In luence o local o de ing in he pe mea ion o Temozolomide h ough he
b ain plasma ic memb ane
Yanhong Ge
a
, Huixia Lu
b,*
, Jo di Ma í
a,*
a
Depa men o Physics, Poly echnic Uni e si y o Ca alonia-Ba celona Tech, B4-B5 No he n Campus UPC, Ba celona 08034, Ca alonia, Spain
b
Ins i u de Ciencia de Ma e ials de Ba celona (ICMAB-CSIC), Campus de la UAB, Bella e a 08193, Ca alonia, Spain
ARTICLE INFO
Keywo ds:
Temozolomide
B ain plasma memb ane
F ee ene gy ba ie
Molecula dynamics
Adap i e biasing o ce
ABSTRACT
Temozolomide, a small-molecule d ug, is p ima ily used o ea glioblas oma, a umo ha a acks bo h he
spinal co d and b ain. Unde s anding how Temozolomide in e ac s wi h di e en lipids wi hin he b ain cell
memb ane a he a omic le el can help elucida e i s abili y o pe mea e h ough cell memb anes. In his s udy, we
cons uc ed a simpli ied b ain plasma memb ane model o explo e he mic oscopic s uc u e and dynamics o
Temozolomide using all-a om mic osecond-scale molecula dynamics simula ions. Temozolomide is ypically
ound in he sol en -aqueous luid su ounding he b ain memb ane, bu i can access he memb ane in e ace
egula ly and e en ually bind o lipids o he choline and ce eb oside classes. To in es iga e he ee ene gy
ba ie s o Temozolomide ela ed o i s c ossing o b ain-like plasma memb anes, we employed adap i e biasing
o ce me hods. These simula ions e ealed ha he ee ene gy ba ie s anged be ween 28 and 50 kcal/mol a
empe a u es be ween 310 K and 323 K. Ou indings sugges ha Temozolomide canno c oss he memb ane by
pu e di usion a no mal human body empe a u e, bu ha ising he empe a u e signi ican ly inc eases he
p obabili y o ba ie c ossing. This is p ima ily due o he c ucial ole played by choles e ol and lipids o he
ce eb oside class. These esul s can be used o op imise he molecula design o Temozolomide and de elop new
analogs wi h imp o ed pha macokine ic p ope ies.
1. In oduc ion
Temozolomide (TMZ) is he mos well-known and e ec i e d ug
commonly used o ea pa ien s wi h glioblas oma, a highly malignan
and di icul - o- ea b ain umo [1,2], as well as anaplas ic as ocy-
oma [3]. TMZ is a small-molecule imidazo e azine ha ac s as an
alkyla ing agen wi h demons a ed e icacy agains glioblas oma [4].
Howe e , se e al ba ie s —mos no ably he e iculoendo helial sys em
and he blood-b ain ba ie (BBB) [5]— limi he abili y o many d ugs
o each he b ain and e ec i ely a ge umo s. O e coming hese ob-
s acles is essen ial o success ul ea men [6–8]. Unde no mal condi-
ions, TMZ, which has a sho plasma hal -li e (∼1.8 h), is comple ely
abso bed in he gas oin es inal ac ollowing o al adminis a ion [9].
Subsequen ly, he p od ug apidly hyd olyzes a physiological pH in o
i s ac i e bu uns able me aboli e, 5-(3-me hyl iazen-1-yl)imidazole-4-
ca boxamide (MTIC, hal -li e ∼2 min) [9–11], which hen deg ades in o
5-aminoimidazole-4-ca boxamide and he me hyldiazonium ion [12]. In
he inal s ep, he me hyldiazonium ca ion p e e en ially me hyla es
DNA a guanine and adenine esidues [13,14]. I is impo an o no e
ha he ac i a ion o TMZ occu s wi hin a na ow pH ange close o
physiological pH [15].
The MTIC compound o med in he plasma is unable o c oss he BBB
and mus ins ead be gene a ed locally wi hin he b ain. Howe e , TMZ
i sel is gene ally conside ed o ha e good BBB pe meabili y [16,17].
Despi e i s e ec i eness, p olonged TMZ he apy can esul in se e e and
unp edic able myelosupp ession, which has limi ed i s u he clinical
de elopmen [18]. To enhance TMZ concen a ions in he b ain, many
esea che s ha e explo ed nanoscale d ug deli e y sys ems as a p om-
ising s a egy [11]. Fo example, Bouzinab e al. [19] de eloped a
nanodeli e y pla o m based on a biocompa ible p o ein nanocage ha
imp o ed TMZ deli e y o cance cells, demons a ing inc eased e icacy
o he imidazo e azine class o ea ing glioblas oma mul i o me and
po en ially o he malignancies. Meanwhile, o he esea ch g oups a e
ocusing on he de elopmen o no el compounds, including TMZ de-
i a i es [20–22]. Recen ly, Yin e al. [23] epo ed a new TMZ-de i ed
compound, 5-aminoimidazole-4-ca boxamide (AICA).
The e icacy o TMZ is also limi ed by i s es ic ed abili y o c oss
b ain and umo cell memb anes o each i s a ge si e [24]. Nume ous
* Co esponding au ho s.
E-mail add esses: [email p o ec ed] (H. Lu), [email p o ec ed] (J. Ma í).
Con en s lis s a ailable a ScienceDi ec
Biophysical Chemis y
jou nal homepage: www.else ie .com/loca e/biophyschem
h ps://doi.o g/10.1016/j.bpc.2025.107457
Recei ed 23 Janua y 2025; Recei ed in e ised o m 8 May 2025; Accep ed 10 May 2025
Biophysical Chemis y 324 (2025) 107457
A ailable online 26 May 2025
0301-4622/© 2025 The Au ho s. Published by Else ie B.V. This is an open access a icle unde he CC BY-NC-ND license ( h p://c ea i ecommons.o g/licenses/by-
nc-nd/4.0/ ).
s udies ha e in es iga ed how ac o s such as age and egional b ain
dis ibu ion in luence he lipid composi ion o neu ons and b ain cells
[25–28]. The human b ain is composed o specialized cells wi h mem-
b anes s uc u ed in mul iple laye s, each con aining dis inc lipid
componen s [29]. TMZ is p ima ily used in he ea men o glioblas-
oma, which mos commonly de elops in he sup a en o ial egion o he
b ain [30]. Fluo escence and spec opho ome y expe imen s ha e
shown ha TMZ has a high a ini y o b ain memb anes, which may
hinde i s pene a ion due o s ong in e ac ions wi h lipid head g oups
loca ed a he blood-b ain ba ie in e ace [9,24]. Howe e , o he bes
o ou knowledge, no s udies ha e ye examined he s uc u al in-
e ac ions be ween TMZ and biological b ain plasma memb anes. In his
wo k, we conside a ealis ically complex lipid model o he human
neu onal/b ain plasma memb ane (BPM), which accu a ely ep esen s
he majo componen s o he ac ual sys em. A schema ic ep esen a ion
o he memb anes o ming he ou e laye s o b ain cells is shown in
Fig. 1. I is impo an o no e ha , in gene al, he inne and ou e lea le s
o he BPM di e in hei lipid composi ion.
