1
Scien i ic RepoR s | (2019) 9:1647 | h ps://doi.o g/10.1038/s41598-018-37599-z
www.na u e.com/scien i ic epo s
Fas NMR me hod o p obe sol en
accessibili y and diso de ed egions
in p o eins
And é F. Faus ino
1,5, Glauce M. Ba bosa2, Micael sil a4, Miguel A. R. B. Cas anho1,
And ea . Da poian2, Eu ico J. Cab i a
4, Nuno C. san os 1, Fabio C. L. Almeida2,3 &
I o C. Ma ins
1
Unde s anding p o ein s uc u e and dynamics, which go e n key cellula p ocesses, is c ucial o
basic and applied esea ch. In insically diso de ed p o ein (IDP) egions display mul i unc ionali y
ia al e na i e ansien con o ma ions, being key playe s in disease mechanisms. IDP egions a e
abundan , namely in small i uses, allowing a la ge numbe o unc ions ou o a small p o eome. The
ela ion be ween p o ein unc ion and s uc u e is hus now seen om a di e en pe spec i e: as IDP
egions enable ansien s uc u al a angemen s, each con o me can play di e en oles wi hin he cell.
Howe e , as IDP egions a e ha d and ime-consuming o s udy ia classical echniques (op imized o
globula p o eins wi h unique con o ma ions), new me hods a e equi ed. He e, employing he dengue
i us (DENV) capsid (C) p o ein and he immunoglobulin-binding domain o s ep ococcal p o ein G,
we desc ibe a s aigh o wa d NMR me hod o di e en ia e he sol en accessibili y o single amino
acid N-H g oups in s uc u ed and IDP egions. We also gain insigh s in o DENV C lexible old egion
biological ac i i y. The me hod, based on minimal pH changes, uses he well-es ablished 1H-15N HsQC
pulse sequence and is easily implemen able in cu en p o ein NMR ou ines. The da a gene a ed a e
simple o in e p e , wi h his apid app oach being an use ul i s -choice IDPs cha ac e iza ion me hod.
Nuclea magne ic esonance (NMR) spec oscopy is he echnique o excellence o ob ain s uc u al and dynam-
ics a omic esolu ion in o ma ion o mac omolecules, especially p o eins1. NMR is compa ible wi h oom em-
pe a u e solu ion measu emen s, a majo ad an age o e o he high- esolu ion s uc u al echniques (X- ay
c ys allog aphy and c yo-elec on mic oscopy). Via NMR, p o ein s uc u es can now be de e mined wi h
chemical shi s da a alone (employing he CS-Rose a package)2–6, which is ex emely impo an when me ely
spa se da a is a ailable3,4. No ewo hy, NMR p o ides p o ein dynamics in o ma ion in physiological condi-
ions, ia he p obing o di e en mac omolecula mo ion imescales7–9. Backbone amide hyd ogen exchange
expe imen s a e pa icula ly in o ma i e o p o ein N-H sol en accessibili y, being ela ed o bo h s uc u e
and dynamics (sensi i e o he millisecond imescale)9–13. Ne e heless, complemen a y da a is equen ly s ill
necessa y14–18, especially o s udy in insically diso de ed p o ein (IDP) egions17,19. The e o e, he e is a majo
unme demand o as s aigh o wa d me hodologies o analyze IDP egions, which play key oles in heal h and
disease mechanisms.
He e, based on he amide hyd ogen exchange p ocess10–12,20–24, we epo a simple and as me hod o gain
in o ma ion on sol en accessibili y o each amino acid esidue N-H g oup o a p o ein. Bene i ing om ou
p e ious wo k25–30, we usesd dengue i us (DENV) capsid (C) p o ein as a model (Fig.1a, PDB ID 1R6R)26,31,32
since i possesses h ee dis inc s uc u al egions (Fig.1b and TableS1): he diso de ed N- e minal, he lexible
old and he conse ed old. The nomencla u e “conse ed old” e e s o a s uc u ally pe sis en old ha was
1Ins i u o de Medicina Molecula , Faculdade de Medicina, Uni e sidade de Lisboa, A . P o . Egas Moniz, 1649-028,
Lisbon, Po ugal. 2ins i u o de Bioquímica Médica Leopoldo de Meis, Uni e sidade ede al do Rio de Janei o, Rio
de Janei o, 21941-902, RJ, B azil. 3cen o nacional de Ressonância Magné ica nuclea , Uni e sidade ede al do
Rio de Janei o and Na ional Ins i u e o S uc u al Biology and Bioimage (CENABIO), Rio de Janei o, 21941-902,
RJ, B azil. 4UciBiO, Depa amen o de Química, aculdade de ciências e ecnologia, Uni e sidade no a de Lisboa,
Quin a da To e, 2829-516, Mon e de Capa ica, Po ugal. 5P esen add ess: iBe , ins i u o de Biologia expe imen al
e Tecnológica, Apa ado 12, 2780-901, Oei as, Po ugal. Co espondence and eques s o ma e ials should be
add essed o n.c.S. (email: [email p o ec ed]) o .c.L.A. (email: [email p o ec ed]) o i.c.M. (email:
[email p o ec ed])
