Electron microscopy and calorimetry of proteins in supercooled water

1
Vol.:(0123456789)
Scien i ic Repo s | (2022) 12:16512 | h ps://doi.o g/10.1038/s41598-022-20430-1
www.na u e.com/scien i ic epo s
Elec on mic oscopy
and calo ime y o p o eins
in supe cooled wa e
Jo ge H. Melillo1,5, Eliza e a Nikulina2,5, Maia a A. I ia e‑Alonso2, Sil ina Ce eny1,3* &
Alexande M. Bi ne 2,4*
Some o he bes nuclea ing agen s in na u e a e ice‑nuclea ing p o eins, which boos ice g ow h
be e han any o he ma e ial. They can induce imme sion eezing o supe cooled wa e only a ew
deg ees below 0 °C. An open ques ion is whe he his abili y also ex ends o he deposi ion mode, i.e.,
o wa e apo . In his wo k, we used h ee p o eins, apo e i in, InaZ (ice nuclea ion ac i e p o ein
Z), and myoglobin, o which he i s wo a e classi ied as ice‑nuclea ing p o eins o he imme sion
eezing mode. We s udied he ice nuclea ion abili y o hese p o eins by di e en ial scanning
calo ime y (imme sion eezing) and by en i onmen al scanning elec on mic oscopy (deposi ion
eezing). Ou da a show ha InaZ c ys allizes wa e di ec ly om he apo phase, while apo e i in
i s condenses wa e in he supe cooled s a e, and subsequen ly c ys allizes i , jus as myoglobin,
which is unable o nuclea e ice.
Supe cooled wa e (SCW) is o u mos in e es o en i onmen al p ocesses, mainly in he a mosphe e1,2. Unde -
s anding SCW o ma ion is c ucial in a ious echnological p ocesses3–5, usually o a oid undesi ed ice o ma-
ion. SCW a ambien p essu e o ms below he eezing poin , whe e pu e wa e can be p e en ed om eezing
o long pe iods, e.g., in ha d con inemen 6 o in so con inemen 7–9. SCW becomes p og essi ely less s able
a lowe empe a u es. In e y pu e condi ions, SCW can emain liquid un il − 38°C10,11, which co esponds o
homogeneous c ys alliza ion, i.e., in his case, ice nuclea ion s a s inside bulk wa e (see Fig.1a). Fo expe i-
men ally p obing o SCW many me hods ha e been de eloped, e.g., eezing indi idual and cascade d ople s12,13,
e apo a ion om eezing supe cooled sessile d ople s14, condensa ion om he gas phase (wa e apo ) on
biphilic su aces15, pulsed-lase –hea ing echniques10, and simple op ical de ec ion o eezing in ml con aine s16.
The signi icance o SCW in he a mosphe e is mainly ela ed o ae osols in connec ion wi h cloud o ma-
ion. The mechanism by which wa e apo condensa es o o m ice clouds has been in es iga ed o decades
because i is essen ial o unde s anding cloud o ma ion, which in u n is one o he decisi e ac o s in clima e
and wea he 17. Me ely cooling wa e apo does no p oduce ice. Ra he , ice o ma ion equi es o e coming he
nuclea ion ba ie . A possible way (depending on wea he condi ions) is imme sion eezing, whe e a d ople o
SCW combines wi h a solid pa icle, and subsequen ly eezes (see Fig.1b). Typical imme sion eezing expe i-
men s ha e been ca ied ou wi h he Zü ich ice nuclea ion chambe (ZINC)18,19, wi h he po able imme sion
mode cooling chambe (PIMCA)20,21, wi h he d ople ice nuclei coun e Zu ich (DRINCZ)22, wi h di e en ial
scanning calo ime e s (DSCs)23–25, and wi h op ical mic oscopes equipped wi h a ious cooling sys ems26–29. The
essen ial ea u e is always cooling o an aqueous solu ion o suspension a a con olled a e, ypically a ambien
p essu e, o ind he c ys alliza ion empe a u e (Tc ys ). In o he wo ds, an aqueous liquid mix u e is (supe )
cooled un il i c ys allizes.
In con as o imme sion eezing, he deposi ion nuclea ion mode s a s om wa e apo , and he c ys al-
liza ion p ocess can ollow wo di e en ou es, as shown in Fig.1c27,30. One mechanism is he deposi ion o SCW
on a (solid) pa icle, ollowed by ice nuclea ion and g ow h, ul ima ely esul ing in a bulk ice c ys al31. The o he
one is di ec apo o ice nuclea ion deposi ion on he pa icle. Deposi ion nuclea ion wi h bo h mechanisms
is ep esen ed in Fig.1c. One o he mos ad anced ins umen s o s udy such gas–solid ansi ions wi h high
empo al and spa ial esolu ion is he en i onmen al scanning elec on mic oscope (ESEM)32–36.