In a pionee ing s udy on compu a ional modeling o he b ain plasma
memb ane, Ing´
ol sson e al. [31] de eloped a coa se-g ained (CG) model
o he BPM, inco po a ing ens o lipids om a ious classes, based on
lipidomic da a om neu ons and b ain issue. In he p esen wo k, we
ha e d awn upon he lipid ypes iden i ied in Ing´
ol sson e al.’s s udy
[31], selec ing he mos abundan and ep esen a i e lipid om each
class. This app oach has enabled us o pe o m all-a om le el simula-
ions while a oiding he need o model an imp ac ically la ge numbe o
a oms. In his wo k, we ha e ocused ou e o s on he s udy o he ee
ene gy ba ie s in ol ed in he ansloca ion o TMZ h ough he
ex e nal laye o he BPM. In summa y, he p esen s udy is i s de o ed
o he analysis o he p e e en ial si es o TMZ a he aqueous solu ion
( ep esen ing blood plasma), and a he in e ace o he b ain cell
memb ane. Second, he main esul o his pape is he calcula ion o he
ee ene gy ba ie s ha TMZ should pass h ough he BPM as modelled
in a ealis ic way. The wo k is o ganised as ollows: me hods and
compu a ional de ails a e desc ibed wi h all de ails in Sec ion “Me hod”;
he main esul s and he co esponding discussion a e desc ibed in
Sec ion “Resul s and Discussion” and he main conclusions a e epo ed
in Sec ion “Conclusions”. Finally, some key complemen a y p ope ies
a e epo ed in he “Supplemen a y Da a” (SD).
2. Me hods
We ob ained he s uc u e o TMZ (PubChem CID: 5394) om he
Na ional Cen e o Bio echnology In o ma ion PubChem da abase (h p
://www.ncbi.nlm.nih.go /pccompound) [9]. A ealis ic model o he
b ain cell memb ane has been gene a ed wi h he well-known
CHARMM-GUI web-based ool [32,33]. Ske ches o he backbone
s uc u es o TMZ and he main componen s o he b ain memb ane a e
ep esen ed in Fig. 2. The model memb ane used in his wo k is a
simpli ied bu eliable ep esen a ion o he mammalian b ain plasma
memb ane, comp ising wo lea le s (ou e and inne laye s) wi h
di e en composi ions. A pionee ing compu a ional modeling based on
CG simula ions [31] ound signi ican di e ences be ween di e en cell
memb anes, highligh ing he composi ion o hose pa icipa ing in he
b ain memb ane. Gene ally, cell memb anes a e o med by a wide di-
e si y o lipids. In he case o he b ain, i s ex e nal memb ane is o med
by a mix u e o di e se lipids, s e ols, and o he o ganic subs ances, such
as: phospha idylcholine (PC) lipids, phospha idyle hanolamine (PE)
lipids, sphingomyelin (SM) lipids, phospha idylse ine (PS) lipids, gly-
colipids (GM), ce eb osides, phospha idylinosi ol (PI) lipids, phospha ic
acids (PA), phospha idylinosi ol phospha es (PIP), ce amides (CER),
lysophospha idylcholine (LPC) lipids, lysophospha idyle hanolamine
(LPE) lipids, and he s e ol diacylglyce ol (DAG). Besides, choles e ol
(CHOL), is cen al in all memb anes [34], cons i u ing almos hal o all
lipids in he b ain memb ane.
We conside ed he model p oposed by Ing´
ol sson e al. [31] o
analyze he in e ac ions o lipids wi h TMZ. To simpli y he s udy, we
selec ed he mos abundan lipid o ep esen one lipid ype, neglec ing
hose wi h minimal pe cen ages. Subsequen ly, we cons uc ed a 400-
lipid memb ane based on he pe cen ages indica ed in Table 1. I is
impo an o no e ha we employed all-a om molecula dynamics (MD)
simula ions, p o iding accu acy a he a omic-le el.
The NAMD2 so wa e package [35] wi h he epa ame e ized
CHARMM-36 m o ce ield [36–39] was used in all MD simula ions a
he empe a u es o 310, 316.5 and 323 K. TMZ and lipids a e ully
sol a ed by 53,556 TIP3P wa e molecules and sodium chlo ide a 0.15
Fig. 1. TMZ c ossing b ain cell memb anes. Schema ic diag am o he p ocess o TMZ c ossing he BPM. The b ain bilaye cell memb ane con ains a wide a ie y
o lipids, wi h he ypes and p opo ions o lipids in he uppe and lowe laye s showing some signi ican di e ences.
Y. Ge e al.
Biophysical Chemis y 324 (2025) 107457
2
M concen a ion, yielding a sys em o 95,235 a oms. The ini ial sys em
size is 101Å ×101Å ×100Å, and pe iodic bounda y condi ions a e all
applied in X, Y, and Z di ec ions. The XY plane is aken along he su ace
o he BPM, whe eas Z is he di ec ion no mal o he ins an aneous XY
plane. The Van de Waals in e ac ions included a smoo hing unc ion
s a ing a 10Å and i s cu o was 12Å, wi h a pai lis dis ance o 16Å.
Long- ange elec os a ic o ces we e compu ed using pa icle mesh
Ewald [40], wi h a g id space o 1Å. Pe iodic bounda y condi ions in all
spa ial di ec ions a e conside ed. A Lange in he mos a [41] was used
wi h a damping coe icien o 1 ps−1 o con ol he empe a u e, and he
p essu e was se a 1 a m and egula ed by a Nose-Hoo e Lange in
pis on [42] wi h Lange in dynamics [43] a an oscilla ion pe iod o 50
s. A e 100 ns equilib a ion pe iods a he NVT ensemble, 0.5
μ
s a-
jec o ies wi h a imes ep o 2 s we e gene a ed a he NPT ensemble o
each sys em. All bonds in ol ing hyd ogens we e se o ixed leng hs,
allowing luc ua ions o bond dis ances and angles o he emaining
a oms.
3. Resul s and discussion
3.1. Physical cha ac e is ics o BPM and TMZ a i s in e ace
To desc ibe he phase s a es o he BPM simula ed in he p esen
wo k, we ha e compu ed he so-called deu e ium o de pa ame e SCD,
as i was de ined in e e ences [44–45], which can help us o e icien ly
cha ac e ise he o de ing inside he hyd a ed lipid bilaye . SCD is de ined
o each CH2 g oup o he wo ails o each lipid class as ollows:
SCD =1
2(3<cos2θCD >−1),(1)
whe e θCD is he angle be ween he di ec ion no mal o he su ace o he
memb ane and a CH-bond. I is ema kable ha SCD can be ob ained
om 2H NMR expe imen s [46], making i a sui able p ope y o e i y
he eliabili y o he simula ions. The a e aged esul s a e shown in
Fig. 3 o bo h ail chains o se e al classes o lipids conside ed in his
wo k. The esul s indica e ha all classes o lipids in es iga ed show
maxima o SCD be ween 0.37 and 0.43 uni s, namely wi h lowe alues
ound a 323 K and highe a 310 K. These alues la ge han 0.4 a e in
o e all good ag eemen wi h he esul s o Ing´
ol sson e al. [31] o he
ull b ain cell CG model epo ed by such au ho s. Since SCD is an indi-
ca ion o he deg ee o o de ing (linea i y) o he acyl chains, we con i m
ha he acyl chains o all ypes o lipids exhibi sligh ly g ea e o de ing
a 310 K, which becomes mo e diso de ed a 323 K, as expec ed. On he
o he hand, we can obse e wo addi ional ea u es: (1) he o de ing a
Fig. 2. Backbone s uc u es. Ske ches o molecula s uc u es o TMZ and he main classes o lipids conside ed in his wo k: choles e ol (CHOL), di-palmy oil-
phospha idylcholine (DPPC), 1-palmi oyl-2-oleoyl-sn-glyce o-3-phosphoe hanolamine (POPE), 1-palmi oyl-2-oleoyl-sn-glyce o-3-phospho-L-se ine (POPS), N-(hex-
adecanoyl)-hexadecasphing-4-enine-1-phosphocholine (DPSM) and 1,2-dipalmi oyl-sn-glyce o-3-succina e (DPGS). The la e a e ce eb osides and hey a e lipids
exclusi ely loca ed a he b ain cells. Pa o hyd ogen‑ca bon bonds a e no shown, o he sake o cla i y. The highligh ed si es will be e e ed o in he ex by he
same labels.
