Phospho y osine couples pep ide binding and SHP2 ac i a ion ia a
dynamic allos e ic ne wo k
Michelangelo Ma asco
a,1
, John Ki kpa ick
a,b,1
, Vi o ia Nanna
a
, Jus yna Siko ska
b
, Te esa Ca lomagno
a,b,
⇑
a
Leibniz Uni e si y Hanno e , Cen e o Biomolecula D ug Resea ch and Ins i u e o O ganic Chemis y, Schneide be g 38, 30167 Hanno e , Ge many
b
Helmhol z Cen e o In ec ion Resea ch, G oup o NMR-based S uc u al Chemis y, Inho ens asse 7, 38124 B aunschweig, Ge many
a icle in o
A icle his o y:
Recei ed 22 Feb ua y 2021
Recei ed in e ised o m 14 Ap il 2021
Accep ed 16 Ap il 2021
A ailable online 20 Ap il 2021
Keywo ds:
SHP2
PD-1
NMR spec oscopy
Molecula dynamics
Allos e ic coupling
abs ac
SHP2 is a ubiqui ous p o ein y osine phospha ase, whose ac i i y is egula ed by phospho y osine (pY)-
con aining pep ides gene a ed in esponse o ex acellula s imuli. I s c ys al s uc u e e eals a closed,
au o-inhibi ed con o ma ion in which he N- e minal S c homology 2 (N-SH2) domain occludes he ca -
aly ic si e o he phospha ase (PTP) domain. High-a ini y mono-phospho yla ed pep ides p omo e ca -
aly ic ac i i y by binding o N-SH2 and dis up ing he in e ac ion wi h he PTP. The mechanism
behind his p ocess is no en i ely clea , especially because N-SH2 is incapable o accommoda ing com-
ple e pep ide binding when SHP2 is in he au o-inhibi ed s a e. He e, we show ha pY pe o ms an essen-
ial ole in his p ocess; in addi ion o i s con ibu ion o o e all pep ide-binding ene gy, pY- ecogni ion
leads o enhanced dynamics o he N-SH2 EF and BG loops ia an allos e ic communica ion ne wo k,
which des abilizes he N-SH2–PTP in e ac ion su ace and simul aneously gene a es a ully accessible
binding pocke o he C- e minal hal o he phosphopep ide. Subsequen ly, ull binding o he phospho-
pep ide is associa ed wi h he s abiliza ion o ac i a ed SHP2. We demons a e ha his allos e ic ne -
wo k exis s only in N-SH2, which is di ec ly in ol ed in he egula ion o SHP2 ac i i y, while he C-
e minal SH2 domain (C-SH2) unc ions p ima ily o ec ui high-a ini y biden a e phosphopep ides.
Ó2021 The Au ho (s). Published by Else ie B.V. on behal o Resea ch Ne wo k o Compu a ional and
S uc u al Bio echnology. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i e-
commons.o g/licenses/by-nc-nd/4.0/).
1. In oduc ion
Cellula signaling cascades equi e p ecise spa io empo al con-
ol o mul iple binding e en s and hus ely on coo dina ed uning
o a ini ies and speci ici ies wi hin he ele an molecula ne -
wo ks [1]. In signaling pa hways based on y osine phospho yla-
ion, phospho y osine (pY) esidues, gene a ed by p o ein
y osine kinases (PTKs), a e ecognized by specialized p o ein mod-
ules, o which S c homology 2 (SH2) domains a e by a he mos
p e alen and well-cha ac e ized [2,3]. SH2 domains consis o a
globula old o app oxima ely 100 amino acids, wi h a h ee-
s anded an ipa allel b-shee co e lanked by wo
a
-helices
(Fig. 1A) [4]. Pep ides ha ca y a pY ma k (phosphopep ides) usu-
ally bind o hei cogna e SH2 domains in an ex ended con o ma-
ion, wi h he pep ide backbone a anged pe pendicula ly o he
cen al bshee ; he pY is held in place by a ne wo k o sal b idges,
whe e he in a ian esidue A gbB5 (seconda y-s uc u e and
amino-acid nomencla u e acco ding o Eck and cowo ke s [5],Fig-
u e S1A) plays a cen al ole. Fu he in e ac ions wi h esidues
adjacen o pY con e speci ici y, and hus make SH2 domains mod-
e a ely selec i e, despi e hei s uc u al in a iance [6,7]. In mos
cases, a hyd ophobic pa ch lanked by he EF and BG loops makes
ex ensi e con ac s wi h phosphopep ide esidues C- e minal o he
pY [8].
In e es ingly, some SH2 domains ha e e ol ed he abili y o
egula e he ac i i y o ca aly ic domains ha a e pa o he same
p o ein. P o ein y osine phospha ase SHP2, a 70-kDa p o ein
encoded by he p pn11 gene on ch omosome 12 [9], is comp ised
o wo SH2 domains a anged in andem ( e med N-SH2 and C-
SH2), a ca aly ic PTP domain and a diso de ed C- e minal ail wi h
hi he o unclea unc ion (Figu e S1B) [10]. In i s basal s a e, SHP2
is almos inac i e, due o he inhibi o y in amolecula in e ac ion
o he N-SH2 and PTP domains; he binding o phosphopep ides
igge s a con o ma ional change ha opposes he binding o he
N-SH2 domain o he PTP ca aly ic si e, he eby ac i a ing he
phospha ase (Figu e S1C) [11]. No ably, biden a e pep ides, wi h
wo p ope ly spaced pY esidues, a e much s onge ac i a o s
han mono-phospho yla ed pep ides [12,13].
h ps://doi.o g/10.1016/j.csbj.2021.04.040
2001-0370/Ó2021 The Au ho (s). Published by Else ie B.V. on behal o Resea ch Ne wo k o Compu a ional and S uc u al Bio echnology.
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/).
⇑
Co esponding au ho a : Leibniz Uni e si y Hanno e , Cen e o Biomolecula
D ug Resea ch and Ins i u e o O ganic Chemis y, Schneide be g 38, 30167
Hanno e , Ge many.
E-mail add ess: [email p o ec ed] (T. Ca lomagno).
1
Equal con ibu ion
Compu a ional and S uc u al Bio echnology Jou nal 19 (2021) 2398–2415
jou nal homepage: www.else ie .com/loca e/csbj
As an almos unique example o a p o ein y osine phospha ase
wi h posi i e a he han nega i e signaling unc ion, SHP2 is
in ol ed in se e al signaling pa hways, including Ras/MAPK, PI3K
and PD-1 (Fig. 1B) [14–16]. Immune checkpoin ecep o PD-1
(P og ammed Dea h-1) ec ui s and ac i a es SHP2 wi h wo phos-
pho yla ed y osines, Y223 and Y248, embedded in he Immune
Ty osine Inhibi o y Mo i (ITIM) and he Immune Ty osine Swi ch
Mo i (ITSM), espec i ely [17,18]. This in e ac ion esul s in a
educ ion o he unc ionali y o T lymphocy es and may also be
exploi ed by cance cells o e ade immune su eillance [14].In
o he cellula pa hways, phospho y osine si es o he IRS1 (Insulin
Recep o Subs a e 1) diso de ed C- e minus, such as pY1172 and
pY1222, ac i a e SHP2 [19]. IRS1 is a la ge sca olding p o ein ha
plays a key ole in ansmi ing signals o igina ing om insulin and
insulin-like g ow h ac o (IGF) [20]; ne e heless, he ole o he
SHP2–IRS1 in e ac ion is poo ly cha ac e ized. Finally, he diso -
de ed C- e minus o SHP2 i sel also con ains wo phospho yla ion
si es, Y542 and Y580 (Figu e S1B) [9], whose ole is s ill unde
deba e: hey a e hough o ec ui o he p o eins, including G b2
[21–23], bu hei cis in e ac ion wi h he N-SH2 and C-SH2
domains o SHP2 has also been p oposed [24,25].
The wo SH2 domains o SHP2 ha e a unique ea u e, sha ed
only wi h he SH2 domains o he closely ela ed y osine phos-
pha ase SHP1, namely he p esence o a glycine ins ead o an a gi-
nine a posi ion
a
A2. As a consequence, he phosphopep ides ha
bind hese SH2 domains ca y a hyd ophobic esidue a posi ion
Fig. 1. Binding o SHP-2 SH2 domains o phosphopep ides o di e en composi ion. (A) The a chi ec u e o SHP2 N-SH2 bound o IRS1-pY1172 highligh s he ea u es o a
ypical SH2 domain–phosphopep ide complex; seconda y s uc u e elemen s (black labels) and pep ide esidues ( ed labels) a e named acco ding o he scheme in oduced
by Eck e al., 1993 (see also igu e S1A). (B) Schema ics o he h ee p o eins in ol ed in his s udy, showing he si es subjec o y osine phospho yla ion. (C) Uppe le panel:
sequences o he pep ides used in his s udy. Uppe igh : K
D
alues de e mined o PD-1-, IRS1- and SHP2-de i ed phosphopep ides binding ei he N-SH2 o C-SH2. The K
D
alues de e mined by su ace plasmon esonance a e aken om he li e a u e [19]; hose de e mined by ITC we e measu ed in his wo k. Lowe panel:
1
H,
15
N-HSQC spec a
o N-SH2 and C-SH2 in he p esence o inc easing concen a ions o pep ide. The species p esen in each sample a e indica ed abo e he indi idual spec a. The p o ein
concen a ion was 200
l
M and all spec a we e measu ed on a 600 MHz spec ome e a 298 K. (Fo in e p e a ion o he e e ences o colo in his igu e legend, he eade is
e e ed o he web e sion o his a icle.)
M. Ma asco, J. Ki kpa ick, V. Nanna e al. Compu a ional and S uc u al Bio echnology Jou nal 19 (2021) 2398–2415
2399
pY–2 ( wo amino acids be o e pY) [26,27]. Apa om his common
ea u e, he N-SH2 and C-SH2 domains a e unc ionally e y di e -
en and hei pep ide binde s ha e non-o e lapping consensus
mo i s: C-SH2 is selec i e o pep ides wi h sequence (T/V/I/y)
XpY(A/s/ )X(I/V/L), while N-SH2 is mo e p omiscuous and i s bind-
ing pa ne s can be di ided in o ou di e en classes, based on
hei consensus sequence [28]. Fu he mo e, N-SH2 is di ec ly
esponsible o he unc ional egula ion o SHP2: in he basal s a e,
he DE loop o N-SH2 is deeply inse ed in he ca aly ic pocke o
he PTP domain, blocking phospha ase ac i i y. Phosphopep ide
binding o N-SH2 is associa ed wi h a con o ma ional change in
he EF and BG loops, which educes he deg ee o su ace comple-
men a i y be ween he N-SH2 and PTP domains and p e en s hei
in amolecula in e ac ion (Figu e S1C). Con e sely, C-SH2 appea s
o play a ole in he PTP ac i a ion only ia biden a e phosphopep-
ides, which simul aneously engage bo h SH2 domains o ensu e
maximum s imula ion o SHP2 ac i i y [17]. Unlike N-SH2, he
s uc u e o C-SH2 is la gely una ec ed by pep ide binding [11].