Recei ed: 9 May 2018
Accep ed: 10 Decembe 2018
Published: xx xx xxxx
opeN
www.na u e.com/scien i ic epo s/
2
Scien i ic RepoR s | (2019) 9:1647 | h ps://doi.o g/10.1038/s41598-018-37599-z
ound in he la i i uses capsid p o ein s uc u es o Dengue, Wes Nile and, ecen ly, Zika i uses (wi h PDB
IDs, espec i ely, 1R6R, 1SFK and 5YGH)26,32–34. This app oach employs a small pH a ia ion, p ese ing DENV
C o e all s uc u e and dynamics, which easily allows p obing he backbone N-H g oups’ sol en accessibili y
(using only 1H-15N HSQC peak in ensi ies). The h ee DENV C s uc u e/dynamics egions can be clea ly dis-
inguished ia his me hod, which suppo s no only he cu en unde s anding o DENV C s uc u e/dynamics
p ope ies, bu also he use o his echnique o access and di e en ia e sol en -exposed N-H g oups in IDP
egions. Mo eo e , since he sol en accessibili y o each N-H g oup is ela ed o i s in amolecula H-bond pa -
e n, ou app oach also in o ms on he seconda y s uc u e con en . Impo an ly, his me hodology is eadily
applicable o s udy o he p o eins s uc u e and dynamics. To demons a e his, we also es ed he e ec o a ying
he pH, in a simila 1.5 pH uni s in e al, on he NMR 1H-15N HSQC in ensi ies using ano he model sys em, he
B1 immunoglobulin binding domain o s ep ococcal p o ein G (GB1)35–38. GB1 is pa icula ly in e es ing as i
possesses α-helical and β-shee egions, complemen ing he app oach. As desc ibed ahead, he me hod is eadily
applicable o his p o ein as well, suppo ing i s use.
Resul s
Sui abili y o he p o ein o he pH-based a ia ion app oach used. Ou app oach is based on
small a ia ions o pH wi hin an in e al ha does no lead o majo p o ein con o ma ional changes. DENV C
(Fig.1a,b) was used as a model o s udy he ela ionship be ween he p o ein s uc u e/dynamics and backbone
N-H sol en accessibili y (i.e., he abili y o he N-H hyd ogen o exchange wi h wa e hyd ogens). The in e al
Figu e 1. DENV C s uc u e, o e all dynamics and 1H-15N HSQC spec a om pH 6.0 o 7.5. (a) DENV C
homodime expe imen al s uc u e. This p o ein is posi i ely cha ged, wi h 26 ca ionic and 2 anionic ou o 100
esidues pe monome . F om amino acid esidue 21 o 100, i con ains ou α-helices named α1 o α4 (PDB ID
1R6R32). The i s 20 esidues a e no shown since hey a e in insically diso de ed in solu ion32. (b) Molecula
dynamics simula ion s uc u e o DENV C29, which highligh s he h ee main s uc u e/dynamics egions:
diso de ed N- e minal (blue, esidues 1–22)31,32; lexible old ( ed, esidues 23–44)26; and conse ed old (g een,
esidues 45–100)26,32–34. (c) Supe imposed DENV C 1H-15N HSQC spec a a pH 6.0 (g ay), 6.5 (blue), 6.75
( ed), 7.0 (yellow), 7.25 (g een) and 7.5 (black). Ou app oach equi es only 1H-15N HSQC peak in ensi ies da a.
(d) Zoom on a spec al egion whe e all ypes o esponse o he pH a ia ion a e obse ed: peaks om L44, I59
and K74 a y in in ensi y and chemical shi ; K7, A8 and K73 jus a y in in ensi y; R100 jus a ies in chemical
shi ; and, V26, F84 and R90 nei he a y in in ensi y no in chemical shi . UCSF Chime a 1.9 so wa e54
was used o p o ein s uc u e isualiza ion. The da a ag ees wi h he cu en unde s anding o la i i uses C
p o eins s uc u e/dynamics and biological ac i i y25–34,55.
www.na u e.com/scien i ic epo s/
3
Scien i ic RepoR s | (2019) 9:1647 | h ps://doi.o g/10.1038/s41598-018-37599-z
used is be ween pH 6.0 and 7.5, a physiological ange ha is sui able o mos p o eins, including DENV C, as
desc ibed ahead. Taking ad an age o he pH dependen amide N-H hyd ogen exchange p ocess10–12,20–24, 1H-15N
HSQC spec a o DENV C we e acqui ed a pH 6.0 and 7.5 (Fig.1c,d, g ay and black, espec i ely). Only speci ic
peaks show dec eased in ensi y and/o a chemical shi a ia ion (Fig.1c,d). The a ia ions in in ensi y a ec
mo e peaks and a e mo e p onounced han chemical shi a ia ions, sugges ing no majo con o ma ional change
igge ed by pH. To asce ain ha , we acqui ed 1H-15N HSQC spec a o DENV C a se e al pH alues (6.0, 6.5,
6.75, 7.0, 7.25 and 7.5), assessing he spec al e olu ion as a unc ion o pH (Fig.1c,d). The spec al egion ep e-
sen ed in Fig.1d shows he ou amino acid peaks ha p esen he mos p onounced a ia ion o chemical shi ,
namely L44, K73, K74 and R100. E en o hese esidues, he changes a e minimal, implying he conse a ion o
DENV C o e all a chi ec u e.