Table 1
Composi ion o he wo lea le s (ou e , inne ) o he b ain memb ane, wi h
pe cen ages and numbe o lipids N. We should no e ha DPGS lipids a e only
loca ed a he ou e laye , whe eas POPS lipids a e only loca ed in he inne laye
o he BPM.
Class Species Ou e laye (%) Nou e Inne laye (%) Ninne
PC DPPC 24.2 48 13.6 27
PE POPE 11.0 22 21.3 43
SM DPSM 11.0 22 6.0 12
PS POPS 0.0 0 14.5 29
ce eb osides DPGS 9.4 19 0.0 0
CHOL CHOL 44.4 89 44.6 89
To al –100.0 200 100.0 200
Y. Ge e al.
Biophysical Chemis y 324 (2025) 107457
3
he inne laye ( iangles) is sligh ly g ea e han a he ou e laye
(ci cles) which indica es a mo e igid phase o he inne lea le and (2)
hose lipids ha ing unsa u a ed bonds in one o hei acyl ails, such as
POPE, DPSM, POPS and DPSG show highly diso de ed p o iles o SCD a
he double bond posi ions.
O he ele an pa ame e s o he con ol o he simula ions a e he
a ea pe lipid A and he hickness Δz o he memb ane. A is usually
ob ained conside ing he su ace o he memb ane along he XY plane
di ided by he numbe o lipids and choles e ol [47]. In he p esen case,
he wo lea le s con ain he same numbe o lipids and hei XY plane
su aces a e he same a each ins an o he simula ion, bu since he
classes o lipids in each memb ane a e di e en (see Table 1), small
di e ences could a ise. As a ough app oach, he a e aged a eas pe
lipid o wo memb anes (a lowes and highes empe a u es) we e
compu ed geome ically as desc ibed abo e and a e epo ed in Table 2.
We ob ained alues o A a ound 41 Å2 a 310 K and sligh ly highe a
323 K, indica ing a la ge he mal diso de in he la e case. The ull
p o ile o A as a unc ion o he simula ion ime is epo ed in Fig. S1 o
SD. Ou esul s a e in good o e all ag eemen wi h o he compu a ional
wo ks, such as he e e ence wo k o Ing´
ol sson e al. [31] o he s udy o
Yee e al. [48] on he BPM o a heal hy b ain compa ed o ano he
ha ing Alzheime ’s disease. The alues o A by hese au ho s we e 46
and 42.3Å2 espec i ely, i.e. wi h disc epancies be ween 3 and 11 %,
p obably due o he di e en models conside ed.
Ne e heless, in o de o ob ain signi ican ly mo e accu a e alues,
a eas pe lipid can be ob ained using he me hod o Vo onoi essella ion,
which is able o compu e he con ibu ion o each class o lipid o he
a e aged alue o A. Fo ull de ails o he calcula ions, see Re s.
[49–50]. In he p esen wo k, we ha e employed he sc ip “scipy.
spa ial.Vo onoi” [51] combined wi h he ool “Con ex Hull” [52]. The
compu ed alues o each lea le and a he wo ex eme empe a u es
a e epo ed in Table2. The e we can no ice ha he o e all, weigh ed
a e age alues a e in excellen ag eemen wi h he p e ious geome ical
Fig. 3. O de pa ame e . ∣SCD∣ o wo acyl ails (sn1 and sn2 shown in Fig. 2) o selec ed lipid species conside ed in he memb ane model o b ain cells: DPPC (A-B);
POPE (C
–
D); DPSM (E-F); POPS (G) and DPGS (H). Lipids a he inne laye a e indica ed wi h iangles and hose a he ou e laye a e ep esen ed wi h ci cles.
POPS in panel G is p esen only in he inne laye , while DPGS in panel H is only in he ou e laye , and he ci cle ep esen s sn1 while he iangle ep esen s sn2.
Sys ems a e also ep esen ed a empe a u es o 310 K ( ed) and 323 K (blue). (Fo in e p e a ion o he e e ences o colou in his igu e legend, he eade is
e e ed o he web e sion o his a icle.)
Y. Ge e al.
Biophysical Chemis y 324 (2025) 107457
4
calcula ion, bu using his me hodology, we can sepa a ely ex ac he
con ibu ion o each class o lipid. In his way, we should no ice ha : (1)
he o e all e ec o ising empe a u e is a mode a e g ow h o he
alues o A o all lipid ypes; (2) lipids o he same class can con ibu e
wi h sligh ly di e en a eas pe lipid depending o hei su oundings,
which a e di e en in he wo lea le s; (3) he e ec o choles e ol
(44–45 % o each lea le ) is o ally dominan o e he a e aged alues.
In passing, we can obse e ha he numbe s o he a eas pe lipid co -
esponding o speci ic species epo ed by G een ield e al. [50] o
phospholipids o he phospha idylglyce ol and lysyl-
phospha idylglyce ol classes a e qui e simila o hose epo ed he e, i.
e. a ound 40Å2.
The hickness o he memb ane may p o ide addi ional clues abou
he in luence o choles e ol on he mechanical p ope ies o plasma
memb anes, such as igidi y and capabili y o allowing he mo emen o
species in and ou o he cell. We ha e ob ained he hickness o he
memb ane Δz (de ined and epo ed as a unc ion o ime in Fig. S2 o
SD). The esul s epo ed he e (a ound 42 Åa 310 K) a e again quali-
a i ely close o hose epo ed in Re s. [31] (41 Å) and Re . [48] (47.3
Å) whe e mo e de ailed models we e conside ed, bu only epo ed a
310 K. The ull p o ile o Δz as a unc ion o he simula ion ime is e-
po ed in Fig. S2 o SD, showing luc ua ions o he o de o 5 % a ound
he a e aged alues, epo ed in Table2, whe e we can obse e ha as
he empe a u e inc eases om 310 K o 323 K, hickness sligh ly de-
c eases, while a ea pe lipid inc eases. These esul s indica e a sligh ly
highe deg ee o diso de as empe a u e inc eases, gi en he highe
alues o A and smalle alues o Δz a 323 K as expec ed.
3.2. Radial dis ibu ion unc ions o TMZ bound o choles e ol and
selec ed lipids
A di ec ou e o he cha ac e isa ion o he local s uc u e o each
a omic species o he sys em is usually ob ained by means o no malised
adial dis ibu ion unc ions (RDF) gAB( ) o wo di e en species A and
B (see Eq.(2) o Re . [45], o ins ance), indica ing he p obabili y o
inding a species A a a gi en dis ance o a species B. Among he wide
a ie y o possible RDF ha could be compu ed, we ha e conside ed
only he mos ele an RDF based on he i s coo dina ion shells o TMZ
wi h choles e ol and selec ed lipids. The emaining RDF indica e low
maxima a dis ances signi ican ly longe han he ypical hyd ogen bond
(HB) alues o show a he noisy p o iles which indica e ha he co -
esponding local s uc u es a e no s able enough. The selec ed RDF a e
epo ed in Fig. 4.
As a gene al obse a ion, we no e a well-de ined i s coo dina ion
shell wi h a sha p peak loca ed a ound 1.8–2.0 Å, which is indica i e o
HB be ween TMZ and he su ounding species. This ange o dis ances is
ypical o HB in e ac ions be ween small molecules and o ganic species
(see Re . [53] and e e ences he ein). Among he a ious lipids, he i s
peak in he adial dis ibu ion unc ion (RDF) o he in e ac ion
Table 2
A ea pe lipid (in Å2) and memb ane hickness Δz (in Å) a 310 K and 323 K.
Es ima ed e o s, compu ed om s anda d de ia ions, a e in pa en heses. Each
lipid class has been compu ed using Vo onoi essella ion.