He e, we s udy he SH2 domains o SHP2 o unde s and how
small a ia ions in p o ein sequence and dynamics in luence bind-
ing speci ici y and how binding o pY-con aining pep ides is cou-
pled o he egula ion o enzyma ic ac i i y. We selec h ee
SHP2-binding mo i s, co esponding o phosphopep ides de i ed
om IRS1 (pY1172 and pY1222), PD-1 (ITIM and ITSM) and he dis-
o de ed C- e minus o SHP2 i sel (pY542 and pY580) and a ional-
ize he di e ences in hei binding a ini ies. By employing a
combina ion o solu ion-s a e nuclea magne ic esonance (NMR)
spec oscopy and molecula dynamics (MD) simula ions, we s udy
he s uc u al and dynamic p ope ies o he SHP2 SH2 domains in
isola ion and in he p esence o ei he pY o pY-con aining pep ides
PD-1 ITIM and ITSM, and e eal a dynamic ne wo k ha couples
phospho y osine binding o SHP2 ac i a ion. We ind ha he cen-
al b-s ands bC and bD media e a signi ican coupling be ween
he pY-binding si e and he EF and BG loops o N-SH2. A simila
allos e ic coupling has p e iously been ound in an analysis o
he con o ma ional space o pep ide-bound N-SH2 [29,30]. He e
we demons a e ha pY-binding alone is esponsible o ac i a ing
his allos e ic ne wo k by inducing coupled dynamics o he BC, EF
and BG loops, as well as o he ip o bD p eceding he DE loop.
Some con o ma ional s a es induced by pY-binding dis a o he
in e ac ion o he N-SH2 domain wi h he PTP domain wi hin
SHP2, indica ing ha he pY-binding componen o he phospho-
pep ide in e ac ion wi h N-SH2 is p incipally esponsible o
inc easing he ac ional popula ion o open, ca aly ically ac i e
SHP2. The coupling be ween he pY-binding si e and he EF loop
is weake in C-SH2 han in N-SH2, in ag eemen wi h hei di e -
en oles in SHP2 ac i a ion.
2. Expe imen al p ocedu es
2.1. P o ein exp ession and pu i ica ion
The DNA sequences o N-SH2 (SHP2
1–105
) and C-SH2
(SHP2
106–220
) we e cloned in o he pETM22 exp ession ec o
(EMBL Collec ion) and exp essed as usion p o eins wi h a
clea able His
6
- hio edoxin ag. The ‘‘supe binde ” N-SH2 cons uc
(N-SH2 N37V/T42A/K55V) was p oduced wi h he QuikChange
Ligh ning Si e-Di ec ed Mu agenesis p o ocol (Agilen Technologies).
All cons uc s we e con i med by DNA sequencing.
Recombinan p o ein exp ession was pe o med in Tune (DE3)
cells (Me ck): a e ans o ma ion, he cells we e g own a 37 °Cin
2-li e lasks (250 pm) un il an op ical densi y (OD
600
) o 0.6–0.8
was eached. A e wa ds, he cul u es we e quickly cooled in a
wa e –ice mix, and p o ein exp ession was induced wi h 0.1 mM
isop opyl b-d-1- hiogalac opy anoside (IPTG) and con inued o
20 h a 20 °C. P epa a ion o unlabeled samples ( o ITC) was
achie ed by g owing he cells in Lu ia-Be ani b o h (LB), whe eas
he samples ha equi ed iso opic en ichmen o NMR we e p e-
pa ed by g owing he bac e ia in minimal medium con aining
15
-
NH
4
Cl (1 g/l, Camb idge Iso ope Labo a o ies) and
13
C-D-glucose
(4 g/l, Camb idge Iso ope Labo a o ies) as he sole ni ogen and
ca bon sou ces, espec i ely. Once p o ein exp ession was com-
ple e, he cul u es we e ha es ed, pelle ed and ozen a –20 °C
un il u he use.
P o ein pu i ica ion began wi h he lysis o he cells by sonica-
ion in lysis bu e (1 M NaCl, 50 mM T is-HCl, 5% glyce ol, 10 mM
imidazole, 5 mM b-me cap oe hanol, pH 7.6) supplemen ed wi h
one able o EDTA- ee cOmple e
TM
p o ease inhibi o s (Roche),
100 mg o lysozyme (Ro h) and 50 mg o DNAse (NEB). A e cen-
i uga ion (18000 g o one hou a 4 °C) and il a ion h ough a
s e ile 0.2-mm sy inge il e , he supe na an was loaded on a
HisT ap HP column (GE Heal hca e), p e iously equilib a ed wi h
wash bu e (1 M NaCl, 50 mM T is-HCl, 5% glyce ol, 10 mM imida-
zole, 5 mM b-me cap oe hanol, pH 7.6). A e loading, he column
was washed wi h 10 column- olumes (CV) o wash bu e and he
His- agged p o ein was elu ed wi h 5 CV o elu ion bu e (1 M
NaCl, 50 mM T is-HCl, 5% glyce ol, 500 mM imidazole, 5 mM b-
me cap oe hanol, pH 7.6). The His
6
- hio edoxin ag was emo ed
by incuba ing he elua e wi h 3C p o ease (1:100 p o ease:p o ein
a io) a 4 °C o e nigh , while dialyzing agains 2 L o wash bu e .
The ollowing day, ano he HisT ap a ini y ch oma og aphy was
pe o med and he p o ein o in e es was e ie ed in he low-
h ough, which was hen concen a ed o a inal olume o 1–
2 ml and loaded on a HiLoad 16/600 Supe dex 75 pg column (GE
Heal hca e), p e iously equilib a ed wi h NMR bu e (100 mM
MES, 150 mM NaCl, 3 mM TCEP, 0.05% NaN
3
, pH 6.8). The ac ions
con aining he pu e p o ein we e pooled, he p o ein was concen-
a ed o he desi ed concen a ion and ei he used immedia ely o
lash- ozen wi h liquid ni ogen o long- e m s o age a –80 °C.
2.2. Ligands
The pep ides used in his s udy (PD-1 ITIM: FSVDpYGELDFQ,
PD-1 ITSM: EQTEpYATIVFP, IRS1-pY1172: GLNpYIDLDLV,
IRS1-pY1222: LSTpYASINFQ, SHP2-pY542: KRKGHEpYTNIKYS,
SHP2-pY580: EDSARVpYENVGLE) we e pu chased om Caslo ApS
(Lyngby, Denma k). All o hem we e syn hesized wi h
N- e minal ace yla ion and C- e minal amida ion. S ock solu ions
we e p epa ed by dissol ing he lyophilized pep ide in he app o-
p ia e bu e o a nominal concen a ion o 5 mM. I necessa y, he
pH was co ec ed by addi ion o small olumes o 10 M NaOH.
Small aliquo s o 20
l
L we e p epa ed and s o ed a –20 °C un il
u he use. In all cases, pep ide concen a ion in he s ock solu-
ions was con i med by in eg a ion o app op ia e me hyl-g oup
p o on peaks by NMR. Phospho y osine (O-Phospho-L- y osine)
was pu chased om To on o Resea ch Chemicals (ca alog numbe
P365000).
2.3. Iso he mal i a ion calo ime y (ITC)
ITC expe imen s we e ca ied ou on a Nano ITC machine (TA
Ins umen s) in 100 mM MES (pH 6.8), 150 mM NaCl a 25 °C. Pep-
ides (1–1.5 mM) we e i a ed on o p o eins (100–200
l
M) in a
300-
l
L sample cell while s i ing a 200 pm. The da a (Figu e S2)
we e analyzed wi h NanoAnalyze ( 3.6.0, T
A
Ins umen s
)
and he
K
D
alues we e calcula ed om a single eplica e using a single-
binding si e model.
M. Ma asco, J. Ki kpa ick, V. Nanna e al. Compu a ional and S uc u al Bio echnology Jou nal 19 (2021) 2398–2415
2400
2.4. NMR spec oscopy
NMR expe imen s we e eco ded on B uke A ance III HD spec-
ome e s a
1
H ield-s eng hs o 600 MHz and 850 MHz, equipped
wi h c yogenic in e se HCN p obe-heads (N
2
- and He-cooled,
espec i ely) and unning Topspin 3.2 so wa e. All expe imen s
we e eco ded a a empe a u e o 298 K. NMR samples we e p e-
pa ed using uni o mly-
15
N-labelled p o ein a concen a ions o
0.3–0.6 mM in NMR bu e (100 mM MES, 150 mM NaCl, 3 mM
TCEP, 0.05% NaN
3
, pH 6.8).
2D
1
H-
15
N co ela ion maps we e eco ded as
1
H,
15
N-HSQC
spec a, acqui ed wi h S a es-TPPI o equency-disc imina ion
and wi h wa e supp ession achie ed ia a combina ion o WATER-
GATE and wa e lip-back pulses o p ese e he wa e magne iza-
ion [31–33].
In all i a ion expe imen s, he p o ein concen a ion was
200 mM and he amide chemical shi s o U-
15
N-labeled N-SH2 o
C-SH2 we e moni o ed by acqui ing a se ies o
1
H,
15
N-HSQC spec-
a in he p esence o inc easing s oichiome ic a ios o pep ide.
Chemical-shi pe u ba ions (CSPs) we e calcula ed acco ding o:
CSP ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
2
D
2
H
þ0:15
D
2
N

whe e
D
d
H
and
D
d
N
deno e he chemical-shi di e ences in he
1
H
and
15
N dimensions, espec i ely.
Backbone-amide
15
N elaxa ion a es (R
1
and R
1
q
) we e mea-
su ed a a
1
H ield-s eng h o 600 MHz using es ablished
p o on-de ec ed expe imen s based on a g adien -selec ed,
sensi i i y-enhanced, e ocused
1
H,
15
N-HSQC sequence [34–37].
The wa e signal was p ese ed using lip-back pulses and weak
bipola g adien s du ing he indi ec chemical shi e olu ion ime
o main ain he wa e magne iza ion along he +zaxis. The elax-
a ion delays we e a ied be ween 20 ms and 1.2 s o he R
1
expe -
imen , and be ween 3 ms and 100 ms o he R
1
q
expe imen . In he
R
1
sequence, N–H c oss- elaxa ion pa hways we e supp essed by
applica ion o
1
H amide-selec i e IBURP-1 in e sion pulses [38]
a in e als o 10 ms du ing he elaxa ion delay. In he R
1
q
sequence, c oss- elaxa ion was supp essed by applica ion o
be ween one and ou
1
H amide-selec i e in e sion pulses du ing
he
15
N spin-lock elaxa ion pe iod [39] and he
15
N magne iza ion
was explici ly aligned wi h he spin-lock ield [40], which was
applied a a ield-s eng h o 2.5 kHz. Backbone-amide {
1
H}
15
N
s eady-s a e he e onuclea NOEs we e measu ed using he s an-
da d me hod [41,42]. Wa e magne iza ion was p ese ed in he
e e ence spec um as desc ibed abo e. Sa u a ion o he amide
p o on magne iza ion was achie ed using a 5-s ain o high-
powe 180°pulses applied a 10.9-ms in e als [43]. The e e ence
and sa u a ed spec a we e eco ded in an in e lea ed ashion,
using a long ecycle delay o 15 s o ensu e ull eco e y o he
wa e magne iza ion a he s a o each inc emen o he e e ence
expe imen .
