any interaction of carbohydrate
 Journal Pre-proof

polymers to biomacromolecules.

During the binding mechanism, the cyclodextrin cavity with bound crystal water competes

with solvent to access the binding site, dehydrate its cavity, and perform inclusion complex with

residue or short peptides. On a molar basis this cyclodextrin desolvation mechanism is more

effective in protein destabilization than the desolvation mechanism of urea and guanidinium salts

under similar conditions [39, 41]. Although cyclodextrins do not entirely avoid irreversible

processes as previously proposed [25], the mechanism looks promising and offers proteins some

of
protection against aggregation [9] and chemical and thermal denaturation [15]. The exact effect of

ro
cyclodextrin on a given protein will always be related to the particular structure of this protein.

-p
Especially to the residue sequence containing aromatic and hydrophobic residues, to the solvent
re
accessibility at the surface for possible protonation, and last but not least to the amount of strong

interactions stabilizing the protein internal structure (H-bonds in  -helices, disulfide bridges).
lP
na

Acknowledgement
ur

The conceptualization of this work, acquisition of financial support, first manuscript drafting and
Jo

supervision of A. J. Díaz-Salazar are honorary acknowledged to Miguel Costas (deceased in

December, 2024). This work was supported by DGAPA-UNAM (PAPIIT grant number

IN204616) and FQ-UNAM (PAIP 5000-9018). Daniel Ondo gratefully acknowledges financial

support for this work from Institutional funding of the University of Chemistry and Technology,

Prague.

Supplementary material
 Journal Pre-proof

Supplementary data associated with this article (full derivation of binding model, calculated SASA

of residue) can be found, in the online version, at

References

[1] H. Dodziuk (Ed.), Cyclodextrins and Their Complexes: Chemistry, Analytical Methods,

Applications, WILEY-VCH, 2006.

of
[2] M. V. Rekharsky, Y. Inoue, Complexation Thermodynamics of Cyclodextrins, Chem. Rev.

ro
98 (1998) 1875–1917. doi:10.1021/cr970015o.

[3]
-p
A. Cooper, Effect of cyclodextrins on the thermal stability of globular proteins, J. Am.

Chem. Soc. 114 (1992) 9208–9209. doi:10.1021/ja00049a074.

re
[4] Z. Hammoud, N. Khreich, L. Auezova, S. Fourmentin, A. Elaissari, H. Greige-Gerges,
lP

Cyclodextrin-membrane interaction in drug delivery and membrane structure

maintenance, Int. J. Pharm. 564 (2019) 59–76.

na

doi:10.1016/[Link].2019.03.063.
ur

[5] L. Szente, É. Fenyvesi, Cyclodextrin-lipid complexes: Cavity size matters, Struct. Chem.
Jo

28 (2017) 479–492. doi:10.1007/s11224-016-0884-9.

[6] P. D. Ross, M. V. Rekharsky, Thermodynamics of hydrogen bond and hydrophobic

interactions in cyclodextrin complexes, Biophys. J. 71 (1996) 2144–2154.

doi:10.1016/S0006-3495(96)79415-8.

[7] A. Olvera, S. Perez-Casas, M. Costas, Heat capacity contributions to the formation of

inclusion complexes, J. Phys. Chem. B 111 (2007) 11497–11505.

doi:10.1021/jp072098a.

[8] D. E. Otzen, B. R. Knudsen, F. Aachmann, K. L. Larsen, R. Wimmer, Structural basis for

 Journal Pre-proof

cyclodextrins’ suppression of human growth hormone aggregation, Protein Sci. 11 (7)

(2002) 1779–1787. doi:10.1110/ps.0202702.

[9] M. N. Shinde, R. Khurana, N. Barooah, A. C. Bhasikuttan, J. Mohanty, Sulfobutylether- 

-cyclodextrin for inhibition and rupture of amyloid fibrils, J. Phys. Chem. C 121 (36)

(2017) 20057–20065. doi:10.1021/[Link].7b07286.