P ope y 310 K 323 K
Thickness (Å) 42.6 (0.2) 42.1 (0.2)
A ea pe lipid (Å2)41.2 (0.2) 42.4 (0.3)
A ea pe lipid (in Å2, ou e lea le )
DPGS 39.1 (0.7) 41.7 (1.2)
DPPC 47.5 (0.2) 47.8 (0.1)
POPE 45.8 (0.6) 46.8 (0.3)
DPSM 40.5 (0.4) 42.0 (0.5)
CHOL 37.4 (0.2) 38.7 (0.1)
Weigh ed a e age 41.3 (0.1) 42.4 (0.1)
A ea pe lipid (in Å2, inne lea le )
POPS 43.5 (1.3) 45.0 (1.2)
DPPC 44.9 (0.1) 46.7 (0.2)
POPE 44.6 (0.2) 46.1 (0.3)
DPSM 42.7 (0.7) 42.2 (0.4)
CHOL 37.6 (0.4) 38.7 (0.3)
Weig hed a e age 41.2 (0.1) 42.5 (0.1)
Fig. 4. Radial dis ibu ion unc ions. RDF o di e en a oms in TMZ wi h se e al lipids a 310 K (black) and 323 K ( ed) sys ems. In panel A, gN5 mz−Hchol( ) e e s o
he RDF o he a om named N5 (see Fig. 2) in TMZ and he hyd ogen o he hyd oxyl g oup o choles e ol. In panels B
–
F he selec ed a oms a e also indica ed wi h
he same labels as in Fig. 2. (Fo in e p e a ion o he e e ences o colou in his igu e legend, he eade is e e ed o he web e sion o his a icle.)
Y. Ge e al.
Biophysical Chemis y 324 (2025) 107457
5
be ween N5 o TMZ and he POPE/POPS lipids is no ably highe han o
he o he s, especially o he hyd ogens a he headg oups o POPE/
POPS (see Fig. 4, panels C and D). This sugges s s onge binding o TMZ
o phospha idyle hanolamine (POPE) and phospha idylse ine (POPS)
lipids. Con e sely, panels B and E in Fig. 4 show ha DPGS, a speci ic
ce eb oside cha ac e is ic o he ou e lea le o he b ain cell mem-
b ane, also con ibu es o TMZ binding h ough hyd ogens H3 and H4.
No ably, a ela i ely s onge in e ac ion be ween TMZ and DPGS is
obse ed a he highe empe a u e (323K). The noisy p o iles in panels
B and E indica e ha he TMZ-DPGS in e ac ion has a sho e li e ime
compa ed o he TMZ-POPE/POPS binding, as long-li ed in e ac ions
(such as HB) esul in smoo he RDF p o iles, whe eas sho -li ed in-
e ac ions unde go equen b eaking and e o ming e en s.
In he pa icula case o TMZ-DPGS hyd ogen-bonding we can
obse e mainly HB o med by ni ogen N5 o TMZ wi h H3 and H4 om
DPGS and an oxygen‑hyd ogen in e ac ion o H1 o TMZ wi h O2 in
POPS, bu only well de ined a 310 K. Compa ing sys ems a 310 K and
323 K, i can be no iced ha empe a u e plays a mino ole in mos
cases. In a ew cases, such as TMZ-DGPS bindings (see panels B and E o
Fig. 4) we can no ice an enhancemen o he i s maxima a 323 K,
which can be a ibu ed o mo e equen and du able e en s o TMZ-
DPGS pai ings han a 310 K, when TMZ shows a bigge a ini y o
bind choles e ol (panel A in Fig. 4) and POPS lipids (panel F o Fig. 4).
3.3. Es ima ion o ee ene gy ba ie s o TMZ c ossing
F om a gene al pe spec i e, he calcula ion o he Helmhol z o Gibbs
ee ene gy di e ences o binding p ocesses o con igu a ional changes
is a di icul ask and i equi es a conside able amoun o compu e ime
and a p ecise knowledge o he hype su ace o he po en ial ene gy o
he sys em [54]. This can be explo ed by means o me hods such as
me adynamics [55–57], hyb id quan um mechanics/molecula me-
chanics me hods [58] o ansi ion pa h sampling [59–62]. Howe e , a
usual way o ob ain ee ene gy es ima ions is h ough he calcula ion o
po en ials o mean o ce (PMF) h ough a a ie y o me hods such as
umb ella sampling [63], cons ained molecula dynamics [64] o
adap i e biasing o ce me hods, among many o he s.
In his Sec ion we compu ed ee ene gy p o iles by means o he
adap i e biasing o ce (ABF) me hod [65,66], whe e he PMF is de i ed
om in eg a ing he mean o ce Fz o e he eac ion coo dina es ξ0
[67–69]. The o al wo k h ough he memb ane is:
ΔW= − [∫ξ inal
ξo iginal
〈Fz(ξ0) 〉dξ0],(2)
whe e ξo iginal is he o iginal eac ion coo dina e o he TMZ and ξ inal is
he inal eac ion coo dina e o he TMZ. 〈Fz(ξ0) 〉 is he a e age o ce in
he z-di ec ion exe ed on he TMZ a ξ0. The ABF me hod o e s he
ad an age o no equi ing he es ima ion o ee ene gy ba ie s and
p o iding mo e di e se sampling. In ypical simula ions, he subs a e is
d i en in a speci ic di ec ion by he o al o ce ac ing on i . Wi h ABF, a
biasing o ce in he opposi e di ec ion is in oduced, c ea ing a balance
o he subs a e and leading o andom mo ing. Ne e heless, ABF
me hods may exhibi inaccu acies a ising om insu icien sampling and
sensi i i y o he choice o eac ion coo dina e, which can hinde
con e gence in complex sys ems [68,70,71].
The eac ion coo dina es in he p esen case ep esen he dis ance
along he z-axis om he cen e o mass o TMZ o a “dummy a om”
loca ed a he poin (0.0, 0.0, 0.0). The eac ion coo dina es a e de e -
mined by in oducing a dummy a om ins ead o eal molecules o a oms
o p e en de ia ion caused by hei mo emen . These collec i e a i-
ables in small size imp o e he pe o mance o ABF calcula ion. The
eac ion coo dina es a e di ided in o se e al windows. TMZ will be
con ined wi hin speci ic windows by ha monic o ces in he Z-axis di-
ec ion a he bounda ies, which enhances he sampling e iciency.
Th ee TMZ memb ane-pene a ing ABF sys ems we e se up a 310 K,
323 K and an addi ional in e media e empe a u e a 316.5 K, o gain a
comp ehensi e unde s anding o he impac o sys em empe a u e.
Each window has a wid h o 5Å, and he PMF is calcula ed e e y 0.1Å.
The o ce cons an a he window bounda ies is 20 kcal/mol/Å2. A
sample coun o he imes ha TMZ has been placed a di e en posi-
ions in he sys em in one ABF calcula ion o each empe a u e is e-
po ed in Fig. S3 o SD, in o de o show he eliabili y o he ABF
calcula ions and ha su icen sampling was achie ed in all cases.