Backbone-amide
15
N elaxa ion-dispe sion p o iles we e mea-
su ed wi h he
15
N CPMG sequence o Hansen e al.[44] a s a ic
1
H ield-s eng hs o 600 MHz and 850 MHz (Figu es S3&S4). Du -
ing he CPMG pe iod, ans e se
15
N magne iza ion is main ained
as pu e in-phase cohe ence ia he applica ion o high-powe
1
H
con inuous-wa e (CW) decoupling, in which he exac
1
HCW
ield-s eng h is a ied sligh ly acco ding o he
15
N CPMG ield-
s eng h so ha
m
(
1
H-CW) = 2k
m
CPMG
o all
m
CPMG
(kin ege ).
The pulse-sequence was implemen ed wi h bo h
1
H and
15
N
empe a u e-compensa ion elemen s so ha RF hea ing e ec s
we e cons an o all
m
CPMG
ield-s eng hs and o he e e ence
plane. Wi h he excep ion o he da a o he N-SH2 (apo) sample
a 600 MHz, all CPMG da a-se s we e measu ed using a cons an -
ime delay o 50 ms, and he CPMG ield-s eng h was a ied
be ween 20 and 1200 Hz. Fo he N-SH2 (apo) sample a
600 MHz, he cons an - ime delay was 40 ms, and he CPMG
ield-s eng h was a ied be ween 25 and 1000 Hz.
All NMR spec a we e p ocessed/ isualized wi h a combina ion
o NMRPipe 10.1 [45] and CcpNm Analysis 2.4 [46]. Relaxa ion
spec a we e p ocessed wi h pa ial Lo en zian- o-Gaussian
apodiza ion in bo h equency dimensions and limi ed linea -
p edic ion in he
15
N dimension. Peak in ensi ies we e quan i ied
by lineshape- i ing wi h FuDA [47].
Backbone-amide R
1
and R
1
q
a es and {
1
H}
15
N s eady-s a e
he e onuclea NOEs we e analysed wi hin he model- ee ame-
wo k o Lipa i & Szabo [48,49] using he p og am TENSOR2 [50]
o i he elaxa ion da a o he model- ee spec al-densi y and
ex ac he global o a ional di usion enso , local o de -
pa ame e s (S
2
) and, whe e necessa y, local co ela ion imes.
The di e ences in he amide-speci ic con o ma ional en opies
be ween wo s a es A and B we e calcula ed acco ding o he
equa ion:
D
S¼k
B
log 1S
2
B
1S
2
A
!
which holds ei he o unco ela ed mo ions o indi idual ec o s o
o o de pa ame e s S
2
> 0.75 bo h in he p esence and in he
absence o co ela ed mo ions [51,52].
CPMG elaxa ion-dispe sion p o iles we e analysed using he
so wa e package ChemEx [53]. P o iles eco ded a bo h s a ic
ield-s eng hs we e i ed simul aneously on a esidue-by-
esidue basis o a simple wo-s a e exchange model wi h a iable
pa ame e s k
ex
(exchange- a e), p
b
(mino -s a e popula ion),
D
x
ab
( equency-di e ence be ween majo and mino s a es) and R
0
2
(exchange- ee ans e se elaxa ion a e) and using an {Îx,Îy,Îz}
spin-ope a o basis-se o in eg a ion o he spin-e olu ion du ing
he CPMG elemen . The o e all exchange con ibu ions, R

2;ex
, we e
ex ac ed om he back-calcula ed p o iles as desc ibed in he
main ex .
2.5. Homology modeling
The models o N-SH2–IRS1-pY1172, N-SH2–IRS1-pY1222,
C-SH2–IRS1-pY1172, C-SH2–IRS1-pY1222, N-SH2–SHP2-pY542,
N-SH2–SHP2-pY580, C-SH2–SHP2-pY542, C-SH2–SHP2-pY580
and C-SH2–ITIM we e gene a ed wi h Modelle , e sion 9.23
[54,55]. The empla e s uc u es used o he modeling we e
N-SH2–ITIM (PDB code 6ROY), N-SH2–ITSM (PDB code 6ROZ),
C-SH2–ITSM (PDB code 6R5G) and N-SH2–GAB1 (PDB code
4QSY). The oo -mean-squa e de ia ions (RMSDs) om he
empla e s uc u es a e gi en in Table S1.
2.6. Molecula dynamics
Molecula dynamics (MD) simula ions we e pe o med wi h
he GROMACS 2019.3 so wa e package [56,57], using he
AMBER99SB-ILDN pa ame e se [58]. The ini ial a omic coo di-
na es we e aken om c ys allog aphic (N-SH2: PDB code 1AYD,
N-SH2–pY: modi ied om PDB code 6ROZ) o NMR (C-SH2: PDB
code 2SHP, C-SH2–pY: modi ied om PDB code 6R5G) s uc u es.
The ‘‘supe binde ” N-SH2 domain (N-SH2 N37V/T42A/K55V) and
i s complex wi h pY we e gene a ed by in-silico mu agenesis wi h
PyMOL (The PyMOL Molecula G aphics Sys em, Ve sion 2.0
Sch ödinge , LLC). The opology and pa ame e s o he pY pep ide
(Ac-GpYG-NH
2
) we e ob ained om he GAFF o ce- ield [59] wi h
he ACPYPE se e [60]. Each p o ein o p o ein complex was
placed a he cen e o a dodecahed al box illed wi h TIP3P wa e
molecules [61], wi h a leas 1 nm o sol en on all sides. Long-
ange elec os a ic in e ac ions we e compu ed wi h he pa icle-
M. Ma asco, J. Ki kpa ick, V. Nanna e al. Compu a ional and S uc u al Bio echnology Jou nal 19 (2021) 2398–2415
2401
mesh Ewald (PME) scheme [62], while sho - ange non-bonded
in e ac ions we e unca ed a 1 nm. A e sol a ion and neu al-
iza ion wi h an app op ia e numbe o coun e -ions, each sys em
was minimized once by s eepes descen . A e wa ds, i e indepen-
den eplicas wi h di e en assignmen s o he ini ial eloci ies
we e p epa ed om he ene gy-minimized sys ems. Each eplica
unde wen wo 1-ns ounds o sol en equilib a ion (NVT and
NPT ensembles) o b ing i o he a ge empe a u e o 310 K
and a ge p essu e o 1 ba , while p o ein and pep ide a omic
posi ions we e es ained wi h a ha monic po en ial. P o ein
bond-leng hs in ol ing H we e cons ained wi h LINCS [63], while
he bond-leng hs and bond-angles o wa e molecules we e con-
s ained wi h SETTLE [64]. Tempe a u e coupling was achie ed
by eloci y escaling wi h a s ochas ic e m [65] and, when neces-
sa y, he p essu e was ixed wi h he Pa inello-Rahman ba os a
[66]. A e sol en elaxa ion, each eplica unde wen a 250-ns
un o p oduc i e MD simula ion in he NPT ensemble. A e
emo al o he ini ial 5 ns, he i e ajec o ies we e me ged o
yield an agg ega ed 1.225-ms ajec o y o analysis. Equilib a ion
o he s uc u es was comple e a e 5 ns (Figu e S5).
F ee-ene gy landscapes (FEL) we e calcula ed using he GRO-
MACS ool ‘‘sham”, which gene a es Gibbs ee ene gy landscapes
by Bol zmann in e sion o mul i-dimensional dis ibu ion his-
og ams o he quan i ies o in e es .
2.7. Co ela ion ne wo k analysis
The Bio3D 2.3 package [67] was used o he analysis o co e-
la ed mo ions. Co ela ions o a omic displacemen we e calcu-
la ed o each C
a
pai as Linea Mu ual In o ma ion (LMI) alues
and s o ed in o i e di e en ma ices (one pe eplica, a e
emo al o he ini ial 50 ns), acco ding o he ollowing equa ion:
LMI x
i
;x
j

¼0:5 ln de C
i
ðÞþln de C
j

ln de C
ij

whe e C
i
and C
ij
a e he ma ginal-co a iance ma ices o he displacemen s
o C
a
a om iand o C
a
a oms i&j, espec i ely. 50 ns we e
emo ed ins ead o only he 5 ns necessa y o equilib a ion due
o ha dwa e limi a ions in handling he long ajec o ies. A consen-
sus ma ix con aining a e age LMI alues was buil om he i e
indi idual LMI ma ices, in which ze o alues we e assigned o a
C
a
pai i hei LMI was no g ea e han 0.5 o all eplicas and/
o i hey we e sepa a ed by mo e han 10 Å in a leas 70% o
cumula i e ajec o y ames; applica ion o his il e equi ed
p e ious calcula ion o a con ac map ha s o ed all in e -C
a
pai wise dis ances [68,69]. In he cons uc ion o he co ela ion
ne wo k, he nodes ep esen C
a
a oms and node-pai s a e con-
nec ed h ough edges ha a e weigh ed by minus he loga i hm
o hei LMI alues. Communi y analysis, node cen ali y and
sub-op imal pa h calcula ions we e pe o med wi h Bio3D.
Be weenness cen ali y (de ined as he numbe o unique sho es
pa hs c ossing a node) was used as he measu e o node cen ali y.
The pa ame e s o he analysis o sub-op imal pa hs we e he
sou ce and sink nodes, as well as he o al numbe o pa hs o be
calcula ed, which was se o 250.
3. Resul s
3.1. Binding speci ici y: A compa ison o he esidues C- e minal o pY
The sequence de e minan s o he a ini y o phosphopep ides
o SH2 domains ha e been he subjec o nume ous s udies [70–
74]. He e, we use NMR spec oscopy, a ailable s uc u es and
homology modelling o a ionalize he di e ences amongs a se
o six phosphopep ides ha bind wi h di e en a ini ies o SHP2
N-SH2 and C-SH2 domains. Fou o hese phosphopep ides a e
known ac i a o s o SHP2: PD-1 ITIM and ITSM, and IRS1-pY1172
and -pY1222 (Fig. 1C); he o he wo comp ise he SHP2 C-
e minal pY-mo i s, con aining pY542 and pY580, whose unc ion
is s ill con o e sial.
The a ini ies o ITIM, ITSM, IRS1-pY1172 and IRS1-pY1222 o
SHP2 N-SH2 and C-SH2 had been de e mined p e iously, as had
he s uc u es o N-SH2–ITIM, N-SH2–ITSM and C-SH2–ITSM
[17,19]. He e, we gene a ed homology models o he complexes
N-SH2–IRS1-pY1172, N-SH2–IRS1-pY1222, C-SH2–IRS1-pY1172,
C-SH2–IRS1-pY1222 and C-SH2–ITIM, as desc ibed in Me hods.
IRS1-pY1172 is he s onges binde o N-SH2, ollowed by ITSM
(wi h one o de -o -magni ude lowe a ini y) and hen ITIM and
IRS1-pY1222 (wi h wo o de s-o -magni ude lowe a ini y). NMR
spec a indica e ha nei he SHP2-pY542 no SHP2-pY580 bind
signi ican ly o he N-SH2 domain (Fig. 1C), hus con adic ing
he hypo hesis ha SHP2 could be au o-ac i a ed by i s C-
e minal ail.
C-SH2 has he highes a ini y o ITSM and IRS1-pY1222, which
bo h ma ch he consensus binding sequence, ollowed by IRS1-
pY1172 and ITIM wi h a ini ies in he mic omola ange. The
NMR spec a demons a e no binding o C-SH2 o SHP2-pY580,
while SHP2-pY542 bound o CSH2 wi h a dissocia ion cons an
K
D
o 28 mM, as de e mined by ITC (Fig. 1C and Figu e S2).