[10] W. Oi, M. Isobe, A. Hashidzume, A. Harada, Macromolecular recognition: Discrimination

between human and bovine serum albumins by cyclodextrins, Macromol. Rapid Commun.

of
32 (6) (2011) 501–505. doi:10.1002/marc.201000699.

ro
[11] M. Textor, C. Vargas, S. Keller, Calorimetric quantification of linked equilibria in

-p
cyclodextrin/lipid/detergent mixtures for membrane-protein reconstitution, Methods 76
re
(2015) 183–193. doi:10.1016/[Link].2015.01.002.
lP

[12] F. L. Aachmann, D. E. Otzen, K. L. Larsen, R. Wimmer, Structural background of

cyclodextrin-protein interactions, Protein Engineering Design and Selection 16 (12) (2003)

na

905–912. doi:10.1093/protein/gzg137.

[13] T. Yamamoto, T. Kobayashi, K. Yoshikiyo, Y. Matsui, T. Takahashi, Y. Aso, A 1h nmr

ur

spectroscopic study on the tryptophan residues of lysozyme included by glucosyl- 

Jo

-cyclodextrin, J. Mol. Struct. 920 (1–3) (2009) 264–269.

doi:10.1016/[Link].2008.10.058.

[14] T. Yamamoto, N. Fukui, A. Hori, Y. Matsui, Circular dichroism and fluorescence

spectroscopy studies of the effect of cyclodextrins on the thermal stability of chicken egg

white lysozyme in aqueous solution, J. Mol. Struct. 782 (1) (2006) 60–66.

doi:10.1016/[Link].2005.01.024.

[15] S. Tavornvipas, F. Hirayama, S. Takeda, H. Arima, K. Uekama, Effects of cyclodextrins

 Journal Pre-proof

on chemically and thermally induced unfolding and aggregation of lysozyme and basic

fibroblast growth factor, J. Pharm. Sci. 95 (12) (2006) 2722–2729.

doi:10.1002/jps.20715.

[16] T. Kamiyama, M. Satoh, T. Tateishi, T. Nojiri, D. Takeuchi, T. Kimura, Effects of

modified  -cyclodextrin on thermal stability and conformation of lysozyme,

Thermochim. Acta 532 (2012) 10–14. doi:10.1016/[Link].2011.03.025.

[17] P. L. Privalov (Ed.), MICROCALORIMETRY OF MACROMOLECULES: The Physical

of
Basis of Biological Structures, John Wiley & Sons, 2012.

ro
[18] M. Bastos (Ed.), Biocalorimetry: Foundations and Contemporary Approaches, CRC Press,

2016.
-p
re
[19] B. Ibarra-Molero, A. N. Naganathan, J. M. Sanchez-Ruiz, V. Muñoz, Modern Analysis of
lP

Protein Folding by Differential Scanning Calorimetry, Vol. 567 of Methods in

Enzymology, Elsevier, 2016, Ch. 12, pp. 281–318.

na

doi:10.1016/[Link].2015.08.027.

[20] E. Ennifar (Ed.), Microcalorimetry of Biological Molecules, Vol. 1964 of Methods in

ur

Molecular Biology, Humana Press, 2019.

Jo

[21] J. M. Sanchez-Ruiz, Ligand effects on protein thermodynamic stability, Biophys. Chem.

126 (2007) 43–49. doi:10.1016/[Link].2006.05.021.

[22] C. H. Robert, A. Colosimo, S. J. Gill, Allosteric formulation of thermal transitions in

macromolecules, including effects of ligand binding and oligomerization, Biopolymers 28

(1989) 1705–1729. doi:10.1002/bip.360281006.

[23] S. Wu, Y. Ding, G. Zhang, Mechanic insight into aggregation of lysozyme by

ultrasensitive differential scanning calorimetry and sedimentation velocity, J. Phys. Chem.

 Journal Pre-proof

B 119 (52) (2015) 15789–15795. doi:10.1021/[Link].5b08190.

[24] A. Desai, C. Lee, L. Sharma, A. Sharma, Lysozyme refolding with cyclodextrins:

structure–activity relationship, Biochimie 88 (10) (2006) 1435–1445.

doi:10.1016/[Link].2006.05.008.