The a e aged esul s o he wo ee ene gy ba ie s a e shown in
Fig. 5 and he ou componen indi idual ABF ajec o ies used o ob ain
he a e ages and e o ba s a e epo ed in Fig.S4 o SD. As gene al
ea u es, we can obse e ha :
1. All PMF p o iles show a ee ene gy maximum when TMZ eaches he
cen e o he memb ane, ega dless o he di ec ion aken ei he
s a ing om he ex e nal aqueous solu ion o om he in e nal
cy oplasm;
2. Fluc ua ions o all ee ene gy p o iles a e qui e la ge a all empe -
a u es, especially when TMZ is close o he cen e o he memb ane,
indica ing impo an con igu a ional changes, mainly ela ed o he
di e en in e ac ions o TMZ wi h su ounding lipids;
3. A he lowes empe a u e (310 K, i.e. no mal human body empe -
a u e) he ee ene gy ba ie equi ed by TMZ o each he cen e o
he BPM is he la ges , o he o de o 50 ±5 kcal/mol;
4. The wo se s o ee ene gy p o iles (i.e. when TMZ is c ossing ou e
and inne laye s) a e e y simila each o he wi h an almos pe ec
ma ch be ween he wo di ec ions, bu showing a small bias o la ge
alues (o a ound 1–3 kcal/mol) when TMZ is mo ed om he in-
e nal solu ion and o ced o c oss he inne laye o he BPM;
5. We obse e a clea mono onic beha iou , wi h la ge ee ene gy
ba ie s a he lowes empe a u e and a endency o dec ease when
empe a u e is aised. In addi ion, we ha e ound a nea educ ion o
a ac o ∼1.8–1.4 ( om 49 o 28 kcal/mol o c oss he ou e laye
and om 50 o 35 o c oss he inne laye ) when he empe a u e o
he sys em is aised om 310 o 323 K. This would indica e a
signi ican ly be e abili y o TMZ o c oss he BPM when empe a-
u e is highe han he no mal one;
6. As he empe a u e ises, he di e ence in he ee ene gy ba ie ha
TMZ needs o su pass o en e /lea e he BPM inc eases: o ins ance,
a 323 K, he ba ie g ows om 28 kcal/mol o c oss he ou e laye
o 35 kcal/mol o c oss he inne laye , making i p og essi ely mo e
di icul o TMZ o be expelled om he BPM once i has en e ed.
We can a ionalize hese indings as ollows. Fi s , ega ding he
la ge e o ba s obse ed in he ABF simula ions, we explo ed he
p e e en ial con o ma ions o TMZ in he case o he la ges e o ba s
(310K), whe e TMZ mo es om he ex e nal solu ion o he cen e o he
BPM h ough he ou e laye , as seen in he blue line o Fig. 5A. The main
in e ac ions o TMZ o he wo ex eme ABF p o iles a e epo ed in
Fig. S5 o he Suppo ing In o ma ion (SD). In he highes ee ene gy
p o ile (blue line), TMZ o ms h ee s ong hyd ogen bonds (HB) wi h
wo choles e ol molecules, whe eas in he lowes ee ene gy p o ile ( ed
line), TMZ o ms wo HBs wi h wo DPPC molecules. This sugges s ha
he s onges bindings o TMZ can a y subs an ially ac oss di e en
ABF ajec o ies, esul ing in quan i a i ely di e en ee ene gy ba -
ie s. On he o he hand, he e is also he possibili y o TMZ unde going
ec ossings, i.e., mo ing backwa d a he han c ossing he ull bilaye ,
e u ning o he solu ion whe e he ajec o y s a ed. The small di e -
ences in he ee ene gy ba ie s epo ed in Fig. 5 sugges ha bo h
o wa d and backwa d mo ions s a ing nea he cen e o he BPM a e
equally p obable. To in es iga e his, we an ou independen unbiased
MD simula ions whe e TMZ was ini ially placed nea he cen e o he
BPM (z∼0 Å), using he coo dina es om he ABF ajec o ies. The
esul s o TMZ’s posi ion along he z-axis a e shown in Fig. S6 o he SD.
Al hough he s a is ics o hese auxilia y MD ajec o ies a e limi ed,
he goal was no o compu e p ecise c ossing a es (which would equi e
Y. Ge e al.
Biophysical Chemis y 324 (2025) 107457
6
unning hund eds o independen simula ions) bu o assess he possi-
bili y o ec ossings. In wo o he ou ajec o ies, TMZ c ossed he
ou e laye owa ds he “ex e nal” solu ion wi hin 20 ns. In he
emaining wo ajec o ies, TMZ c ossed he inne laye , aking be ween
40 and 50 ns. This sugges s a s onge in e ac ion be ween TMZ and he
headg oups o lipids in he inne lea le han wi h hose in he ou e
lea le . These indings a e consis en wi h he esul s shown in Fig. S4 o
he SD, whe e i is sligh ly mo e ene ge ically cos ly o TMZ o c oss he
inne lea le in e ace han he ou e lea le .
4. Summa y and conclusions
In he p esen s udy, we conduc ed a comp ehensi e compu a ional
analysis o TMZ pe mea ion ac oss a b ain-like plasma memb ane
bilaye . We in oduced a TMZ molecule a he in e ace o a model BPM
and sys ema ically obse ed i s in e ac ions wi h sol en wa e , ions,
and he dis inc egions o he memb ane, which include bo h he ou e
and inne laye s. Ou esul s con i m he p ope equilib a ion o he
sys em, and s uc u al da a indica e ha he cha ac e is ic o de
pa ame e used o e i y he sys em’s liquid-o de ed s a e, as well as
o he p ope ies such as he a ea pe lipid and memb ane hickness,
show good quali a i e ag eemen wi h esul s epo ed by Ing´
ol sson
e al. [31] and Yee e al. [48], whe e sligh ly di e en BPM models we e
employed. In addi ion, we ha e obse ed ha he model BPM memb ane
designed in he p esen wo k has a nea ly plana geome y o bo h
lea le s, wi hou any signi ican ly la ge cu a u es o s ess on he
memb ane (see Fig. S7 o SD) obse ed du ing he whole ime ange (0,5
μ
s) o he simula ions.
Ou ee ene gy p o iles o TMZ ansmemb ane c ossing ob ained
om adap i e biasing o ce me hods indica e ha TMZ has he highes
a ini y o s aying a he in e ace be ween he BPM and he aqueous
solu ion, a he han spon aneously c ossing he memb ane ia simple
di usion. This sugges s ha in he eal sys em, some agen s such as
memb ane ABC anspo e s [72] o SLC anspo e s [73,74] may play a
key ole in he c ossing o TMZ h ough he ex e nal laye o he b ain
cell memb ane. In e es ingly, when TMZ has eached he in e nal hy-
d ophobic egions o he BPM, small luc ua ions in i s posi ion can lead
o: (1) a success ul c ossing o he in e nal egions and e en ually each
he i s cy oplasmic egion o he b ain o (2) o ec ossings owa ds he
ex e nal in e ace o he BPM and e en ual e u n o he ex e nal
aqueous solu ion. Ou ABF calcula ions indica e ha TMZ needs abou
50 kcal/mol o c oss he BPM a he no mal human body empe a u e o
310 K, whe eas his ba ie dec eases o abou 28 kcal/mol a 323 K.
Da a ob ained by RDF indica e ha he o ma ion o s ong TMZ-
choles e ol, TMZ-POPS, TMZ-POPE and TMZ-DPGS pai ings a e he
main esponsible o TMZ hyd ogen bonding in e ac ions ha d i e he
pe mea ion o TMZ h ough he BPM ba ie .
Finally, we would like o ema k ha op imizing he molecula
design o TMZ h ough ad anced d ug deli e y sys ems can enhance i s
pha macokine ic p ope ies. This is no mally done by imp o ing s a-
bili y, p olonging ci cula ion ime, and inc easing b ain bioa ailabili y.
Fo ins ance, TMZ-loaded PEGyla ed liposomes ha e demons a ed a
4.2- old inc ease in b ain d ug concen a ion and a 1.6- old inc ease in
blood AUC compa ed o he ee d ug [75]. Simila ly, TMZ encapsula ed
in niosomal o mula ions has shown enhanced b ain a ge ing and
educed sys emic exposu e, leading o imp o ed he apeu ic e icacy
[76]. As an al e na i e, om ou esul s we obse e ha a educ ion o
he size o he ee ene gy ba ie ha TMZ needs o c oss o access he
b ain could be ca ied ou by educing i s abili y o o m HB wi h lipids.