We hen consul ed he pKa alues o i a able esidues (Fig.S1)39, since acid-base equilib ium could cause
con o ma ional changes ha would di icul he in e p e a ion o he esul s. DENV C heo e ical isoelec ic
poin is a pH 12.6 and, impo an ly, i s sequence does no con ain amino acid esidues i a able wi hin he pH
ange s udied (Fig.S1, o ange ba ). We also measu ed he NMR ans e se 15N amide elaxa ion a es (R2) in
bo h pH condi ions (Fig.S2), which demons a ed ha he e a e no con o ma ional ansi ions igge ed by pH.
This pa ame e is sensi i e o al e a ions on he size/shape o a p o ein (since, in globula p o eins, i gene ally
inc eases wi h he p o ein hyd odynamic diame e ), as well as o local luc ua ions in he lexibili y o pa icula
amino acid esidues7,9,40–42. I is clea om Fig.S2 ha he R2 alues ob ained o DENV C a e o e all in a ian in
his pH ange. The e o e, aking all o he abo e in o accoun (namely Figs.1c,d, S1 and S2), he o e all DENV C
s uc u al a angemen is main ained. Thus, he pH-induced 1H-15N HSQC spec al di e ences a e solely due o
amide hyd ogen exchange wi h wa e 10–12,20–24, which epo s on sol en accessibili y.
P obing sol en accessibili y o he p o ein backbone. The spec al changes obse ed a e consis en
wi h an amide hyd ogen exchange p ocess (i.e., whe e N-H g oups exchange hei hyd ogens wi h wa e hyd o-
gens)10–12,20–24. Such p ocess only occu s i N-H g oups a e exposed o he sol en and no in an in amolecula
hyd ogen bond. The e o e, hese changes di ec ly epo on N-H g oups’ sol en accessibili y. A cons an empe -
a u e, his exchange p ocess occu s a a a e ha inc eases 10 old pe pH uni 10–12,20–24. Thus, he e, by inc easing
he pH om 6.0 o 7.5, he hyd ogen exchange a e cons an s inc ease 31.6 old (i.e., 10(7.5−6.0)). This causes a
dec ease o he NMR peak in ensi y, since he ac ha he amide p o on s a s o jump mo e equen ly back and
o h be ween he wa e and he amide si es leads o an enhanced decay o he ans e se magne iza ion du ing
acquisi ion. Spec al changes a e he e o e dependen on he ex en o he inc ease o he N-H exchange a e
cons an . Fo he mos sol en accessible N-H g oups, peaks may e en disappea om he spec um a pH 7.5
(e.g., esidues K7, A8 and L44 in Fig.1d). These spec al changes can be highly in o ma i e i p ope ly explo ed,
epo ing on s uc u al and dynamic p ope ies o p o eins. As such, we s udied hem he e, o de elop a me hod
ha p o ides insigh s in o p o ein s uc u e and unc ion, bo h a he indi idual amino acid and domain le el.
To es ablish his new me hodology, we i s compa ed he maximum a ia ion o pH alues es ed, by plo ing
he in ensi ies a pH 6.0 and 7.5 as a unc ion o p o ein sequence (Fig.2a, g ay and back, espec i ely). A pH 6.0,
he N-H g oups o N- e minal egion esidues display highe peak in ensi ies, consis en wi h hei diso de ed
na u e26,31,32. To simpli y he analysis and compensa e o di e ences in ini ial in ensi y (In pH 6.0), esul s we e
no malized by he a io be ween he in ensi ies a pH 7.5 and 6.0 (In pH 7.5/In pH 6.0; Fig.2b). The whole N- e minal
egion and speci ic esidues loca ed in he α1 and nea loop egions dec ease hei in ensi y as he pH inc eases
o 7.5. These indings a e wo h conside ing in he con ex o DENV C h ee main s uc u al egions (Fig.1b and
TableS1). B ie ly, a pH 7.5, he peak in ensi y o some esidues is less han hal o hei ini ial alues (Fig.2b),
namely: R5 o R22 (excep P12) in he N- e minal egion; V23 and T25 in he D1 domain; Q27 and T30 in α1;
S34, R41, G42 and L44 in L1-2; A49 in α2; I59 and G64 nea L2-3; K74, S75 and K76 nea L3-4; and, R99 in he
C- e minal domain. As such, he h ee main s uc u e/dynamics egions o DENV C a e clea ly dis inguishable,
as u he de ailed ahead.