[25] S. Jóhannsdóttir, P. Saokham, A. Magnúsdóttir, T. Loftsson, Cyclodextrin complexes of a

globular protein and a lipophilic oligopeptide: the effect of structure and physicochemical

properties, Pharmazie 72 (2017) 575–580. doi:10.1691/ph.2017.7061.

of
[26] C. N. Pace, F. Vajdos, L. Fee, G. Grimsley, T. Gray, How to measure and predict the molar

ro
absorption coefficient of a protein, Protein Sci. 4 (11) (1995) 2411–2423.

doi:10.1002/pro.5560041120.
-p
re
[27] M. Kfoury, S. Fourmentin, State of the art in cyclodextrin solubility enhancement. are
lP

green solvents the solution?, J. Mol. Liq. 410 (2024) 125599.

doi:10.1016/[Link].2024.125599.
na

[28] S. Vega, O. Abian, A. Velazquez-Campoy, A unified framework based on the binding

polynomial for characterizing biological systems by isothermal titration calorimetry,

ur

Methods 76 (2015) 99 – 115. doi:10.1016/[Link].2014.09.010.

Jo

[29] E. J. Bekos, J. A. Gardella, F. V. Bright, The binding of free oligopeptides to cyclodextrins:

The role of the tyrosine group, J. Inclusion Phenom. Macrocyclic Chem. 26 (1996)

185–195. doi:10.1007/BF01053537.

[30] C. Kahle, U. Holzgrabe, Determination of binding constants of cyclodextrin inclusion

complexes with amino acids and dipeptides by potentiometric titration, Chirality 16 (2004)

509–515. doi:10.1002/chir.20068.

[31] Z. Sebestyén, A. Buvári-Barcza, J. Rohonczy, ph-dependent complex formation of amino

 Journal Pre-proof

acids with  -cyclodextrin and quaternary ammonium  -cyclodextrin, J. Inclusion

Phenom. Macrocyclic Chem. 73 (2012) 199–210.

doi:10.1007/s10847-011-0043-2.

[32] W. Pfeil, P. Privalov, Thermodynamic investigations of proteins: I. standard functions for

proteins with lysozyme as an example, Biophys. Chem. 4 (1) (1976) 23–32.

doi:10.1016/0301-4622(76)80003-8.

[33] M. J. Abraham, T. Murtola, R. Schulz, S. Páll, J. C. Smith, B. Hess, E. Lindahl, Gromacs:

of
High performance molecular simulations through multi-level parallelism from laptops to

ro
supercomputers, SoftwareX 1–2 (2015) 19–25.

doi:10.1016/[Link].2015.06.001.
-p
re
[34] C. Sauter, F. Otalora, J.-A. Gavira, O. Vidal, R. Giege, J. M. Garcia-Ruiz, Structure of
lP

tetragonal hen egg white lysozyme at 0.94 A from crystals grown by the counter diffusion

method, Acta Cryst. D 57 (8) (2001) 1119–1126.

na

doi:10.1107/s0907444901008873.

[35] C. R. Søndergaard, M. H. M. Olsson, M. Rostkowski, J. H. Jensen, Improved treatment of

ur

ligands and coupling effects in empirical calculation and rationalization of pka values, J.
Jo

Chem. Theory Comput. 7 (7) (2011) 2284–2295. doi:10.1021/ct200133y.

[36] M. H. M. Olsson, C. R. Søndergaard, M. Rostkowski, J. H. Jensen, Propka3: Consistent

treatment of internal and surface residues in empirical pkapredictions, J. Chem. Theory

Comput. 7 (2) (2011) 525–537. doi:10.1021/ct100578z.

[37] M. Cueto, M. J. Dorta, O. Munguía, M. Llabrés, New approach to stability assessment of

protein solution formulations by differential scanning calorimetry, Int. J. Pharm. 252

(2003) 159–166. doi:10.1016/S0378-5173(02)00627-0.

 Journal Pre-proof

[38] A. Blumlein, J. J. McManus, Reversible and non-reversible thermal denaturation of

lysozyme with varying ph at low ionic strength, Biochimica et Biophysica Acta - Proteins

and Proteomics 1834 (10) (2013) 2064–2070.

doi:10.1016/[Link].2013.06.001.