Bo h N-SH2 and C-SH2 in e ac wi h phosphopep ides in a wo-
p onged manne , wi h one binding pocke , delimi ed by bB, bC, bD,
a
A and he BC loop, hos ing he pY esidue, and a second
hyd ophobic pocke , delimi ed by CD, DE, EF, BG, bD and
a
B,
accommoda ing esidues C- e minal o pY. The a ini y o he pep-
ides o N-SH2 co ela es s ongly wi h he leng h o he side-
chain in posi ion pY+1, which poin s owa ds esidues T52, I54
and I96 o he second binding pocke . The homology model o
he N-SH2–IRS1-pY1172 complex shows ha he long alipha ic
side-chain o I(+1) (Fig. 2A) makes a mul i ude o an de Waals
con ac s wi h hese esidues; ewe con ac s can be made by he
sho e A(+1) side-chain o ITSM and IRS1-pY1222 (Fig. 2B), while
no in e ac ions occu wi h ITIM G(+1) (Fig. 2C). The T(+1) and E(+1)
esidues o SHP2-pY542 and SHP2-pY580, espec i ely, do no
complemen he hyd ophobic cha ac e o he pocke . Thus, o
N-SH2, he a ini y o he phosphopep ides es ed he e co ela es
wi h he na u e o he pY+1 amino acid.
Despi e he conse a ion be ween he wo SH2 domains o he
hyd ophobic esidues in he egion con ac ing he pY+1 amino acid
(N-SH2/C-SH2: I54/V171; I96/L210), he a ini ies o he pep ides o
C-SH2 do no show he same co ela ion wi h he leng h o he pY+1
side-chain. One possible explana ion is ha he hyd ophobici y o
his egion in C-SH2 is educed by he p esence o E204, whose
side-chain pa ially shields L210 (Fig. 2D). In N-SH2 he same
side-chain (E90/E204 in N-SH2/C-SH2) is mo ed away om I96 by
a sal -b idge wi h R47 (Fig. 2E), which canno be o med in C-SH2
(R47/T153 in N-SH2/C-SH2). We p opose ha he educed
hyd ophobic cha ac e o his p o ein egion lowe s he dependency
o he binding ee-ene gy on he na u e o he pY+1 side-chain.
The side-chain o pY+2 is o ien ed owa ds he sol en and
would no appea o make signi ican con ibu ions o he ee
ene gy o binding. Howe e , molecula dynamics simula ions ha e
shown ha an acidic amino acid a his posi ion, as o example he
aspa ic acid o he s onges N-SH2 binde IRS1-pY1172, can make
a o able elec os a ic in e ac ions wi h K89 and K91 o N-SH2
[30]. Bo h SHP2-pY542 and SHP2-pY580 ha e an aspa agine in
posi ion pY+2, which is equi ed o binding o he SH2 domain
o G b2 wi h a con o ma ion di e en om he ex ended one
[75]. Thus, bo h he lack o a s ong in e ac ion wi h ei he N-
SH2 o C-SH2 o SHP-2 and he p esence o N(+2) sugges ha
he phospho yla ed mo i s o he C- e minal ail o SHP2 unc ion
mainly o ec ui G b2.
A hyd ophobic alipha ic esidue a pY+3 is equi ed o binding
o he SH2 domains o SHP2 (and o mos SH2 domains in gene al)
M. Ma asco, J. Ki kpa ick, V. Nanna e al. Compu a ional and S uc u al Bio echnology Jou nal 19 (2021) 2398–2415
2402

[5,7,76]. In ag eemen , all ou binding phosphopep ides ha e an
isoleucine (IRS1-pY1222 and ITSM) o leucine (IRS1-pY1172 and
ITIM) esidue a his posi ion. In all complexes, his esidue is
encapsula ed in a pen agonal-shaped a ay o hyd ophobic
esidues, namely I54/V170, L65/V181, Y81/Y197, L88/M202 and
I96/L210 in N-SH2/C-SH2 (Fig. 2F).
The ole o he esidue pY+4 is sub le: i s side-chain poin s
owa ds he sol en , bu is e y close o he BG loop, wi h which
i could in e ac . The ITIM mu a ion D(+4)A causes ei he he asso-
cia ion o he dissocia ion a e o he N-SH2–ITIM complex o
inc ease, as demons a ed by he beha io o he N-SH2
1
H-
15
N
NMR peaks upon i a ion o he pep ide ( ansi ion om he
slow-exchange egime in he p esence o ITIM o he
in e media e-exchange egime in he p esence o he mu an ITIM,
Fig. 3B). The same mu a ion educes a ini y o C-SH2, as demon-
s a ed by he inc ease in he pep ide:p o ein mola a io necessa y
o each sa u a ion o C-SH2 om 1:1 o 2:1 (Fig. 3B). In complex
wi h N-SH2, D(+4) may pa icipa e in ansien elec os a ic in e -
ac ions wi h K89 and K91 o he BG loop (Fig. 3C). Howe e , in
complex wi h C-SH2, hese elec os a ic in e ac ions a e no possi-
ble (K89/V203 and K91/T205 in N-SH2/C-SH2) and he de imen al
e ec o he D(+4)A mu a ion canno be a ionalized in e ms o
binding en halpy. Ne e heless, he hyd ophobic cha ac e o he
BG loop o C-SH2 ma ches he na u e o V(+4) in he s onges
binding pep ide ITSM (Fig. 3D), while IRS1-pY1222 wi h N(+4)
binds less s ongly.
The in e ac ion wi h esidue pY+5 is pa icula o he SH2
domains o SHP2: his esidue is in a iably hyd ophobic and e y
o en a oma ic and in e ac s wi h he same hyd ophobic pa ch ha
su ounds pY+3 (Fig. 2F). The p esence o a non-hyd ophobic
amino acid a his posi ion is s ongly un a o able, as demon-
s a ed by he educ ion in a ini y by wo o de s o magni ude
when compa ing he VSPEPIpYATIDDL pep ide wi h he ASPEPI-
pYATIDFD pep ide [77].
The con ibu ion o esidues beyond pY+5 is negligible: bo h he
ITIM mu a ion Q(+6)A and ex ension o ITIM by wo esidues, W
Fig. 2. Recogni ion o phosphopep ide sidechains a posi ions pY+1, pY+3 and pY+5. (A, B, C) In e ac ions be ween he hyd ophobic clus e o N-SH2, o med by I54, L88 and
I96, and he pY+1 esidue o IRS1-pY1172, ITSM and ITIM, espec i ely. (D) Same egion o he p o ein as in panels A–C, bu o he C-SH2–ITSM complex. E204 in C-SH2
canno o m a sal -b idge wi h T153 ( o which he co esponding esidue in N-SH2 is R47) and hus shields L210 om in e ac ing wi h he pep ide esidue pY+1. (E) Same as
panel B (N-SH2–ITIM complex), highligh ing he sal -b idge be ween E90 and R47, which exposes I96 o he ITSM esidue pY+1. (F) The pY+5 binding si e o N-SH2 bound o
ITSM consis s o a pen agonal a angemen o i e hyd ophobic esidues (I54, L65, Y81, L88 & I96), which also con ac he side-chain o he pY+3 esidue.
M. Ma asco, J. Ki kpa ick, V. Nanna e al. Compu a ional and S uc u al Bio echnology Jou nal 19 (2021) 2398–2415
2403
(+7) and R(+8), cause only mino local changes in he
1
H-
15
N NMR
spec um o bound N-SH2 in compa ison wi h he spec um o N-
SH2 in complex wi h wild- ype ITIM: he di e ences a e localized
o he BG-loop esidues K89 and K91 (Fig. 3E and 3F).
3.2. In luence o ITIM and ITSM binding on he dynamics o SHP2 SH2
domains
To es ablish whe he he binding o ei he ITIM o ITSM a ec s
he dynamics o he SHP2 SH2 domains di e en ly, and hus e eal
po en ial en opic con ibu ions o he binding ene gy, we mea-
su ed as (sub-ns) and slow (200-
l
s–20-ms) ime-scale dynamics
by NMR spec oscopy o bo h he N-SH2 and C-SH2 domains o
SHP2 in hei apo, ITIM-bound and ITSM-bound o ms.
15
NR
1
and R
2
elaxa ion a es and {
1
H}
15
N he e onuclea NOEs
we e measu ed o he backbone amide g oups o N-SH2, N-SH2–
ITIM, N-SH2–ITSM, C-SH2 and C-SH2–ITSM and i ed wi h he
Lipa i-Szabo model- ee o malism, as desc ibed in Me hods. F om
his analysis we ex ac ed he global o a ional co ela ion ime
s
c
,
he deg ee o aniso opy o he o a ional di usion enso , and he
S
2
o de -pa ame e o each amide g oup, which is a measu e o
he magni ude o p o ein in e nal mo ions on he as ime-scale
(S
2
can a y be ween 0 and 1; la ge in e nal mo ions co espond
o smalle alues o S
2
). In o ma ion on some o he esidues o
he apo N-SH2 BC and BG loops was missing because he co e-
sponding peaks we e e y weak ( ide in a). The unbound, ITIM-
bound and ITSM-bound N-SH2 and C-SH2 domains can be
desc ibed as globula and app oxima ely sphe ical en i ies wi h
o a ional co ela ion imes compa ible wi h hose p edic ed om
he monome ic s a e (Table S2). In he absence o phosphopep ide,
bo h N-SH2 and C-SH2 showed only limi ed in e nal mo ion on he
as ime-scale, wi h he excep ion o he long CD loop o C-SH2
(Fig. 4). The a e age S
2
o de -pa ame e o C-SH2 (0.81) was lowe
han ha o N-SH2 (0.86). Addi ion o phosphopep ide did no sig-
ni ican ly pe u b he o e all backbone dynamics (Fig. 4), (a e age
alues o he S
2
o de -pa ame e : N-SH2–ITIM, 0.88; N-SH2–ITSM,
0.87; C-SH2–ITSM, 0.81). Ne e heless, a ew sub le di e ences
we e obse ed in unc ionally ele an a eas o he wo SH2
domains. To be e unde s and hese changes, we con e ed he
di e ences be ween he S
2
alues o he phosphopep ide-bound
and unbound s a es o indi idual esidues o he co esponding
changes in con o ma ional en opy [52]. Despi e neglec ing
Fig. 3. In luence o phosphopep ide side-chains a posi ions pY+4 and pY+6 on binding a ini ies. (A) Sequences o he ITIM mu an s used o s udy he ole o amino acids a
posi ion pY+4 and beyond pY+5. (B)
1
H,
15
N-HSQC spec a o N-SH2 and C-SH2 in he p esence o inc easing concen a ions o wild- ype and D(+4)A ITIM. (C) Al hough no
e iden om he c ys al s uc u e, D(+4) may o m ansien elec os a ic in e ac ions wi h K89 and/o K91 o N-SH2; hese in e ac ions would no be possible wi h he D(+4)
A mu a ion. (D) K89 and K91 in N-SH2 a e eplaced by V203 and T205 in C-SH2, which makes his a ea subs an ially mo e hyd ophobic, in ag eemen wi h he high a ini y o
phosphopep ides possessing a hyd ophobic esidue a pY+4. (E) O e lay o
1
H,
15
N-HSQC spec a o N-SH2 wi h a 2:1 pep ide:p o ein mola a io; he pep ide is ei he wild-
ype ITIM o Q(+6)A ITIM. (F) O e lay o
1
H,
15
N-HSQC spec a o N-SH2 wi h a 2:1 pep ide:p o ein mola a io; he pep ide is ei he wild- ype ITIM o mu an ITIM con aining
one (W) o wo (WR) addi ional amino acids a he C- e minus. The p o ein concen a ion was 200
l
M and all spec a we e measu ed on a 600 MHz spec ome e a 298 K.