A e age sol en accessibili y o p o ein egions. Gi en he abo e, we hen analyzed he in ensi y
changes in he con ex o he p o ein s uc u e and dynamics o he main egions o he p o ein. Fo such pu pose
and al hough each N-H g oup o an indi idual amino acid beha es di e en ly in esponse o pH10–12,20–24, we con-
side ed ha s uc u al ac o s a e mo e de e minan and we a e aged he backbone N-H g oup esponse o pH
ac oss egions. Those ha a e p o ec ed, ei he by being bu ied wi hin he s uc u e o wi hin an in amolecula
hyd ogen bond, will no be a ec ed by pH. The amino acids ha a e no p o ec ed will be esponsi e o pH wi hin
he pH ange es ed he e. The a ios de e mined in Fig.2b, when a e aged ac oss a p o ein egion o domain,
p o ide a single pa ame e o dis inguish be ween s uc u al and dynamics sec ions.
Fig.3 depic s he a e age In pH 7.5/In pH 6.0 o he h ee main s uc u e/dynamics egions (le panel) and o he
seconda y s uc u e domains ( igh panel). The main s uc u al egions a e dis inguished by hei a e age back-
bone sol en accessibili y (Fig.3, le panel): he diso de ed N- e minal backbone is highly exposed, he lexible
old is pa ially accessible, and he conse ed old is mos ly inaccessible o he sol en . Looking a he seconda y
s uc u e domains (Fig.3, igh panel), he a e age alues show ha he backbone o he α0 domain, which is
diso de ed and may ansien ly adop an α-helical seconda y s uc u e29, is highly exposed o he sol en . Among
he α-helices, he backbone o α1 p esen s alues ha a e in be ween he ob ained o α0 and hose o he α2, α3
and α4 backbones, sugges ing an in e media e exposu e o α1 backbone o he sol en , implying a ce ain deg ee
o lexibili y. The e o e, DENV C α1 has mo e eedom o in e ac wi h he sol en , in line wi h ou p e ious s ud-
ies26,29. Mo eo e , he a e age backbone alues o loop egions L2-3 and L3-4 a e, in gene al, simila o hose o
nea by α-helices (Fig.3, igh panel). Thus, his app oach p obes he di e ences in backbone N-H g oups sol en
accessibili y o he main s uc u e/dynamics egions (Fig.3, le panel), as well as o he seconda y s uc u e
domains (Fig.3, igh panel).
www.na u e.com/scien i ic epo s/
4
Scien i ic RepoR s | (2019) 9:1647 | h ps://doi.o g/10.1038/s41598-018-37599-z
1H-15N HSQC peak in ensi ies in unc ion o pH. The changes in he 1H-15N HSQC spec um o each
pH es ed (6.0, 6.5, 6.75, 7.0, 7.25 and 7.5) we e assigned o he espec i e indi idual N-H g oups o he p o ein
(Fig.1c,d), o gi e a comple e pic u e o he peak e olu ion wi h pH. Fig.4a shows he e olu ion o 1H-15N HSQC
peak in ensi ies as a unc ion o pH o h ee esidues (M15, T30 and R97) ep esen a i e o he h ee main s uc-
u al egions. Impo an ly, he sol en accessibili y p obed ia he app oach p esen ed epo s he in e ac ion
o each speci ic amide g oup wi h wa e . I can be used o dis inguish he sol en accessibili y o backbone and
side-chain N-H g oups wi hin he same esidue, as shown o W69 N-H g oups (Fig.S3), whe e he backbone
amide is no a ec ed by pH, while he indole N-H g oup alue a ies signi ican ly. The e o e, each N-H g oup
epo s i s own mic oen i onmen . The highly localized p obing sensi i i y illus a es he me hodology g ea
esolu ion le el, a p ope y ha can be exploi ed o gain i al s uc u al and dynamics in o ma ion.
Besides his, we can analyze he no malized in ensi y o he backbone N-H g oups, as a unc ion o pH, o
each amino acid, wi h s iking di e ences be ween amino acids om di e en egions o he p o ein (Fig.4a). An
a e age o he no malized in ensi y o all he amino acids o each main egion can hen be ob ained (Fig.4b). The
h ee key s uc u al egions o DENV C a e clea ly dis inguished (Fig.4b): he conse ed old su e s no majo
changes (g een), he N- e minal su e s he g ea es change (blue), while he lexible old shows an in e media e
egime ( ed). The lexible old also has la ge e o ba s (Fig.4b, ed), indica i e o highe he e ogenei y among
he cons i u ing esidues sol en accessibili y. Fo each seconda y s uc u e domain, he a e age in ensi ies as a
unc ion o pH a e a ailable in Fig.S4. Since he α0 domain29 is mos ly diso de ed in solu ion26,31,32, i s backbone
a e age sol en accessibili y is highe han o o he α-helical domains, as expec ed. Impo an ly, α1 displays an
in e media e accessibili y and he o he α-helical domains backbones a e gene ally no exchanging he amide
hyd ogen wi h he sol en , in ag eemen wi h Fig.3 da a. Rega ding loop domains, D0 and L1-2 ha e hei back-
bone N-H g oups mos ly in e ac ing wi h he sol en , while o he loops a e oughly unable o pe o m amide
hyd ogen exchange, in acco dance wi h he analysis o Fig.3. The e o e, we can ob ain a single pa ame e ha
desc ibes indi idual and egional exposu e o he sol en , as desc ibed he ea e .