[39] A. Arsiccio, T. Sarter, I. Polidori, G. Winter, R. Pisano, J.-E. Shea, Thermodynamic

modeling and experimental data reveal that sugars stabilize proteins according to an

excluded volume mechanism, J. Am. Chem. Soc. 145 (30) (2023) 16678–16690.

of
doi:10.1021/jacs.3c04293.

ro
[40] T. Imoto, Lysozyme (Mar. 2017).

doi:10.1002/9780470015902.a0000869.pub3.
-p
re
[41] J.-Y. Chang, L. Li, The unfolding mechanism and the disulfide structures of denatured

lysozyme, FEBS Lett. 511 (1–3) (2002) 73–78.

lP

doi:10.1016/s0014-5793(01)03284-7.
na

[42] L. Samuelsen, R. Holm, C. Schönbeck, Cyclodextrin binding constants as a function of ph

for compounds with multiple pka values, Int. J. Pharm. 585 (2020) 119493.
ur

doi:10.1016/[Link].2020.119493.
Jo

[43] L. Samuelsen, R. Holm, C. Schönbeck, Specific buffers affect the stability of a charged

cyclodextrin complex via competitive binding and ionic strength, J. Pharm. Sci. 110 (7)

(2021) 2661–2668. doi:10.1016/[Link].2021.02.012.

[44] K. Matsuyama, S. El-Gizawy, J. H. Perrin, Thermodynamics of binding of aromatic amino

acids to  -,  - and  -cyclodextrins, Drug Dev. Ind. Pharm. 13 (15) (1987) 2687–2691.

doi:10.3109/03639048709022084.

[45] J. Mrozek, B. Banecki, J. Karolczak, W. Wiczk, Influence of the separation of the charged
 Journal Pre-proof

groups and aromatic ring on interaction of tyrosine and phenylalanine analogues and

derivatives with  -cyclodextrin, Biophys. Chem. 116 (3) (2005) 237–250.

doi:10.1016/[Link].2005.04.011.

[46] J. Horsky, J. Pitha, Inclusion complexes of proteins: interaction of cyclodextrins with

peptides containing aromatic amino acids studied by competitive spectrophotometry., J.

Inclusion Phenom. Macrocyclic Chem. 18 (1994) 291–300.

doi:10.1007/BF00708735.

of
[47] H. Yamamura, M. V. Rekharsky, Y. Ishihara, M. Kawai, Y. Inoue, Factors controlling the

ro
complex architecture of native and modified cyclodextrins with dipeptide (z-glu-tyr)

-p
studied by microcalorimetry and nmr spectroscopy: Critical effects of peripheral
re
bis-trimethylamination and cavity size, J. Am. Chem. Soc. 126 (2004) 14224–14233.

doi:10.1021/ja046612r.
lP

[48] F. L. Aachmann, K. L. Larsen, R. Wimmer, Interactions of cyclodextrins with aromatic

na

amino acids: a basis for protein interactions, J. Inclusion Phenom. Macrocyclic Chem. 73

(1–4) (2012) 349–357. doi:10.1007/s10847-011-0071-y.

ur
Jo
 Journal Pre-proof

A. Jessica Díaz-Salazar: Conceptualization, Data curation, Formal analysis, Writing – original

draft,
Writing – review & editing
Daniel Ondo: Conceptualization, Formal analysis, Visualization, Writing- Original draft, Review
and Editing

of
ro
-p
re
lP
na
ur
Jo
 Journal Pre-proof

Declaration of interests

☒The authors declare that they have no known competing financial interests or personal relationships
that could have appeared to influence the work reported in this paper.

☐The authors declare the following financial interests/personal relationships which may be considered
as potential competing interests:

of
ro
Graphical Abstract

[Image]
-p
re
lP
na
ur
Jo
 Journal Pre-proof

Highlights
 Lysozyme thermal stability probed by DSC in presence of native cyclodextrins (CDs)
 Stability decreases in all cases, in order β-CD > α-CD > γ-CD
 Two versions of ligand binding model applied
 Binding parameters mostly correlate with molecular structure and hydration aspects
 CDs partially inhibit irreversible aggregation, β-CD > α-CD ≈ γ-CD at pH 2

of
ro
-p
re
lP
na
ur
Jo