OPEN
1Donos ia In e na ional Physics Cen e (DIPC), Paseo Manuel de La dizabal 4, 20018 Donos ia-San
Sebas ián, Spain. 2CIC nanoGUNE (BRTA), A . Tolosa 76, 20018 Donos ia-San Sebas ián, Spain. 3Cen o de
Física de Ma e iales (CFM, CSIC-UPV/EHU)-Ma e ial Physics Cen e (MPC), Paseo Manuel de La dizabal 5,
20018 Donos ia-San Sebas ián, Spain. 4Ike basque, Basque Founda ion o Science, Pl. Euskadi 5, 48009 Bilbao,
Spain. 5
These au ho s con ibu ed equally: Jo ge H. Melillo and Eliza e a Nikulina. *email: sil ina.ce eny@
ehu.es; [email p o ec ed]
2
Vol:.(1234567890)
Scien i ic Repo s | (2022) 12:16512 | h ps://doi.o g/10.1038/s41598-022-20430-1
www.na u e.com/scien i ic epo s/
Al hough ice nuclea ion by imme sion and deposi ion consis s o wa e c ys alliza ion on a “ o eign” pa icle,
bo h modes a e independen . They can p o ide di e en in o ma ion abou he nuclea ion p ope ies o he
pa icle. Fo example, a ecen s udy showed ha he nuclea ion abili y o he aluminosilica e eldspa and o
qua z, wo essen ial ae osol mine als, bo h based on SiO2 la ices, di e s be ween imme sion and deposi ion
mode27. In ac , i is no e en i ial o compa e bo h modes.
Figu e2 shows he wa e phase diag am. The liquid- o-ice i s o de ansi ion co esponds o imme sion
eezing (da k yellow a ow), and he apo - o-ice ansi ion o deposi ion eezing (g een a ow), espec i ely.
Howe e , he apo - o-ice ansi ion can p oceed in wo ways, ep esen ed by he ed and he blue cu e in he
phase diag am (Fig.2). These cu es a e based on an imp o ed e sion o he Clausius-Clapey on equa ion1.
Figu e1. Ice c ys alliza ion mechanisms. (a) Homogeneous ice nuclea ion. (b) and (c) He e ogeneous ice
nuclea ion; (b) shows imme sion eezing nuclea ion and (c) deposi ion nuclea ion. C ea ed wi h Blende 3.0
(www. blend e . o g/).
Figu e2. Wa e phase diag am. The da k yellow a ow co esponds o imme sion eezing expe imen s, and
he g een a ow o deposi ion eezing. The ed cu e indica es he ice sa u a ion p essu e (hice = 1), and he blue
cu e he sa u a ion p essu e o supe cooled wa e (hSCW = 1). C ea ed wi h O iginP o 9.0 (www. o igi nlab.
com/).
3
Vol.:(0123456789)
Scien i ic Repo s | (2022) 12:16512 | h ps://doi.o g/10.1038/s41598-022-20430-1
www.na u e.com/scien i ic epo s/
The ed cu e ep esen s he ice sa u a ion p essu e Pice, and he blue one is he SCW sa u a ion p essu e PSCW,
espec i ely. We can de ine he ela i e humidi y o ice as:
whe e P is he absolu e p essu e. Simila ly, we can de ine he ela i e humidi y wi h espec o SCW as hSCW by
using PSCW ins ead o Pice(see Eq. (1)).When hSCW ≤ 1 and hice ≥ 1, wa e apo condensa es o ice, and when
hSCW ≥ 1, wa e apo o ms i s SCW and subsequen ly ice. Ne e heless, i is no always possible o ob ain he
me as able SCW om deposi ion eezing; mos impo an , ice nuclea ion mus be supp essed.
Al hough imme sion and deposi ion eezing modes a e di e en , ice nuclea ion p oceeds in he p esence
o a “ o eign” solid pa icle. This is usually assumed o be a mine al (ino ganic), bu can also be composed o
o ganic ma e 37. Recen wo k on cloud nuclea ion sugges s ha bioo ganic molecules, such as p o eins, play a
decisi e ole38–42. On he one hand, his migh be su p ising because biomolecules had adi ionally no been
conside ed as ae osols; on he o he hand, ice-nuclea ing p o eins (INPs) a e widesp ead in na u e. They induce
ice o ma ion in SCW, e.g., he bac e ium Pseudomonas sy ingae43 employs he ice nuclea ion ac i e p o ein Z
(Ina Z) on i s ou e memb ane o igge eezing a empe a u es up o − 2°C44 (ice g ow h ex ending ou wa d
om he bac e ial su ace is use ul o p o ide nu ien s). InaZ is he mos e icien and one o he bes s udied
INPs. I comp ises ~ 1200 amino acids45 wi h many epea s, which a e a anged in a β-helical olding, which in
u n is based on s acked β-shee s. The con o ma ion is simila o ha o insec an i eeze p o ein s uc u es46,47
(which p o ide eeze p o ec ion inside cells). The β-shee s esul o bo h p o ein ypes in ex ended and almos
la su aces. Fu he simila i ies a e ypical ice-binding sequences (TxT mo i s48,49) and he abili y o a anging
he ex e nal hyd a ion laye in o an ice-like s uc u e50,51, despi e he opposi e unc ion o ice nuclea ion. How-
e e , InaZ has a unique p ope y, namely he capaci y o p omo e ene gy ans e ac oss he wa e in e ace51.