Figu e 2. DENV C 1H-15N HSQC peak in ensi ies a pH 6.0 and pH 7.5, and hei a io. (a) DENV C 1H-15N
HSQC peak in ensi ies a pH 6.0 (g ay ba s) and pH 7.5 (black ba s), and (b) a io o HSQC peak in ensi ies
a pH 7.5 and pH 6.0 (In pH 7.5/In pH 6.0). E o ba s ep esen s anda d e o (SE). The symbols in a encode he
eason why he espec i e esidues could no be analyzed by NMR: ‘#’ o esidues ha a e no assigned, ‘o’ o
o e laps, ‘*’ o absen esonances due o line b oadening, and ‘P’ o p olines. The ho izon al line on b ma ks
he a io equal o 1. The main s uc u al ea u es a e indica ed on he op o he igu e: he h ee s uc u e/
dynamics egions26,31,32, he seconda y s uc u e domains29,31,32, and he p o ein p ima y sequence. Colo ed
columns a e a guide o he da a co esponding o each seconda y s uc u e domain (pink columns ep esen
he expe imen ally de e mined α-helices31,32, while he cyan column co esponds o he ansien α-helix
sugges ed by ou p e ious wo k29).
www.na u e.com/scien i ic epo s/
5
Scien i ic RepoR s | (2019) 9:1647 | h ps://doi.o g/10.1038/s41598-018-37599-z
Linea i y o in ensi ies e sus pH. The backbone N-H peak in ensi ies o indi idual amino acid esidues
ollow a oughly linea dec ease wi h pH (Fig.4a), which is also obse ed o he a e age o he main egions
(Fig.4b) and domains (Fig.S4). As such, an app oxima ion was used by i ing he ollowing empi ical linea
equa ion o he da a:
≈×−. +
.
In
In Slope(pH 60)1
(1)
pH 60
whe e In is he 1H-15N HSQC peak in ensi y a a gi en pH and In pH 6.0 is he a e age in ensi y om 3 inde-
penden measu emen s a pH 6.0. The i ing o his equa ion e ie es he slope, which is a pa ame e ha
desc ibes he a e age alue o he de i a i e d(In /In pH 6.0)/dpH h oughou he pH in e al p obed. A o mal
app oach was also de ised based on he li e a u e10–12,20–24, which can be ound on he Supplemen a y No e (o
he Supplemen a y In o ma ion ile), leading o he pH dependencies o bo h In /In pH 6.0 and d(In /In pH 6.0)/dpH.
Impo an ly, his simple slope-based (linea ) app oach e ie es a single i ing pa ame e ha en i ely desc ibes
he end, being independen o ex e nal pa ame e s es ima ion ha some imes a e di icul o de e mine (i.e.,
k c o alues; o de ails, please consul he Supplemen a y No e). In p ac ice, he mo e nega i e is he slope, he
mo e suscep ible o exchange is he co esponding N-H g oup. Slopes and In pH 7.5/In pH 6.0 alues a e compa a-
ble, as explained he ea e . Since slopes a e o igina ed om measu emen s a se e al pH alues, hey a e a be e
pa ame e o ep esen each N-H g oup sol en accessibili y, being o use o mo e ad anced applica ions, and
we e employed hence o h.
DENV C s uc u e/dynamics and he slope in o ma ion. Slope alues we e calcula ed ia equa ion 1
o each analyzable DENV C backbone N-H g oup (Fig.5a). The a e age slope alues o he h ee majo egions
and o he seconda y s uc u e domains (Fig.S4) we e hen compu ed (Fig.5b). The in o ma ion ob ained is
simila o he one de i ed om he In pH 7.5/In pH 6.0 alues (compa e Figs.5a and 2b, and also Figs.5b and 3). In
Fig.5a, i is easy o dis inguish he indi idual N-H g oups ha a e ully exposed, in e media ely exposed o bu ied
away om he sol en . This is also clea in Fig.5b o he h ee main egions and he se e al seconda y s uc u e
domains. No ewo hy, wi hin a gi en egion, in e connec ing loops seem o be mo e dynamic and exposed han
adjacen α-helical domains, in consonance wi h he p o ein s uc u e. Slope alues o 0 (Fig.5a) a e om N-H
g oups o esidues ha canno change hei hyd ogen wi h he sol en (co esponding o In pH 7.5/In pH 6.0 alues
o 1, in Fig.2b). Slopes wi h absolu e alue highe han 0.7 (Fig.5a) a ise om N-H g oups which a e pe o m-
ing H-bonds wi h he wa e (co esponding o In pH 7.5/In pH 6.0 alues o 0, in Fig.2b). A slope h eshold o −0.4
dis inguishes he mo e sol en exposed N-H g oups (Fig.5a, yellow) om hose less exposed (Fig.5a, g ay). A
de ailed analysis o Fig.5a using his h eshold e eals ha he mos sol en accessible N-H g oups a e om
esidues R5 o R22, T25, Q27, T30, S34, G42, L44, I59, G64, S75, K76 and R99. These esidues a e acco dingly
Figu e 3. A e age o he NMR peak in ensi ies a io o he h ee s uc u e/dynamics egions and seconda y
s uc u e domains o DENV C. The NMR peak in ensi ies a io be ween pH 7.5 and pH 6.0 o each esidue
(da a om Fig.2b) we e a e aged ac oss he esidues ha comp ise each o he h ee s uc u e/dynamics
egions (le panel) and seconda y s uc u e domains ( igh panel) o DENV C. E o ba s a e SE.