M. Ma asco, J. Ki kpa ick, V. Nanna e al. Compu a ional and S uc u al Bio echnology Jou nal 19 (2021) 2398–2415
2404
con ibu ions om luc ua ions wi h co ela ion ime g ea e han
s
c
and en opy changes due o he ligand o he sol en , his anal-
ysis can e eal in e es ing esidue-speci ic e ec s. The binding o
ITIM and ITSM was associa ed wi h an o e all nega i e con o ma-
ional en opy change on he imescale p obed by he NMR mea-
su emen s (up o 50 ps). This global educ ion o as mo ions
was mo e p onounced o N-SH2 compa ed o C-SH2 and o ITIM
compa ed o ITSM:
D
S(N-SH2–ITIM s N-SH2) = 32.6 J mol
–1
K
1
;
Fig. 4. Fas dynamics o he N-SH2 and C-SH2 domains in hei apo s a es and in complexes wi h phosphopep ides. S
2
o de pa ame e s ex ac ed om he i o he
15
NR
1
,R
2
and he e onuclea NOE da a o he Lipa i-Szabo model- ee pa ame iza ion o as mo ions. A. S
2
o de pa ame e s o N-SH2, N-SH2–ITIM and N-SH2–ITSM. B. S
2
o de
pa ame e s o C-SH2 and C-SH2–ITSM. The esidue numbe s a e gi en on he x-axes. Seconda y-s uc u e elemen s a e indica ed along he sequence.
M. Ma asco, J. Ki kpa ick, V. Nanna e al. Compu a ional and S uc u al Bio echnology Jou nal 19 (2021) 2398–2415
2405
D
S(N-SH2–ITSM s N-SH2) = 19.5 J mol
–1
K
1
;
D
S(C-SH2–ITIM s
C-SH2) = 15.9 J mol
–1
K
1
;
D
S(C-SH2–ITSM s C-SH2) = 3.7 J mo
l
–1
K
1
(o e all en opy-changes we e calcula ed as a sum o he
esidue-speci ic en opy-changes, excluding esidues wi h
S
2
< 0.7). Locally, bo h ITIM and ITSM inc eased he con o ma ional
en opy o he N-SH2 EF loop, while ITSM dec eased ha o he
same loop in C-SH2 (Fig. 5A). Mo eo e , binding o bo h pep ides
igidi ied he BG loop o bo h domains (Fig. 5A). Upon inspec ion
o indi idual si es, we obse ed ha K89, in he BG loop o N-
SH2, showed an en opy inc ease in complex wi h ITSM bu no
wi h ITIM. A possible explana ion is ha in complex wi h ITSM,
he a ea a ound K89 is comple ely dehyd a ed, p esumably
because o he high hyd ophobici y o he pep ide (PDB code
6ROZ); in con as , in complex wi h ITIM (PDB code 6ROY), K89
coo dina es a wa e molecule, hus becoming less mobile
(Fig. 5B). In conclusion, he as dynamics o he N-SH2– and C-
SH2–phosphopep ide complexes show some pep ide- and
domain-speci ic e ec s, which can con ibu e o sub le di e ences
in he binding a ini ies.
Nex , we p obed he dynamics in he high-
l
s–ms imescale,
measu ing ans e se elaxa ion dispe sion expe imen s on bo h
isola ed N-SH2 and C-SH2 and hei complexes wi h ITIM and
ITSM. We used a
15
N Ca l-Pu cell-Meiboom-Gill (CPMG) expe i-
men , eco ded a bo h 600 and 850 MHz, wi h CPMG ield-
s eng hs a ying be ween 20 and 1200 Hz applied o a
cons an - ime pe iod o 50 ms. Unde hese condi ions, con o ma-
ional exchange dynamics can be measu ed wi h exchange a es,
k
ex,
in he ange ~50–5000 Hz. The expe imen al elaxa ion dispe -
sion p o iles o indi idual esidues we e i o a wo-si e exchange
model using ChemEx [53]. The exchange con ibu ions o he
ans e se elaxa ion a es, R

2;ex
, we e hen calcula ed as he di e -
ence be ween he i ed ans e se elaxa ion a es a he mini-
mum and maximum CPMG ield-s eng hs
(R

2;ex
¼R
2
m
min
CPMG

R
2
m
max
CPMG

). The alue o he R

2;ex
pa ame e
can be oughly co ela ed wi h he ex en o slow ime-scale
dynamics a each amino-acid posi ion.
In apo N-SH2, highe han a e age R

2;ex
alues we e measu ed
o he DE and EF loops and o he ip o bD be o e he DE loop.
In he BG loop, wo amino acids showed enhanced dynamics,
oge he wi h he ip o he
a
B helix be o e he loop (Fig. 5C1
and Supplemen a y Figu e S3). In he
1
H-
15
N co ela ion spec a,
a small numbe o amide g oups o he BC and BG loops we e
absen , p obably due o as exchange o he amide p o on wi h
he sol en , and hei elaxa ion dispe sion could no be measu ed.
O e all, apo N-SH2 showed enhanced mobili y on he high-
l
s–ms
imescale in egions ha bind he pep ide esidues C- e minal o
he phospho y osine. Binding o bo h ITIM and ITSM diminished
he dynamics in bo h he DE and EF loops and he lanking bD.
The slow dynamics o he ip o
a
B was quenched in bo h he N-
SH2–ITIM and N-SH2–ITSM complexes, while slow dynamics was
induced by bo h ITIM and ITSM in he
89
KE
91
K s e ch (Fig. 5C1
and Supplemen a y Figu e S3). In addi ion, ITIM and ITSM had di -
e en e ec s on he exchange-con ibu ions o H85 and L88, which
we e highe in he N-SH2–ITSM han in he N-SH2–ITIM complex.
By con as , he inc ease o he slow dynamics o he
89
KE
91
K
s e ch was mo e p ominen in N-SH2–ITIM han in N-SH2–ITSM.
As o he as dynamics, hese a iable e ec s may be due o he
di e en hyd ophobici ies o he C- e minal hal es o ITIM and
ITSM.
Simila o he N-SH2 domain, exchange con ibu ions o he
15
N
ans e se elaxa ion a es o he C-SH2 domain we e qui e small
o e all. The peaks o a small numbe o amide g oups o he CD
loop we e absen , again p obably due o as exchange o he amide
p o on wi h he sol en , and hei elaxa ion dispe sion could no
be measu ed. In he es o he p o ein, he N- e minal ail, he
DE and he EF loops and he BG loop oge he wi h he C-
e minus showed enhanced dynamics (Fig. 5C2 and Supplemen a y
Figu e S4). The dynamics o he BG loop and he C- e minal pa o
he p o ein we e quenched by ITSM binding (Fig. 5C2 and Supple-
men a y Figu e S4), in ag eemen wi h a s abiliza ion o his egion
due o he in e ac ion wi h he C- e minal hal o he pep ide and
in quali a i e ag eemen wi h he as dynamics da a (Fig. 5A). The
dynamics o he DE and EF loops we e nea ly unchanged in he
p esence o he pep ide, wi h he excep ion o V181 and G182,
which a e in di ec p oximi y o I(+3) in he hyd ophobic speci-
ici y pocke and we e pa ially igidi ied by pep ide binding, again
in ag eemen wi h he as dynamics da a (Fig. 5A).
In conclusion, pep ide binding conside ably educes he slow
mo ions o he DE and EF loops o apo N-SH2, while i has a milde
e ec on he less mobile DE and EF loops o apo C-SH2. By con as ,
he slow dynamics o he apo C-SH2 BG loop is quenched by pep-
ide binding, while ha o he N-SH2 BG loop is changed in a
esidue-speci ic manne . These dis inc beha io s o he N-SH2
and C-SH2 domains may con ibu e o modula e hei in e ac ions
wi h phosphopep ides in di e en manne s, acco ding o hei di -
e en unc ions.
3.3. A ne wo k o s uc u al in e ac ions connec s he pY-binding si e
wi h emo e s uc u al elemen s
Nex , we asked whe he pY in he phosphopep ide has a ole in
inducing he s uc u al changes o N-SH2 ha a e esponsible o
i s educed a ini y o he PTP domain, o whe he i unc ions
solely o inc ease a ini y. To add ess his ques ion, we i a ed
pY on o he SHP2 N-SH2 domain and ollowed he CSPs o he
1
H-
15
N peaks (Fig. 6A). pY binds in he as -exchange egime wi h
aK
D
o 565 mM, as calcula ed by quan i ica ion o he NMR CSPs
(Fig. 6B). Binding o pY a ec ed he chemical shi s o esidues
loca ed a om he pY-binding si e, such as hose in he EF and
DE loops (Fig. 6C). This obse a ion sugges s he p esence o an
allos e ic ne wo k, which p opaga es he s uc u al changes
induced in he BC loop by pY-binding o dis an egions o he N-
SH2 domain.
To u he con i m he exis ence o his allos e ic ne wo k, we
gene a ed he N-SH2 binding-si e mu an N37V/T42A/K55V, which
was demons a ed o bind pY wi h inc eased a ini y (pY ‘‘su-
pe binde ”, SB N-SH2) [78]. NMR i a ions con i m ha SB N-
SH2 binds pY wi h a lowe K
D
han wild- ype N-SH2 (K
D
= 160
l
M,
Fig. 4,Fig. 6B). Despi e he ac ha all he mu a ions we e local-
ized o he pY-binding si e, compa ison o he NMR spec a o
apo SB N-SH2 and apo wild- ype N-SH2 showed CSPs o dis an
esidues (Fig. 6D), con i ming he p esence o an allos e ic s uc-
u al ne wo k connec ing he pY-binding si e wi h o he s uc u al
elemen s o he N-SH2 domain. In e es ingly, SB N-SH2 could no
longe disc imina e be ween ITIM and ITSM and bound bo h pep-
ides wi h simila a ini ies (Fig. 6E). This inding suppo s he p e-
ious wo k o Kaneko e al. [78], who p oposed ha a ela i ely
weak a ini y o SH2 domains o pY alone is necessa y o disc im-
ina e be ween speci ic and non-speci ic binding pa ne s and a oid
c oss alk be ween un ela ed pa hways.