This e ec could be achie ed by eplacing o masking HB dono o
accep o g oups ( o ins ance by con e ing hyd oxyl o amide g oups
in o less pola unc ionali ies o by in oducing s e ic hind ance a ound
HB-p one si es. In summa y, hese changes can dec ease nonspeci ic
in e ac ions wi h memb ane lipids, po en ially imp o ing memb ane
pe meabili y and pha macokine ics.
CRediT au ho ship con ibu ion s a emen
Yanhong Ge: W i ing – e iew & edi ing, W i ing – o iginal d a ,
Visualiza ion, Valida ion, So wa e, Me hodology, In es iga ion, Fo mal
analysis, Da a cu a ion. Huixia Lu: W i ing – e iew & edi ing, W i ing
– o iginal d a , Visualiza ion, Valida ion, Supe ision, So wa e, Re-
sou ces, Me hodology, In es iga ion, Fo mal analysis, Concep ualiza-
ion. Jo di Ma í: W i ing – e iew & edi ing, W i ing – o iginal d a ,
Visualiza ion, Valida ion, Supe ision, So wa e, Resou ces, Me hodol-
ogy, In es iga ion, Fo mal analysis, Concep ualiza ion.
Fig. 5. F ee ene gy p o iles o TMZ c ossing he b ain cell memb ane bilaye . (A): TMZ om solu ion owa ds he cen e o he BPM, c ossing ou e laye , (B):
TMZ om he solu ion owa ds he cen e o he BPM, c ossing he inne laye . The o igin o he ee ene gy is aken o TMZ being sol a ed in he aqueous solu ion,
o each o he h ee empe a u es. Each PMF cu e is he a e age esul ob ained om ou independen ABF simula ions, so ha he shadowed a eas indica e e o
ba s and hey a e ob ained as he la ges de ia ions be ween he a e age and he componen p o iles.
Y. Ge e al.
Biophysical Chemis y 324 (2025) 107457
7
Decla a ion o compe ing in e es
The au ho s decla e ha hey ha e no known compe ing inancial
in e es s o pe sonal ela ionships ha could ha e appea ed o in luence
he wo k epo ed in his pape .
Acknowledgemen s
The au ho s acknowledge inancial suppo p o ided by he Spanish
Minis y o Science, Inno a ion and Uni e si ies. Huixia Lu hanks he
inancial suppo by he “Ma ga i a Salas” g an which is unded by he
Eu opean Union–Nex Gene a ionEU. This publica ion is a pa o he I +
D +i p ojec wi h e e ence PID2021-124297NB-C32, ounded by
MCIN/AEI/10.13-039/501100011033 and “FEDER Una mane a de
hace Eu opa”. Yanhong Ge is a Ph.D. ellow om he China Schola ship
Council (g an 202306230043). J.M. hanks he Gene ali a de Ca alunya
o he suppo h ough he g an G up de Rece ca SGR-Ca 2021
Condensed, Complex and Quan um Ma e G oup e e ence 2021SGR-
01411 and o he Poly echnic Uni e si y o Ca alonia-Ba celona Tech
h ough he unding AGRUPS. The au ho s hank ully acknowledge
compu e esou ces om Ma eNos um5 supe compu e as well as
echnical suppo p o ided by BSC (RES-BCV-2024-2-0006).
Appendix A. Supplemen a y da a
Supplemen a y da a o his a icle can be ound online a h ps://doi.
o g/10.1016/j.bpc.2025.107457.
Da a a ailabili y
All ela ed iles including an example used o calcula e he binding
ee ene gy o his wo k can be ound in ou eposi o y o Zenodo:
h ps://doi.o g/10.5281/zenodo.12745251.
Re e ences
[1] H.S. F iedman, T. Ke by, H. Cal e , Temozolomide and ea men o malignan
glioma, Clin. Cance Res. 6 (7) (2000) 2585–2597.
[2] A.C. Tan, D.M. Ashley, G.Y. L´
opez, M. Malinzak, H.S. F iedman, M. Khas aw,
Managemen o glioblas oma: s a e o he a and u u e di ec ions, CA Cance J.
Clin. 70 (4) (2020) 299–312.
[3] W.A. Yung, M.D. P ados, R. Yaya-Tu , S.S. Rosen eld, M. B ada, H.S. F iedman,
R. Alb igh , J. Olson, S.M. Chang, A.M. O’Neill, e al., Mul icen e phase ii ial o
emozolomide in pa ien s wi h anaplas ic as ocy oma o anaplas ic
oligoas ocy oma a i s elapse, J. Clin. Oncol. 17 (9) (1999) 2762–2771.
[4] A. Thomas, M. Tanaka, J. T epel, W.C. Reinhold, V.N. Rajapakse, Y. Pommie ,
Temozolomide in he e a o p ecision medicine, Cance Res. 77 (4) (2017)
823–826.
[5] D. Wu, Q. Chen, X. Chen, F. Han, Z. Chen, Y. Wang, The blood–b ain ba ie :
s uc u e, egula ion, and d ug deli e y, Signal T ansduc . Ta ge . The . 8 (1)
(2023) 217.
[6] W. Risau, H. Wolbu g, De elopmen o he blood-b ain ba ie , T ends Neu osci. 13
(5) (1990) 174–178.
[7] P. Ballabh, A. B aun, M. Nede gaa d, The blood–b ain ba ie : an o e iew:
s uc u e, egula ion, and clinical implica ions, Neu obiol. Dis. 16 (1) (2004) 1–13.
[8] O. Van Tellingen, B. Ye kin-A ik, M. De Gooije , P. Wesseling, T. Wu dinge , H. De
V ies, O e coming he blood–b ain umo ba ie o e ec i e glioblas oma
ea men , D ug Resis . Upda . 19 (2015) 1–12.
[9] M. Rubio-Camacho, J.A. Encina , M.J. Ma ínez-Tom´
e, R. Esquemb e, C.R. Ma eo,
The in e ac ion o emozolomide wi h blood componen s sugges s he po en ial use
o human se um albumin as a biomime ic ca ie o he d ug, Biomolecules 10 (7)
(2020) 1015.
[10] H. S obel, T. Baisch, R. Fi zel, K. Schilbe g, M.D. Siegelin, G. Ka pel-Massle , K.-
M. Deba in, M.-A. Wes ho , Temozolomide and o he alkyla ing agen s in
glioblas oma he apy, Biomedicines 7 (3) (2019) 69.
[11] D. Pe enko, V. Chuba e , N. Syz an se , N. Ismail, V. Me kulo , S. Sologo a,
E. G igo e skikh, E. Smolya chuk, R. Alyau din, Temozolomide e icacy and
me abolism: he implici ele ance o nanoscale deli e y sys ems, Molecules 27
(11) (2022) 3507.
[12] I.C. Lopes, S.C.B. de Oli ei a, A.M. Oli ei a-B e , Temozolomide chemical
deg ada ion o 5-aminoimidazole-4-ca boxamide–elec ochemical s udy,
J. Elec oanal. Chem. 704 (2013) 183–189.
[13] M.J. Tisdale, An i umou imidazo e azines—x : ole o guanine o6 alkyla ion in
he mechanism o cy o oxici y o imidazo e azinones, Biochem. Pha macol. 36 (4)
(1987) 457–462.
[14] B.J. Denny, R.T. Wheelhouse, M.F. S e ens, L.L. Tsang, J.A. Slack, Nm and
molecula modeling in es iga ion o he mechanism o ac i a ion o he an i umo
d ug emozolomide and i s in e ac ion wi h dna, Biochemis y 33 (31) (1994)
9045–9051.