www.na u e.com/scien i ic epo s/
6
Scien i ic RepoR s | (2019) 9:1647 | h ps://doi.o g/10.1038/s41598-018-37599-z
depic ed in he p o ein s uc u e (Fig.5c, ma ching yellow and g ay esidues), p o iding di ec in o ma ion on
bo h IDP and o de ed egions o DENV C p o ein, which a e immedia ely dis inguishable. Mo eo e , many
o hese esidues a e loca ed a he beginning o all he p o ein α-helices, which gi es in o ma ion on p o ein
s uc u e.
The ac ha many o he i s α-helical esidues ha e N-H g oups exposed o he sol en is a di ec insigh
in o he na u e o α-helices. In an α-helix, he i s esidue is es ablishing H-bond ia i s C=O g oup wi h he
N-H g oup o he ou h esidue, lea ing i s own N-H g oup ee o H-bonds in ol ed in he α-helix s abiliza-
ion. This means ha he N-H g oups o he i s h ee esidues o α-helices a e ee o es ablish H-bonds wi h
o he nucleophile g oups ( ha se e as hyd ogen bond accep o s) ei he om he p o ein, becoming una ailable
o he sol en , o om he sol en . I hey a e exposed o he wa e , hei amide hyd ogen can exchange wi h
hose om he sol en . This is exac ly wha we obse e in DENV C α-helices, by analyzing he backbone N-H
g oups pe o ming in amolecula H-bonds wi hin he DENV C s uc u e (Fig.S5a). We hen compa ed he
slopes in o ma ion wi h he no malized equency o in amolecula H-bonds pe N-H g oup (Fig.S5b), inding
a clea co ela ion o he slopes wi h DENV C s uc u e. In e es ingly, esidues ha ha e low equency o N-H
in amolecula H-bonds (<0.5) and low slope alues (be ween −0.4 and 0) a e hus ee o pe o m hyd ogen
exchange, bu a e unable o do so. This sugges s ha hey a e no acing he sol en because hey a e bu ied wi hin
he p o ein. In summa y, an N-H g oup om a speci ic esidue needs o be bo h ee o in amolecula H-bonds
and exposed o he sol en in o de o exchange i s hyd ogen wi h he wa e . O e all, ou indings sugges ha he
p obing o he N-H g oups’ sol en accessibili y o a p o ein, ia mino pH changes, may be used as an addi ional
s uc u e and dynamics es ain o help on he calcula ion o p o ein s uc u es.
Figu e 4. No malized peak in ensi ies o indi idual N-H g oups and he a e age among DENV C s uc u e/
dynamics egions, as a unc ion o pH. (a) In /In pH 6.0 a ia ion wi h pH o he backbone N-H g oups o M15,
T30 and R97, ep esen ing esidues in he diso de ed N- e minal, lexible old and conse ed old egions,
espec i ely. (b) A e age in ensi ies a io o he h ee majo s uc u e/dynamics egions o DENV C. In all
g aphs, lines a e i s o equa ion 1 o he da a, om which slopes we e ex ac ed.
www.na u e.com/scien i ic epo s/
7
Scien i ic RepoR s | (2019) 9:1647 | h ps://doi.o g/10.1038/s41598-018-37599-z
Applying he me hod o GB1 p o ein. Ha ing es ablished he me hod applicabili y wi h DENV C, we
p oceeded o es i wi h he B1 immunoglobulin binding domain o s ep ococcal p o ein G (GB1), which con-
ains 56 amino acid esidues and a s uc u e o ou s anded β-shee s wi h one long α-helix on op (Fig.6a), as
shown by X-Ray di ac ion c ys allog aphy as well as by NMR (PDB ID: 2GB1 and 5JXV)35,36,43. GB1 has been
ex ensi ely s udied by di e en biophysical me hods and is one o he smalles s able olded globula domains
known. A pH in e al o 1.5 was assayed as well, bu now changing he pH om 6.5 o 8.0. No majo con o ma-
ional changes we e seen (Fig.6b), only mino local swi ches (Fig.6c,d). As he o e all p o ein s uc u e emains
highly s able wi hin ha pH ange37, we wen u he and es ed i he in ensi ies o he 1H-15N HSQC peaks
e ealed any changes (Fig.7). As o DENV C (Fig.2), by di ec ly compa ing peak in ensi ies a pH 6.5 and 8.0
on GB1 (Fig.7a) o he a io be ween hese in ensi ies (Fig.7b), he majo egions o he p o ein wi h exposed
backbone amide ni ogen a oms can be eadily iden i ied, namely he loops, he ou e s ands o he ou -s anded
β-shee (i.e., β2 and β3) and he beginning o he α-helix, which a e ee o backbone in amolecula H-bonds and
accessible o he sol en .