3.4. Molecula dynamics simula ions e eal he na u e o he allos e ic
ne wo k
To unde s and he na u e o he allos e ic ne wo k ha con-
nec s he pY-binding si e wi h emo e s uc u al elemen s o he
N-SH2 domain, we pe o med molecula dynamics (MD) simula-
ions o N-SH2, N-SH2–pY, SB N-SH2, SB N-SH2–pY, C-SH2 and
C-SH2–pY. A simila s udy has ecen ly been conduc ed on a pool
o phosphopep ide-bound con o ma ions o he SHP2 N-SH2
domain [29,30]. He e, we ocus on unde s anding he ole o pY-
M. Ma asco, J. Ki kpa ick, V. Nanna e al. Compu a ional and S uc u al Bio echnology Jou nal 19 (2021) 2398–2415
2406
in e ac ion su ace (consis ing o he EF and DE loop) and hus p e-
en a e-closing o he N-SH2 domain on he PTP. This esul s in a
slowe a e o closu e, k
close
, in he p esence o pY han in i s
absence. I is also likely ha pY-binding inc eases he a e o an-
si ion o he open con o ma ion, k
open
; howe e , ou MD simula-
ions ca ied ou wi h he isola ed N-SH2 domain gi e no
in o ma ion on he dynamics o his domain when i is connec ed
o he PTP domain. In addi ion o a o ing he open SHP2 con o -
ma ion, he mo ions induced by pY-binding p o ide access o di -
e en con o ma ions o he speci ici y pocke ha can be
selec ed by he C- e minal hal o he pep ide o ull binding
(Fig. 9). Once he C- e minal esidues o he pep ide a e locked in o
he speci ici y pocke , ou NMR expe imen al da a suppo a pic-
u e whe e he dynamics o he DE and EF loops a e quenched,
while he dynamics a longe ime-scales is p ese ed in he BG
loop (Fig. 4 and Fig. 5). A e he N-SH2–phosphopep ide complex
is comple ely o med, i s li e ime (1/k
pep ide-o
) de e mines he e i-
ciency o SHP2 ac i a ion.
The co ela ed dynamics ne wo k appea s o be pa icula o N-
SH2, since in C-SH2 he co ela ion be ween he mo ions o he pY
and speci ici y pocke s is ma kedly less p onounced. The s ingen
selec i i y and weakly-co ela ed dynamics o C-SH2 i well wi h
he no ion ha his domain mainly ac s o ec ui biden a e phos-
phopep ides o SHP2, while N-SH2, wi h i s mo e p omiscuous
binding and co ela ed dynamics, has e ol ed o couple phospho-
pep ide binding wi h egula ion o phospha ase ac i i y [17].
Using MD simula ions, Anselmi and Hub ind ha he s a e o N-
SH2 wi h unzipped bC- and bD-s ands (s a e
a
in hei wo k and
s a es
e
/din ou wo k) is be e a inducing he displacemen o
he N-SH2 domain om PTP [29]. They p opose a mechanism o
ac i a ion o SHP2 whe e he pep ide binds o he closed con o -
ma ion o he phospha ase, bu only pep ides binding o N-SH2
in he
a
s a e induce opening and ac i a ion o SHP2. This hypo h-
esis may seem p oblema ic, as s ong ac i a o s o he phos-
pha ase, such as PD-1 ITSM and PDGFR-1009, bind N-SH2
wi hou unzipping he bC- and bD-s ands (Figu e S12). Ou model
may econcile his disc epancy by p oposing, nex o he model o
Anselmi and Hub [29], a second ac i a ion pa hway (Fig. 9), based
on he ini ial ecogni ion o pY by he closed con o ma ion o SHP2
and ull pep ide binding o he open con o ma ion, whose popula-
ion is inc eased by he pY-binding. Once he pep ide is bound, he
c
s a e o N-SH2 is also incompa ible wi h he closed con o ma ion
o SHP2 because o s e ic clashes be ween he EF loop and he PTP
domain (Figu e S1C). Fu he mo e, he ac ha Anselmi and Hub
ound ha he
a
con o ma ion induces SHP2 opening suppo s
ou hypo hesis ha pY-binding inc eases he a e k
open
by p omo -
ing he ansi ion om he
c
o he
e
/ds a es, which a e simila o
he
a
s a e o Anselmi and Hub [29]. The wo mechanisms o SHP2
ac i a ion may coexis o be mu ually exclusi e, depending on he
pep ide sequence.
5. Conclusions
The N-SH2 domain o SHP2 di e s om many o he SH2
domains (including SHP2 C-SH2) because i s binding o ups eam
phosphopep ides is coupled o he egula ion o enzyma ic ac i i y.
In his con ex , phospho y osine- ecogni ion no only p o ides ee
ene gy o binding, bu also has signi ican e ec s on he global
s uc u e and in e nal dynamics o he domain. In he pY-bound
s a e, he pY-binding pocke and he speci ici y pocke expe ience
co ela ed luc ua ions, which co espond o wo p incipal con o -
ma ional s a es o N-SH2–pY; o hese wo s a es, only one can
accommoda e he phosphopep ide esidues C- e minal o pY+3.
The luc ua ions induced by pY-binding dis up he in e ac ion
in e ace wi h he PTP domain, hus suppo ing SHP2 ac i a ion.
The esul s desc ibed he e explain how high-a ini y monophos-
pho yla ed pep ides can bind he N-SH2 domain o au o-
inhibi ed SHP2, despi e he inaccessibili y o pa o he speci ici y
pocke when N-SH2 is engaged wi h he PTP domain.
In conclusion, we show ha , despi e hei e y simila old, SH2
domains ha e a ange o s uc u al and dynamical di e ences ha
suppo a a ie y o speci ici ies and unc ions. These p ope ies
canno be e ealed by X- ay s uc u es, as hey a e ela ed o he
allos e ic coupling o he dynamics o binding pocke s and s uc-
u al elemen s. Likely, he concep s in oduced he e o he SH2
domains o SHP2 also apply o o he SH2 domains in ol ed in
he egula ion o enzyma ic ac i i y (as, o example, hose o phos-
pholipase C
c
1 o S c- amily kinases) and a e hus o gene al ele-
ance o unde s anding he mechanisms o unc ion o SH2
domains.
Decla a ion o Compe ing In e es
The au ho s decla e ha hey ha e no known compe ing inan-
cial 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
This wo k was unded by he Ge man Science Founda ion DFG
(g an CA 294/20-1 o TC). MM was suppo ed by a ellowship om
he Hanno e School o Biomolecula D ug Resea ch (HSBDR) and
was a membe o he Hanno e Biomedical Resea ch School (HBRS)
and he MD/PhD p og am ‘‘Molecula Medicine”.
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.csbj.2021.04.040.
Re e ences
[1] Pawson T. Speci ici y in signal ansduc ion: om phospho y osine-SH2
domain in e ac ions o complex cellula sys ems. Cell. 2004;116(2):191-203.
Epub 2004/01/28. doi: 10.1016/s0092-8674(03)01077-8. PubMed PMID:
14744431.
[2] Hun e T. Ty osine phospho yla ion: hi y yea s and coun ing. Cu en
opinion in cell biology. 2009;21(2):140-6. Epub 2009/03/10. doi: 10.1016/j.
ceb.2009.01.028. PubMed PMID: 19269802; PubMed Cen al PMCID:
PMCPMC2670436.
[3] Liu BA, Jablonowski K, Raina M, A cé M, Pawson T, Nash PD. The human and
mouse complemen o SH2 domain p o eins-es ablishing he bounda ies o
phospho y osine signaling. Molecula cell. 2006;22(6):851-68. Epub 2006/06/
24. doi: 10.1016/j.molcel.2006.06.001. PubMed PMID: 16793553.
[4] Waksman G, Kuma an S, Lubman O. SH2 domains: ole, s uc u e and
implica ions o molecula medicine. Expe e iews in molecula medicine.
2004;6(3):1-18. Epub 2004/02/28. doi: 10.1017/s1462399404007331. PubMed
PMID: 14987415.
[5] Eck MJ, Shoelson SE, Ha ison SC. Recogni ion o a high-a ini y phospho y osyl
pep ide by he S c homology-2 domain o p56lck. Na u e 1993;362
(6415):87–91. h ps://doi.o g/10.1038/362087a0. PubMed PMID: 7680435.
[6] Songyang Z, Shoelson SE, Chaudhu i M, Gish G, Pawson T, Hase WG, e al. SH2
domains ecognize speci ic phosphopep ide sequences. Cell. 1993;72(5):767-
78. Epub 1993/03/12. doi: 10.1016/0092-8674(93)90404-e. PubMed PMID:
7680959.
[7] Huang H, Li L, Wu C, Schibli D, Colwill K, Ma S, e al. De ining he Speci ici y
Space o he Human S c Homology 2 Domain*. Mol Cell P o eomics 2008;7
(4):768–84. h ps://doi.o g/10.1074/mcp.M700312-MCP200.
[8] Ladbu y JE, A old ST. Ene ge ics o S c homology domain in e ac ions in
ecep o y osine kinase-media ed signaling. Me hods in enzymology.
2011;488:147-83. Epub 2011/01/05. doi: 10.1016/b978-0-12-381268-
1.00007-0. PubMed PMID: 21195228.
[9] Neel BG, Gu H, Pao L. The ’Shp’ing news: SH2 domain-con aining y osine
phospha ases in cell signaling. T ends in biochemical sciences. 2003;28
(6):284-93. Epub 2003/06/27. doi: 10.1016/s0968-0004(03)00091-4. PubMed
PMID: 12826400.
[10] Ho P, Pluskey S, Dhe-Paganon S, Eck MJ, Shoelson SE. C ys al s uc u e o he
y osine phospha ase SHP-2. Cell. 1998;92(4):441-50. Epub 1998/03/10. doi:
10.1016/s0092-8674(00)80938-1. PubMed PMID: 9491886.
M. Ma asco, J. Ki kpa ick, V. Nanna e al. Compu a ional and S uc u al Bio echnology Jou nal 19 (2021) 2398–2415
2413

[11] Ba o d D, Neel BG. Re ealing mechanisms o SH2 domain media ed
egula ion o he p o ein y osine phospha ase SHP-2. S uc u e (London,
England : 1993). 1998;6(3):249-54. Epub 1998/04/29. doi: 10.1016/s0969-
2126(98)00027-6. PubMed PMID: 9551546.
[12] Pluskey S, Wandless TJ, Walsh CT, Shoelson SE. Po en s imula ion o SH-PTP2
phospha ase ac i i y by simul aneous occupancy o bo h SH2 domains. The
Jou nal o biological chemis y 1995;270(7):2897–900.
[13] O inge EA, Bo ield MC, Shoelson SE. Tandem SH2 domains con e high
speci ici y in y osine kinase signaling. J Biol Chem. 1998;273(2):729-35. Epub
1998/02/14. doi: 10.1074/jbc.273.2.729. PubMed PMID: 9422724.
[14] Longo DL, Boussio is VA. Molecula and Biochemical Aspec s o he PD-1
Checkpoin Pa hway. N Engl J Med 2016;375(18):1767–78.
[15] Dance M, Mon agne A, Salles J-P, Ya A, Raynal P. The molecula unc ions o
Shp2 in he Ras/Mi ogen-ac i a ed p o ein kinase (ERK1/2) pa hway. Cell
Signal 2008;20(3):453–9.
[16] Xu E, Schwab M, Ma e e A. Role o p o ein y osine phospha ases in he
modula ion o insulin signaling and hei implica ion in he pa hogenesis o
obesi y-linked insulin esis ance. Re . Endoc ine Me abolic Diso de s 2014;15
(1):79–97. h ps://doi.o g/10.1007/s11154-013-9282-4.
[17] Ma asco M, Be eo i A, Weye shaeuse J, Tho ausch N, Siko ska J, K ausze J,
e al. Molecula mechanism o SHP2 ac i a ion by PD-1 s imula ion. Science.