[15] J. Zhang, M.F.G. S e ens, T.D. B adshaw, Temozolomide: mechanisms o ac ion,
epai and esis ance, Cu . Mol. Pha macol. 5 (1) (2012) 102–114.
[16] K. Kelly, M. Sacapano, K. P osolo ich, J. Ong, K. Black, Blood b ain ba ie
pe meabili y o emozolomide, Cance Res. 65 (9 Supplemen ) (2005) 330.
[17] M.C. de Gooije , N.A. de V ies, T. Buckle, L.C. Buil, J.H. Beijnen, W. Booge d,
O. an Tellingen, Imp o ed b ain pene a ion and an i umo e icacy o
emozolomide by inhibi ion o abcb1 and abcg2, Neoplasia 20 (7) (2018) 710–720.
[18] G. Blackledge, J. Robe s, S. Kaye, R. Taylo , J. Williams, B. De S a ola,
B. Uscinska, A phase ii s udy o mi ozolomide in me as a ic ansi ional cell
ca cinoma o he bladde , Eu . J. Cance Clin. Oncol. 25 (2) (1989) 391–392.
[19] K. Bouzinab, H.S. Summe s, M.F. S e ens, C.J. Moody, N.R. Thomas,
P. Ge shko ich, N. Wes on, M.B. Ash o d, T.D. B adshaw, L. Tu yanska, Deli e y o
emozolomide and n3-p opa gyl analog o b ain umo s using an apo e i in
nanocage, ACS Appl. Ma e . In e aces 12 (11) (2020) 12609–12617.
[20] M. Chak a a y, P. Ganguli, M. Mu aha i, R.R. Sa ka , G.J. Pe e s, Y. Mayu , S udy
o combina o ial d ug syne gy o no el ac idone de i a i es wi h emozolomide
using in-silico and in- i o me hods in he ea men o d ug- esis an glioma,
F on . Oncol. 11 (2021) 625899.
[21] S.D. P ecilla, S.S. Kudu alli, E.A. P a eena, S. Thanga el, T. Ani ha, In eg a ion o
syn he ic and na u al de i a i es e i es he he apeu ic po en ial o emozolomide
agains glioma-an in i o and in i o pe spec i e, Li e Sci. 301 (2022) 120609.
[22] R. Li, D. Tang, J. Zhang, J. Wu, L. Wang, J. Dong, The emozolomide de i a i e 2 -
p400 inhibi s glioma g ow h ia adminis a ion ou e o in a enous injec ion,
J. Neu o-Oncol. 116 (2014) 25–30.
[23] J. Yin, X. Wang, X. Ge, F. Ding, Z. Shi, Z. Ge, G. Huang, N. Zhao, D. Chen, J. Zhang,
e al., Hypoxan hine phospho ibosyl ans e ase 1 me abolizes emozolomide o
ac i a e ampk o d i ing chemo esis ance o glioblas omas, Na . Commun. 14 (1)
(2023) 5913.
[24] M.J. Ramalho, S. And ade, M.´
A.N. Coelho, J.A. Lou ei o, M.C. Pe ei a, Biophysical
in e ac ion o emozolomide and i s ac i e me aboli e wi h biomemb ane models:
he ele ance o d ug-memb ane in e ac ion o glioblas oma mul i o me he apy,
Eu . J. Pha m. Biopha m. 136 (2019) 156–163.
[25] I. K acun, H. Rosne , V. D no sek, M. He e -Lauc, C. Coso i´
c, G. Lauc, Human
b ain gangliosides in de elopmen , aging and disease, In . J. De . Biol. 35 (3)
(1991) 289–295.
[26] L. S enne holm, K. Bos ¨
om, B. Jungbje , L. Olsson, Memb ane lipids o adul
human b ain: lipid composi ion o on al and empo al lobe in subjec s o age 20
o 100 yea s, J. Neu ochem. 63 (5) (1994) 1802–1811.
[27] M. Au eli, S. G assi, S. P ioni, S. Sonnino, A. P ine i, Lipid memb ane domains in
he b ain, Biochim. e Biophys. Ac a (BBA)-Mol. Cell Biolo. Lipids 1851 (8) (2015)
1006–1016.
[28] Y. A ai, J.L. Sampaio, M. Wilsch-B ¨
auninge , A.W. E inge , C. Ha ne , W.
B. Hu ne , Lipidome o midbody eleased om neu al s em and p ogeni o cells
du ing mammalian co ical neu ogenesis, F on . Cell. Neu osci. 9 (2015) 325.
[29] P. Sj¨
o all, J. Lausmaa, B. Johansson, Mass spec ome ic imaging o lipids in b ain
issue, Anal. Chem. 76 (15) (2004) 4271–4278.
[30] I. Chak aba i, M. Cockbu n, W. Cozen, Y.-P. Wang, S. P es on-Ma in,
A popula ion-based desc ip ion o glioblas oma mul i o me in los Angeles coun y,
1974–1999, Cance , In e disciplina y In e na ional Jou nal o he Ame ican
Cance Socie y 104 (12) (2005) 2798–2806.
[31] H.I. Ing´
ol sson, T.S. Ca pen e , H. Bha ia, P.-T. B eme , S.J. Ma ink, F.
C. Ligh s one, Compu a ional lipidomics o he neu onal plasma memb ane,
Biophys. J. 113 (10) (2017) 2271–2280.
[32] S. Jo, T. Kim, V.G. Iye , W. Im, Cha mm-gui: a web-based g aphical use in e ace
o cha mm, J. Compu . Chem. 29 (11) (2008) 1859–1865.
[33] S. Jo, J.B. Lim, J.B. Klauda, W. Im, Cha mm-gui memb ane builde o mixed
bilaye s and i s applica ion o yeas memb anes, Biophys. J. 97 (1) (2009) 50–58.
[34] J. Yang, J. Ma í, C. Cale o, Pai in e ac ions among e na y dppc/popc/choles e ol
mix u es in liquid-o de ed and liquid-diso de ed phases, So Ma e 12 (20)
(2016) 4557–4561.
[35] J.C. Phillips, R. B aun, W. Wang, J. Gumba , E. Tajkho shid, E. Villa, C. Chipo , R.
D. Skeel, L. Kale, K. Schul en, Scalable molecula dynamics wi h namd, J. Compu .
Chem. 26 (16) (2005) 1781–1802.
[36] J.B. Klauda, R.M. Venable, J.A. F ei es, J.W. O’Conno , D.J. Tobias, C. Mond agon-
Rami ez, I. Vo obyo , A.D. MacKe ell J ., R.W. Pas o , Upda e o he cha mm all-
a om addi i e o ce ield o lipids: alida ion on six lipid ypes, J. Phys. Chem. B
114 (23) (2010) 7830–7843.
[37] J.B. Lim, B. Rogaski, J.B. Klauda, Upda e o he choles e ol o ce ield pa ame e s
in cha mm, J. Phys. Chem. B 116 (1) (2012) 203–210.
[38] J.A. Lemkul, J. Huang, B. Roux, A.D. MacKe ell J ., An empi ical pola izable o ce
ield based on he classical d ude oscilla o model: de elopmen his o y and ecen
applica ions, Chem. Re . 116 (9) (2016) 4983–5013.
[39] J. Huang, S. Rausche , G. Naw ocki, T. Ran, M. Feig, B.L. De G oo , H. G ubmülle ,
A.D. MacKe ell J ., Cha mm36m: an imp o ed o ce ield o olded and
in insically diso de ed p o eins, Na . Me hods 14 (1) (2017) 71–73.
[40] U. Essmann, L. Pe e a, M.L. Be kowi z, T. Da den, H. Lee, L.G. Pede sen, A smoo h
pa icle mesh ewald me hod, J. Chem. Phys. 103 (19) (1995) 8577–8593.