Then, wi h he abo e in mind, we es ed he use o he slope o map he p o ein egions mos accessible o he
sol en (Figs.8 and S6), using he same cu -o as o DENV C. The in o ma ion i s well wi h he known pa e n
Figu e 5. Slopes o In /In pH 6.0 e sus pH in he con ex o he DENV C s uc u e. (a) Slope o he in ensi ies
a io e sus pH along he DENV C sequence (an in e se scale is shown since he mo e e icien is he hyd ogen
exchange p ocess, he mo e nega i e is he slope). The h eshold o −0.4 (dashed line) was de ined o iden i y
he DENV C backbone N-H g oups ha a e highly exposed o he sol en (yellow ba s). Fo de ails on he
p o ein s uc u al in o ma ion and symbols, on op and wi hin he g aph ( espec i ely), please e e o he
legend o Fig.2. (b) A e age slopes o he h ee s uc u e/dynamics egions (b, le panel) and o he seconda y
s uc u e domains (b, igh panel) o DENV C. E o ba s in a and b ep esen SE. (c) DENV C esidues in
which he backbone N-H is highly exposed o he wa e we e highligh ed wi hin he p o ein s uc u e (yellow
egions). Clea ly, om a and c, all he esidues o he diso de ed N- e minal egion, some speci ic esidues on
he lexible old egion and esidues in he beginning o he α-helices a e able o exchange hei backbone amide
hyd ogen wi h he wa e .
www.na u e.com/scien i ic epo s/
8
Scien i ic RepoR s | (2019) 9:1647 | h ps://doi.o g/10.1038/s41598-018-37599-z
o GB1 sol en -accessible su ace a ea (SASA) along he sequence44, suppo ing he me hodology employed.
Mo eo e , ou esul s a e in acco dance wi h H/D exchange s udies, as he same egions iden i ied as being mo e
accessible o exchange wi h he sol en a e he mos sensi i e o pH38. Mo e p o ec ed esidues ha exchange
h ough a global un olding mechanism (e.g., esidues K4, L5, A26, F30 and T44) o a local high ene gy un olding
mechanism (e.g., esidues K28, Y33, N35 and T55) display minimal changes wi h pH, while he egions ha co -
espond o as exchanging non H-bonded N-H g oups a e clea ly isible (e.g., esidues T17, E19 and V21). This
is also suppo ed by he in amolecula H-bonds equency analysis (Fig.S6), simila ly o DENV C (Fig.S5). All
his in o ma ion u he alida es he me hodology employed and sugges s i s applicabili y in o he s udies, as
discussed ahead.
Discussion
Taking ad an age o he amide hyd ogen exchange wi h wa e 10–12,20–24, we es ablished a new NMR app oach
o de e mine p o ein backbone sol en accessibili y. Sol en accessibili y co ela es wi h he gene al s uc u e/
dynamics egions o a p o ein, wi h he explana ion being s aigh o wa d: in mo e dynamic egions he N-H
g oups a e mo e suscep ible o hyd ogen exchange wi h he sol en , as hey a e no in ol ed in s able seconda y
s uc u e elemen s. We may hus p obe sol en accessibili y by sligh ly a ying he pH o a p o ein solu ion, as
demons a ed by he s udy wi h DENV C, whe e a 1.5 pH uni s in e al was employed (pH 6.0 o 7.5). The main
p o ein egions (Fig.1b) can be clea ly dis inguished ia hei indi idual amino acids (Figs.2b and 5a) and hei
a e aged backbone sol en exposu e (Figs.3, 4b and 5b). We can disc imina e be ween he exposed diso de ed
N- e minal, he pa ially accessible lexible old and he mos ly inaccessible conse ed old, in acco dance wi h
he cu en unde s anding o DENV C p ope ies26,29,31,32. No ewo hy, he lexible old in e media e beha io ,
p e iously p edic ed26, is con i med he e (Figs.3, 4b and 5b, ed). Fu he mo e, a single esidue esolu ion,
he sol en accessibili y o each N-H g oup clea ly co ela es wi h i s no malized equency o in amolecula
H-bonds in DENV C s uc u e (Fig.S5). Such co ela ion is speci ic o each seconda y s uc u e elemen (such as
α-helices and β-shee s), depending on he ela i e posi ion in he p o ein s uc u e. The same app oach was hen
applied o a di e en sys em, GB1, a well s uc u ed p o ein domain, he B1 domain o immunoglobulin p o ein
Figu e 6. GB1 s uc u e and 1H-15N HSQC spec a om pH 6.5 o 8.0. (a) GB1 expe imen al s uc u e. GB1
is a domain o immunoglobulin G binding p o ein ha is nega i ely cha ged, wi h 6 ca ionic and 10 anionic
esidues ou o 56. I consis s o ou β-shee s, named β1 o β4 (colo ed in o ange), plus one α-helix, named
α1 (colo ed in pink), connec ed by sho loops (PDB ID 5JXV43). (b) Supe imposed GB1 1H-15N HSQC
spec a a pH 6.5 (blue), 7.3 ( ed), 7.6 (g een) and 8.0 (black). Majo chemical shi s co esponding o la ge
con o ma ional ea angemen s a e no obse ed. (c and d) Zoom on spec al egions whe e all ypes o
esponses o he pH a ia ion a e obse ed: peaks ha a y in in ensi y and/o display a small chemical shi
pe u ba ion. Mos peaks nei he a y in in ensi y no in chemical shi . UCSF Chime a 1.9 so wa e54 was
used o p o ein s uc u e isualiza ion43.