Ad ances 2020;6(5):eaay4458. h ps://doi.o g/10.1126/sciad .aay4458.
[18] Peled M, Toche a AS, Sandigu sky S, Nayak S, Philips EA, Nichols KE, e al.
A ini y pu i ica ion mass spec ome y analysis o PD-1 unco e s SAP as a
new checkpoin inhibi o . P oc Na l Acad Sci U S A 2018;115(3):E468–77.
[19] Sugimo o S, Wandless TJ, Shoelson SE, Neel BG, Walsh CT. Ac i a ion o he
SH2-con aining p o ein y osine phospha ase, SH-PTP2, by phospho y osine-
con aining pep ides de i ed om insulin ecep o subs a e-1. J Biol Chem
1994;269(18):13614–22. Epub 1994/05/06 PubMed PMID: 7513703.
[20] Sun XJ, Ro henbe g P, Kahn CR, Backe JM, A aki E, Wilden PA, e al. S uc u e
o he insulin ecep o subs a e IRS-1 de ines a unique signal ansduc ion
p o ein. Na u e. 1991;352(6330):73-7. Epub 1991/07/04. doi: 10.1038/
352073a0. PubMed PMID: 1648180.
[21] Benne AM, Tang TL, Sugimo o S, Walsh CT, Neel BG. P o ein- y osine-
phospha ase SHPTP2 couples pla ele -de i ed g ow h ac o ecep o be a o
Ras. P oc Na l Acad Sci U S A 1994;91(15):7335–9. PubMed PMID: 8041791.
[22] Li W, Nishimu a R, Kashishian A, Ba ze AG, Kim WJ, Coope JA, e al. A new
unc ion o a phospho y osine phospha ase: linking GRB2-Sos o a ecep o
y osine kinase. Mol Cell Biol 1994;14(1):509–17.
[23] Vogel W, Ull ich A. Mul iple in i o phospho yla ed y osine phospha ase
SHP-2 engages binding o G b2 ia y osine 584. Cell G ow h Di e 1996;7
(12):1589–97. Epub 1996/12/01 PubMed PMID: 8959326.
[24] Lu W, Gong D, Ba -Sagi D, Cole PA. Si e-speci ic inco po a ion o a
phospho y osine mime ic e eals a ole o y osine phospho yla ion o SHP-
2 in cell signaling. Molecula cell. 2001;8(4):759-69. Epub 2001/10/31. doi:
10.1016/s1097-2765(01)00369-0. PubMed PMID: 11684012.
[25] Sun J, Lu S, Ouyang M, Lin L-J, Zhuo Y, Liu Bo, e al. An agonism be ween
binding si e a ini y and con o ma ional dynamics unes al e na i e cis-
in e ac ions wi hin Shp2. Na Commun 2013;4(1). h ps://doi.o g/10.1038/
ncomms3037.
[26] Huye G, Ramachand an C. The speci ici y o he N- e minal SH2 domain o
SHP-2 is modi ied by a single poin mu a ion. Biochemis y. 1998;37(9):2741-
7. Epub 1998/04/16. doi: 10.1021/bi9722913. PubMed PMID: 9485424.
[27] Huye G, Li ZM, Adam M, Huckle WR, Ramachand an C. Di ec de e mina ion
o he sequence ecogni ion equi emen s o he SH2 domains o SH-PTP2.
Biochemis y. 1995;34(3):1040-9. Epub 1995/01/24. doi: 10.1021/
bi00003a039. PubMed PMID: 7530043.
[28] Sweeney MC, Wa eille AS, Pa k J, Bu cha JP, T idandapani S, Pei D. Decoding
p o ein-p o ein in e ac ions h ough combina o ial chemis y: sequence
speci ici y o SHP-1, SHP-2, and SHIP SH2 domains. Biochemis y. 2005;44
(45):14932-47. Epub 2005/11/09. doi: 10.1021/bi051408h. PubMed PMID:
16274240.
[29] Anselmi M, Hub JS. An allos e ic in e ac ion con ols he ac i a ion mechanism
o SHP2 y osine phospha ase. Sci Rep 2020;10(1). h ps://doi.o g/10.1038/
s41598-020-75409-7.
[30] Anselmi M, Calliga i P, Hub JS, Ta aglia M, Bocchin uso G, S ella L. S uc u al
De e minan s o Phosphopep ide Binding o he N-Te minal S c Homology 2
Domain o he SHP2 Phospha ase. J Chem In Model 2020;60(6):3157–71.
h ps://doi.o g/10.1021/acs.jcim.0c00307.
[31] Bodenhausen G, Ruben DJ. Na u al abundance ni ogen-15 NMR by enhanced
he e onuclea spec oscopy. Chem Phys Le 1980;69(1):185–9. h ps://doi.
o g/10.1016/0009-2614(80)80041-8.
[32] Sklena V, Pio o M, Leppik R, Saudek V. G adien -Tailo ed Wa e Supp ession
o 1H–15N HSQC Expe imen s Op imized o Re ain Full Sensi i i y. J Magn
Reson, Se A 1993;102(2):241–5. h ps://doi.o g/10.1006/jm a.1993.1098.
[33] Pio o M, Saudek V, Sklena V. G adien - ailo ed exci a ion o single-quan um
NMR spec oscopy o aqueous solu ions. J Biomol NMR. 1992;2(6):661-5. Epub
1992/11/01. doi: 10.1007/BF02192855. PubMed PMID: 1490109.
[34] Dayie KT, Wagne G. Relaxa ion-Ra e Measu emen s o 15N–1H G oups wi h
Pulsed-Field G adien s and P ese a ion o Cohe ence Pa hways. J Magn Reson,
Se A 1994;111(1):121–6. h ps://doi.o g/10.1006/jm a.1994.1236.
[35] Fa ow NA, Muhandi am R, Singe AU, Pascal SM, Kay CM, Gish G, e al.
Backbone Dynamics o a F ee and a Phosphopep ide-Complexed S c Homology
2 Domain S udied by 15N NMR Relaxa ion. Biochemis y 1994;33
(19):5984–6003. h ps://doi.o g/10.1021/bi00185a040.
[36] Kay LE, To chia DA, Bax A. Backbone dynamics o p o eins as s udied by
ni ogen-15 in e se de ec ed he e onuclea NMR spec oscopy: applica ion o
s aphylococcal nuclease. Biochemis y 1989;28(23):8972–9. h ps://doi.o g/
10.1021/bi00449a003.
[37] Mulde FAA, de G aa RA, Kap ein R, Boelens R. An O - esonance Ro a ing
F ame Relaxa ion Expe imen o he In es iga ion o Mac omolecula
Dynamics Using Adiaba ic Ro a ions. J Magn Reson 1998;131(2):351–7.
h ps://doi.o g/10.1006/jm e.1998.1380.
[38] Geen H, F eeman R. Band-selec i e adio equency pulses. Jou nal o Magne ic
Resonance (1969). 1991;93(1):93-141. doi: h ps://doi.o g/10.1016/0022-
2364(91)90034-Q.
[39] Ko zhne DM, Sk ynniko NR, Mille O, To chia DA, Kay LE. An NMR
Expe imen o he Accu a e Measu emen o He e onuclea Spin-Lock
Relaxa ion Ra es. J Am Chem Soc 2002;124(36):10743–53. h ps://doi.o g/
10.1021/ja0204776.
[40] Hansen DF, Kay LE. Imp o ed magne iza ion alignmen schemes o spin-lock
elaxa ion expe imen s. J Biomol NMR 2007;37(4):245–55. h ps://doi.o g/
10.1007/s10858-006-9126-6.
[41] Li YC, Mon elione GT. Sol en Sa u a ion-T ans e E ec s in Pulsed-Field-
G adien He e onuclea Single-Quan um-Cohe ence (PFG-HSQC) Spec a o
Polypep ides and P o eins. J Magn Reson, Se B 1993;101(3):315–9. h ps://
doi.o g/10.1006/jm b.1993.1049.
[42] Renne C, Schleiche M, Mo ode L, Holak TA. P ac ical aspec s o he 2D 15N-
[1h]-NOE expe imen . J Biomol NMR. 2002;23(1):23-33. Epub 2002/06/14.
doi: 10.1023/a:1015385910220. PubMed PMID: 12061715.
[43] Fe age F, Cowbu n D, Ghose R. Accu a e Sampling o High-F equency Mo ions
in P o eins by S eady-S a e 15N{1H} Nuclea O e hause E ec
Measu emen s in he P esence o C oss-Co ela ed Relaxa ion. J Am Chem
Soc 2009;131(17):6048–9. h ps://doi.o g/10.1021/ja809526q.
[44] Hansen DF, Vallu upalli P, Kay LE. An Imp o ed 15N Relaxa ion Dispe sion
Expe imen o he Measu emen o Millisecond Time-Scale Dynamics in
P o eins. J Phys Chem B 2008;112(19):5898–904. h ps://doi.o g/10.1021/
jp074793o.
[45] Delaglio F, G zesiek S, Vuis e GW, Zhu G, P ei e J, Bax A. NMRPipe: A
mul idimensional spec al p ocessing sys em based on UNIX pipes. J Biomol
NMR 1995;6(3):277–93. h ps://doi.o g/10.1007/BF00197809.
[46] V anken WF, Bouche W, S e ens TJ, Fogh RH, Pajon A, Llinas M, e al. The
CCPN da a model o NMR spec oscopy: de elopmen o a so wa e pipeline.
P o eins. 2005;59(4):687-96. Epub 2005/04/09. doi: 10.1002/p o .20449.
PubMed PMID: 15815974.
[47] Hansen DF. Func ional and Da a Analysis (FuDA).
[48] Lipa i G, Szabo A. Model- ee app oach o he in e p e a ion o nuclea
magne ic esonance elaxa ion in mac omolecules. 1. Theo y and ange o
alidi y. J Am Chem Soc 1982;104(17):4546–59. h ps://doi.o g/
10.1021/ja00381a009.
[49] Lipa i G, Szabo A. Model- ee app oach o he in e p e a ion o nuclea
magne ic esonance elaxa ion in mac omolecules. 2. Analysis o expe imen al
esul s. J Am Chem Soc 1982;104(17):4559–70. h ps://doi.o g/
10.1021/ja00381a010.
[50] Dosse P, Hus J-C, Blackledge M, Ma ion D. E icien analysis o
mac omolecula o a ional di usion om he e onuclea elaxa ion da a. J
Biomol NMR 2000;16(1):23–8. h ps://doi.o g/10.1023/A:1008305808620.
[51] Dhulesia A, Bodenhausen G, Abe gel D. P edic ing con o ma ional en opy o
bond ec o s in p o eins by ne wo ks o coupled o a o s. J Chem Phys
2008;129(9):095107. h ps://doi.o g/10.1063/1.2969809.
[52] Li DW, B ueschweile R. A dic iona y o p o ein side-chain en opies om
NMR o de pa ame e s. J Am Chem Soc. 2009;131(21):7226-7. Epub 2009/05/
09. doi: 10.1021/ja902477s. PubMed PMID: 19422234.
[53] Bou ignies G. ChemEx.
[54] ASali A, Blundell TL. Compa a i e p o ein modelling by sa is ac ion o spa ial
es ain s. J Mol Biol. 1993;234(3):779-815. Epub 1993/12/05. doi: 10.1006/
jmbi.1993.1626. PubMed PMID: 8254673.