[41] H.J. Be endsen, J.V. Pos ma, W.F. Van Guns e en, A. Di Nola, J.R. Haak, Molecula
dynamics wi h coupling o an ex e nal ba h, J. Chem. Phys. 81 (8) (1984)
3684–3690.
[42] G.J. Ma yna, D.J. Tobias, M.L. Klein, Cons an p essu e molecula dynamics
algo i hms, J. Chem. Phys. 101 (5) (1994) 4177–4189.
Y. Ge e al.
Biophysical Chemis y 324 (2025) 107457
8
[43] S.E. Felle , Y. Zhang, R.W. Pas o , B.R. B ooks, Cons an p essu e molecula
dynamics simula ion: he lange in pis on me hod, J. Chem. Phys. 103 (11) (1995)
4613–4621.
[44] G.W. S ock on, I.C. Smi h, A deu e ium nuclea magne ic esonance s udy o he
condensing e ec o choles e ol on egg phospha idylcholine bilaye memb anes. i.
Pe deu e a ed a y acid p obes, Chem. Phys. Lipids 17 (2–3) (1976) 251–263.
[45] H. Lu, J. Ma í, E ec s o choles e ol on he binding o he p ecu so
neu o ansmi e yp ophan o zwi e ionic memb anes, J. Chem. Phys. 149 (16)
(2025).
[46] P.L. Yeagle, The Memb anes o Cells, Academic P ess, 2016.
[47] P.R. Pandey, S. Roy, Headg oup media ed wa e inse ion in o he dppc bilaye : a
molecula dynamics s udy, J. Phys. Chem. B 115 (12) (2011) 3155–3163.
[48] S.M. Yee, R.J. Gillams, S.E. McLain, C.D. Lo enz, E ec s o lipid he e ogenei y on
model human b ain lipid memb anes, So Ma e 17 (1) (2021) 126–135.
[49] W. Shinoda, S. Okazaki, A o onoi analysis o lipid a ea luc ua ion in a bilaye ,
J. Chem. Phys. 109 (4) (1998) 1517–1521.
[50] M.L. G een ield, L.M. Ma in, F. Joodaki, Compu ing indi idual a ea pe head
g oup e eals lipid bilaye dynamics, J. Phys. Chem. B 126 (50) (2022)
10697–10711.
[51] H. Yin, G. Cusa is, Ringspy: a py hon package o o onoi mesh gene a ion o
cellula solids wi h adial g ow h pa e n, J. Open Sou ce So w. 8 (83) (2023)
4945.
[52] E.W. Weiss ein, Con ex hull. h ps://ma hwo ld.wol am.com/, 2025.
[53] J. Ma í, H. Lu, Mic oscopic in e ac ions o mela onin, se o onin and yp ophan
wi h zwi e ionic phospholipid memb anes, In . J. Mol. Sci. 22 (6) (2021) 2842.
[54] F. M. Y ebe g, R. H. Swendsen, D. M. Zucke man, Compa ison o ee ene gy
me hods o molecula sys ems, J. Chem. Phys. 125 (18).
[55] A. Laio, M. Pa inello, Escaping ee-ene gy minima, P oc. Na l. Acad. Sci. 99 (20)
(2002) 12562–12566.
[56] J. Yang, C. Cale o, M. Bonomi, J. Ma í, Speci ic ion binding a phospholipid
memb ane su aces, J. Chem. Theo y Compu . 11 (9) (2015) 4495–4499.
[57] H. Lu, J. Ma i, Cellula abso p ion o small molecules: ee ene gy landscapes o
mela onin binding a phospholipid memb anes, Sci. Rep. 10 (1) (2020) 9235.
[58] H.M. Senn, W. Thiel, Qm/mm me hods o biological sys ems, A omis . App oach.
Mode. Biol.: Quan . Chem. o Mol. Simula . (2007) 173–290.
[59] J. Ma í, F.S. Csajka, D. Chandle , S ochas ic ansi ion pa hways in he aqueous
sodium chlo ide dissocia ion p ocess, Chem. Phys. Le . 328 (1–2) (2000) 169–176.
[60] P.G. Bolhuis, D. Chandle , C. Dellago, P.L. Geissle , T ansi ion pa h sampling:
h owing opes o e ough moun ain passes, in he da k, Annu. Re . Phys. Chem.
53 (1) (2002) 291–318.
[61] J. Ma í, F.S. Csajka, T ansi ion pa h sampling s udy o lip- lop ansi ions in
model lipid bilaye memb anes, Phys. Re . E 69 (6) (2004) 061918.
[62] C. Dellago, P.G. Bolhuis, T ansi ion pa h sampling and o he ad anced simula ion
echniques o a e e en s, in: Ad anced compu e simula ion app oaches o so
ma e sciences III, 2009, pp. 167–233.
[63] J. K¨
as ne , Umb ella sampling, Wiley in e disciplina y e iews: compu a ional
molecula , Science 1 (6) (2011) 932–942.
[64] M. Sp ik, G. Cicco i, F ee ene gy om cons ained molecula dynamics, J. Chem.
Phys. 109 (18) (1998) 7737–7744.
[65] J. H´
enin, J. Gumba , C. Chipo , F ee Ene gy Calcula ions Along a Reac ion
Coo dina e: A Tu o ial o Adap i e Biasing Fo ce Simula ions, 2025.
[66] J. H´
enin, C. Chipo , O e coming ee ene gy ba ie s using uncons ained
molecula dynamics simula ions, J. Chem. Phys. 121 (7) (2004) 2904–2914.
[67] E. Da e, D. Rod íguez-G´
omez, A. Poho ille, Adap i e biasing o ce me hod o
scala and ec o ee ene gy calcula ions, J. Chem. Phys. 128 (14) (2025).
[68] E. Da e, A. Poho ille, Calcula ing ee ene gies using a e age o ce, J. Chem. Phys.
115 (20) (2001) 9169–9183.
[69] C. Chipo , J. H´
enin, Explo ing he ee-ene gy landscape o a sho pep ide using an
a e age o ce, J. Chem. Phys. 123 (24) (2025).
[70] C. Chipo , A. Poho ille, F ee Ene gy Calcula ions: Theo y and Applica ions in
Chemis y and Biology, 86 o Sp inge Se ies in Chemical Physics, Sp inge , 2007.
[71] J. Come , J.C. Gumba , J. H´
enin, T. Leli`
e e, A. Poho ille, C. Chipo , The adap i e
biasing o ce me hod: e e y hing you always wan ed o know bu we e a aid o
ask, J. Phys. Chem. B 119 (3) (2015) 1129–1151.
[72] D.C. Rees, E. Johnson, O. Lewinson, Abc anspo e s: he powe o change, Na .
Re . Mol. Cell Biol. 10 (3) (2009) 218–227.
[73] L. Lin, S.W. Yee, R.B. Kim, K.M. Giacomini, Slc anspo e s as he apeu ic a ge s:
eme ging oppo uni ies, Na . Re . D ug Disco . 14 (8) (2015) 543–560.
[74] S.K. Nigam, Wha do d ug anspo e s eally do? Na . Re . D ug Disco . 14 (1)
(2015) 29–44.
[75] J. Vanza, P. Jani, N. Pandya, H. Tandel, Fo mula ion and s a is ical op imiza ion o
in a enous emozolomide-loaded pegyla ed liposomes o ea glioblas oma
mul i o me by h ee-le el ac o ial design, D ug De . Ind. Pha m. 44 (6) (2018)
923–933.
[76] A. De, N. Venka esh, M. Sen hil, B.K.R. Sanapalli, R. Shanmugham, V.V.S.R. Ka i,
Sma niosomes o emozolomide o enhancemen o b ain a ge ing,
Nanobiomedicine 5 (2018), 1849543518805355.
Y. Ge e al.
Biophysical Chemis y 324 (2025) 107457
9