www.na u e.com/scien i ic epo s/
9
Scien i ic RepoR s | (2019) 9:1647 | h ps://doi.o g/10.1038/s41598-018-37599-z
G. We employed ano he physiological in e al o 1.5 pH uni s, om 6.5 o 8.0 (Fig.6). The da a shows ha also
wi h his p o ein he e a e no majo con o ma ional changes. Mo eo e , he key egions o he p o ein a e also
eadily iden i ied (Fig.7), namely a he single amino acid le el (Fig.8).
The e o e, he pH ange es ed (6.0–8.0) is pa icula ly sui able since mos p o eins isoelec ic poin is no
close o he physiological pH (since hey would no be unc ional as hey could p ecipi a e) and, apa om his i-
dines, i a able amino acids a e usually no a ec ed a his pH ange, as in e ed om Fig.S1. E en i mino con-
o ma ional changes occu , hese would be eadily isible in he NMR spec a h ough chemical shi pe u ba ion
(CSP) s udies, allowing i o be aken in o accoun in he analysis. So, he me hodology can be easily applied o
o he p o eins, in pH anges whe e hei gene al s uc u e and dynamics p ope ies a e main ained. A e NMR
assignmen a low pH (e.g., pH 6.0), one can pe o m a pH inc ease on he same sample (a leas one pH uni is
ecommended) and acqui e one mo e 1H-15N HSQC spec um o de e mine immedia ely which N-H g oups and
p o ein egions in e ac wi h he sol en . A single p o ein p epa a ion (a concen a ions a ound 5-10 mg/mL)
can be used and wo measu emen s a di e en pH alues eadily p o ide key in o ma ion. E en i he e is a mild
con o ma ional change igge ed by pH and/o i a ion o some speci ic amino acid esidue(s), a pH a ia ion
can be pe o med by acqui ing 1H-15N HSQC spec a a se e al sligh ly spaced pH alues, o ollow he NMR
peaks e olu ion (CSP analysis). In pa icula , he pH ange used he e (be ween 6.0 and 8.0) has he ad an age o
being physiologically ele an and compa ible wi h he imescale o usual 1H-15N HSQC measu emen s.
Mo eo e , he app oach is pa icula ly use ul o s udy IDP egions o when NMR da a is spa se due o ime o
o he cons ain s. As men ioned, he in ensi y o a signal depends on he line wid h, which is mainly in luenced
by he p o ein co ela ion ime and he chemical exchange egime: sha p o IDPs (due o co ela ion ime below
1 ns, bu po en ially b oadened due o accessibili y o wa e exchange) while amides in he olded pa o a p o ein
will expe ience less line b oadening due o slowe wa e exchange. O e all, i one changes he pH by 1.5 uni s, he
base-ca alyzed hyd ogen exchange inc eases by 101.5, which has la ge e ec s on he in ensi y ( he ecip ocal line
wid h) o he IDP signals and a smalle e ec on he olded pa s, as eadily obse ed he e. O he me hodologies
ha e been desc ibed o imp o e p o ein s uc u e de e mina ion, namely esidual dipola couplings (RDCs)14,17,
di usion enso pa ame e s15, elaxa ion pa ame e s16 and pa amagne ic elaxa ion enhancemen (PRE) p obes17,
among o he s. Ne e heless, since he me hod desc ibed he e is much simple and easie o in e p e , we belie e
ha i will be widely adop ed o p o ein s uc u al and dynamics s udies.
Figu e 7. GB1
1H-15N HSQC peak in ensi ies a pH 6.5 and pH 8.0, and a io be ween peaks. (a) GB1 1H-15N
HSQC peak in ensi ies a pH 6.5 (g ay ba s) and pH 8.0 (black ba s), and (b) a io o HSQC peak in ensi ies
be ween pH 8.0 and pH 6.5 (In pH 8.0/In pH 6.5). E o ba s ep esen s anda d e o (SE). The M1 esidue is no
assigned, while Q2 was no analyzed due o absen esonance, as a esul o line b oadening. All o he amino
acid esidue peaks we e assigned and analyzed. The ho izon al line on b ma ks he a io equal o 1. The main
s uc u al ea u es a e indica ed on he op o he igu e: he seconda y s uc u e domains36, and he p o ein
p ima y sequence. Colo ed columns a e a guide o he da a co esponding o each seconda y s uc u e domain
(o ange and pink columns co espond o expe imen ally de e mined β-shee s and α-helix, espec i ely.