[55] Webb B, Sali A. Compa a i e P o ein S uc u e Modeling Using MODELLER.
Cu P o oc Bioin o ma ics. 2016;54:5 6 1-5 6 37. PubMed PMID: 27322406.
[56] Ab aham MJ, Mu ola T, Schulz R, Páll S, Smi h JC, Hess B, e al. GROMACS:
High pe o mance molecula simula ions h ough mul i-le el pa allelism om
lap ops o supe compu e s. So wa eX 2015;1-2:19–25. h ps://doi.o g/
10.1016/j.so x.2015.06.001.
[57] Van De Spoel D, Lindahl E, Hess B, G oenho G, Ma k AE, Be endsen HJ.
GROMACS: as , lexible, and ee. J Compu Chem. 2005;26(16):1701-18. Epub
2005/10/08. doi: 10.1002/jcc.20291. PubMed PMID: 16211538.
[58] Lindo -La sen K, Piana S, Palmo K, Ma agakis P, Klepeis JL, D o RO, e al.
Imp o ed side-chain o sion po en ials o he Ambe 99SB p o ein o ce
ield. P o eins 2010;78(8):1950–8.
[59] Wang J, Wol RM, Caldwell JW, Kollman PA, Case DA. De elopmen and es ing
o a gene al ambe o ce ield. J Compu Chem. 2004;25(9):1157-74. Epub
2004/04/30. doi: 10.1002/jcc.20035. PubMed PMID: 15116359.
[60] Sousa da Sil a AW, V anken WF. ACPYPE-An echambe py hon pa se in e ace.
BMC Res No es 2012;5(1):367. h ps://doi.o g/10.1186/1756-0500-5-367.
[61] Jo gensen WL, Chand asekha J, Madu a JD, Impey RW, Klein ML. Compa ison
o Simple Po en ial Func ions o Simula ing Liquid Wa e . J Chem Phys
1983;79(2):926–35. h ps://doi.o g/10.1063/1.445869.
[62] Essmann U, Pe e a L, Be kowi z ML, Da den T, Lee H, Pede sen LG. A smoo h
pa icle mesh Ewald me hod. J Chem Phys 1995;103(19):8577–93. h ps://doi.
o g/10.1063/1.470117.
M. Ma asco, J. Ki kpa ick, V. Nanna e al. Compu a ional and S uc u al Bio echnology Jou nal 19 (2021) 2398–2415
2414
[63] Hess B, Bekke H, Be endsen HJC, F aaije JGEM. LINCS: A linea cons ain
sol e o molecula simula ions. J Compu Chem 1997;18(12):1463–72.
h ps://doi.o g/10.1002/(SICI)1096-987X(199709)18:12<1463::AID-JCC4>3.0.
CO;2-H.
[64] Miyamo o S, Kollman PA. Se le: An analy ical e sion o he SHAKE and
RATTLE algo i hm o igid wa e models. J Compu Chem 1992;13(8):952–62.
h ps://doi.o g/10.1002/(ISSN)1096-987X10.1002/jcc. 13:810.1002/
jcc.540130805.
[65] Wang J, Wol RM, Caldwell JW, Kollman PA, Case DA. De elopmen and es ing
o a gene al ambe o ce ield. J Compu Chem. 2004;25(9):1157-74. Epub
2004/04/30. doi: 10.1002/jcc.20035. PubMed PMID: 15116359.
[66] Pa inello M, Rahman A. Polymo phic ansi ions in single c ys als: A new
molecula dynamics me hod. J Appl Phys 1981;52(12):7182–90. h ps://doi.
o g/10.1063/1.328693.
[67] G an BJ, Rod igues APC, ElSawy KM, McCammon JA, Ca es LSD. Bio3d: an R
package o he compa a i e analysis o p o ein s uc u es. Bioin o ma ics
2006;22(21):2695–6. h ps://doi.o g/10.1093/bioin o ma ics/b l461.
[68] Sca abelli G, G an BJ. Kinesin-5 allos e ic inhibi o s uncouple he dynamics o
nucleo ide, mic o ubule, and neck-linke binding si es. Biophys J 2014;107
(9):2204–13. h ps://doi.o g/10.1016/j.bpj.2014.09.019. PubMed PMID:
25418105.
[69] Yao X-Q, Malik RU, G iggs NW, Skj
æ
en L, T ayno JR, Si a amak ishnan S,
e al. Dynamic Coupling and Allos e ic Ne wo ks in he
a
Subuni o
He e o ime ic G P o eins. J. Biological Chem. 2016;291(9):4742–53. h ps://
doi.o g/10.1074/jbc.M115.702605.
[70] Zhou S, Shoelson SE, Chaudhu i M, Gish G, Pawson T, Hase WG, e al. SH2
domains ecognize speci ic phosphopep ide sequences. Cell 1993;72
(5):767–78. h ps://doi.o g/10.1016/0092-8674(93)90404-E.
[71] De Souza D, Fab i LJ, Nash A, Hil on DJ, Nicola NA, Baca M. SH2 Domains om
Supp esso o Cy okine Signaling-3 and P o ein Ty osine Phospha ase SHP-2
Ha e Simila Binding Speci ici ies. Biochemis y 2002;41(29):9229–36.
h ps://doi.o g/10.1021/bi0259507.
[72] Sweeney MC, Wa eille A-S, Pa k J, Bu cha JP, T idandapani S, Pei D. Decoding
P o einP o ein In e ac ions h ough Combina o ial Chemis y: Sequence
Speci ici y o SHP-1, SHP-2, and SHIP SH2 Domains. Biochemis y 2005;44
(45):14932–47. h ps://doi.o g/10.1021/bi051408h.
[73] Imho D, Wa eille A-S, May A, Zacha ias M, T idandapani S, Pei D. Sequence
Speci ici y o SHP-1 and SHP-2 S c Homology 2 Domains: CRITICAL ROLES OF
RESIDUES BEYOND THE pY+3 POSITION *. J Biol Chem 2006;281
(29):20271–82. h ps://doi.o g/10.1074/jbc.M601047200.
[74] Huang H, Li L, Wu C, Schibli D, Colwill K, Ma S, e al. De ining he Speci ici y
Space o he Human S c Homology 2 Domain *. Mol Cell P o eomics 2008;7
(4):768–84. h ps://doi.o g/10.1074/mcp.M700312-MCP200.
[75] Rahuel J, Gay B, E dmann D, S auss A, Ga cia-Eche e ía C, Fu e P, e al.
S uc u al basis o speci ici y o G b2-SH2 e ealed by a no el ligand binding
mode. Na S uc Biol. 1996;3(7):586-9. Epub 1996/07/01. doi: 10.1038/
nsb0796-586. PubMed PMID: 8673601.
[76] Waksman G, Kominos D, Robe son SC, Pan N, Bal imo e D, Bi ge RB, e al.
C ys al s uc u e o he phospho y osine ecogni ion domain SH2 o -s c
complexed wi h y osine-phospho yla ed pep ides. Na u e. 1992;358
(6388):646-53. Epub 1992/08/20. doi: 10.1038/358646a0. PubMed PMID:
1379696.
[77] Hayashi T, Senda M, Suzuki N, Nishikawa H, Ben C, Tang C, e al. Di e en ial
Mechanisms o SHP2 Binding and Ac i a ion A e Exploi ed by Geog aphically
Dis inc Helicobac e pylo i CagA Oncop o eins. Cell Repo s 2017;20
(12):2876–90. h ps://doi.o g/10.1016/j.cel ep.2017.08.080.
[78] Kaneko T, Huang H, Cao X, Li X, Li C, Voss C, e al. Supe binde SH2 domains ac
as an agonis s o cell signaling. Sci Signal 2012;5(243). h ps://doi.o g/
10.1126/scisignal.2003021.
[79] Pawson T, Gish GD, Nash P. SH2 domains, in e ac ion modules and cellula
wi ing. T ends Cell Biol. 2001;11(12):504-11. Epub 2001/11/24. doi: 10.1016/
s0962-8924(01)02154-7. PubMed PMID: 11719057.
[80] Ma asco M, Ca lomagno T. Speci ici y and egula ion o phospho y osine
signaling h ough SH2 domains. J S uc Biol X 2020;4:100026. h ps://doi.o g/
10.1016/j.yjsbx.2020.100026.
[81] Pa soukis N, Duke-Cohan JS, Chaudh i A, Aksoyla HI, Wang Q, Council A, e al.
In e ac ion o SHP-2 SH2 domains wi h PD-1 ITSM induces PD-1 dime iza ion
and SHP-2 ac i a ion. Commun Biol. 2020;3(1):128. Epub 2020/03/19. doi:
10.1038/s42003-020-0845-0. PubMed PMID: 32184441; PubMed Cen al
PMCID: PMCPMC7078208.
[82] Xu X, Sa ikas A, Dias-San aga a DC, Dolios G, La on an PJ, Tsai SC, e al. The
CUL7 E3 ubiqui in ligase a ge s insulin ecep o subs a e 1 o ubiqui in-
dependen deg ada ion. Molecula cell. 2008;30(4):403-14. Epub 2008/05/24.
doi: 10.1016/j.molcel.2008.03.009. PubMed PMID: 18498745; PubMed Cen al
PMCID: PMCPMC2633441.
[83] Keilhack H, Da id FS, McG ego M, Can ley LC, Neel BG. Di e se biochemical
p ope ies o Shp2 mu an s. Implica ions o disease pheno ypes. J Biol Chem.
2005;280(35):30984-93. Epub 2005/07/01. doi: 10.1074/jbc.M504699200.
PubMed PMID: 15987685.
[84] Bocchin uso G, S ella L, Ma inelli S, Flex E, Ca a C, Pan aleoni F, e al.
S uc u al and unc ional e ec s o disease-causing amino acid subs i u ions
a ec ing esidues Ala72 and Glu76 o he p o ein y osine phospha ase SHP-2.
P o eins S uc Func Bioin 2007;66(4):963–74. h ps://doi.o g/10.1002/
p o .21050.
[85] Ta aglia M, Gelb BD. Noonan synd ome and ela ed diso de s: gene ics and
pa hogenesis. Annu Re Genomics Hum Gene . 2005;6:45-68. Epub 2005/08/
30. doi: 10.1146/annu e .genom.6.080604.162305. PubMed PMID: 16124853.
[86] Ma inelli S, To e i P, Tin i M, S ella L, Bocchin uso G, Flex E, e al. Di e se
d i ing o ces unde lie he in a ian occu ence o he T42A, E139D, I282V and
T468M SHP2 amino acid subs i u ions causing Noonan and LEOPARD
synd omes. Human Molecula Gene ics. 2008;17(13):2018-29. doi: 10.1093/
hmg/ddn099.
[87] Papaleo E, Me eghe i P, Fan ucci P, G ando i R, De Gioia L. F ee-ene gy
landscape, p incipal componen analysis, and s uc u al clus e ing o iden i y
ep esen a i e con o ma ions om molecula dynamics simula ions: he
myoglobin case. J Mol G aph Model. 2009;27(8):889-99. Epub 2009/03/07.
doi: 10.1016/j.jmgm.2009.01.006. PubMed PMID: 19264523.
M. Ma asco, J. Ki kpa ick, V. Nanna e al. Compu a ional and S uc u al Bio echnology Jou nal 19 (2021) 2398–2415
2415