Mo eo e , InaZ has di ec applica ions such as eezing echnology and ood p ese a ion, and especially he
induc ion o snow o ma ion (exploi ed in he comme cial p oduc Snomax)52.
The e a e, howe e , p obably many mo e p o ein s uc u es ( e na y and qua e na y), which can p omo e ice
nuclea ion. Fo example, Cascajo-Cas esana e al.16 ecen ly ound ha he apo e i in p o ein cage is a good ice
nuclea o in imme sion eezing expe imen s. Apo e i in is ubiqui ous in many cells om bac e ia o e eb a a.
I has 174 amino acids, a anged in α-helical con o ma ion, bu o ms a 24-me sphe ical cage (4176 amino acids),
which is used by he o ganisms o s o e i on. I s ice nuclea ion abili y is appa en ly no exploi ed by na u e. I
could be based on speci ic p ope ies o he cage su ace (hyd ophilic, nega i ely cha ged, and, compa ed o small
p o eins, o ela i ely small cu a u e)53. Apo e i in has no echnical use, bu is o g ea in e es o nanoscale
science and nano echnological uses, also in biomedicine.
Myoglobin is a small p o ein p esen in high concen a ions in e eb a a; i s ask is oxygen anspo . I
comp ises only 153 amino acids olded in o α-helices; i s mo phology is almos sphe ical wi h a high cu a u e.
Myoglobin was he i s p o ein o be cha ac e ized by X- ay di ac ion. I is o en used as a s anda d example
o a globula p o ein in esea ch; u he mo e, i can be en isaged as he “na u al colo ” in ed mea .
He e, we p esen expe imen s on ice nuclea ion on he abo e-men ioned h ee p o eins, in bo h modes,
imme sion and deposi ion eezing. We in es iga ed apo e i in, and chose InaZ as posi i e con ol because i is
an excellen INP, and myoglobin, which does no nuclea e ice, as nega i e con ol. Ou ques ion is whe he he
p o eins ha e he same nuclea ion abili y in bo h nuclea ion modes. To his end, we analyzed imme sion eez-
ing by di e en ial scanning calo ime y (DSC) in solu ion, and s udied deposi ion eezing by ESEM. The la e
allowed us o ollow he a e o liquid mic od ople s in con ac wi h he (d y) p o eins, and in con ac wi h a solid
su ace (coppe ). Ou esul s indica e ha apo e i in is a good INP in imme sion mode bu no in deposi ion
mode. The e o e, he wo c ys alliza ion modes a e no co ela ed.
Me hods
Sample p epa a ion. Apo e i in (Sigma Ald ich, om equine spleen, ba ch A3641) was dilu ed o
3.4 × 10−4g/ml. To emo e bu e sal s and p e en agg ega ion o he indi idual cages, he dilu ed solu ion was
dialyzed in ammonium bica bona e (NH4HCO3) bu e o 96h. Fo his pu pose, he samples we e suspended
in 10mM ammonium bica bona e (Sigma-Ald ich, 09,830), pH = 7.4 o 7.6, p epa ed wi h Milli-Q wa e . Dialy-
sis was ca ied ou in 10,000 MWCO dialysis casse es (The mo Scien i ic) o 96h, wi h he ammonium bica -
bona e bu e eplaced e e y 24 h54. In he acuum o he ESEM chambe , NH4HCO3 comple ely decomposes
in o gaseous H2O, CO2, and NH3. The ad an age o his p ocedu e is ha po en ial apo e i in disassembly in
pu e wa e (pH = 7, ex emely low ionic s eng h) is a oided.
In addi ion o apo e i in, we employed myoglobin and Snomax solu ions as nega i e and posi i e con ols
o ice nuclea ion, espec i ely. Snomax is a comme cial p oduc ha con ains (among addi i es such as ca bo-
hyd a es and ash) ice-ac i e p o ein complexes om he bac e ium Pseudomonas sy ingae. This snowmaking
addi i e eezes wa e a a empe a u e close o 0°C. Snomax is he e o e conside ed an excellen INP. We used
i wi hou u he pu i ica ion. Myoglobin (Sigma Ald ich, om equine hea ) is no known o nuclea e ice.
Myoglobin was dialyzed in 10,000 MWCO dialysis casse es (The mo Scien i ic) o 24h in wa e , eplacing i
h ee imes du ing he day, and once o e nigh . Bo h Snomax and myoglobin we e dilu ed o 3.4 × 10−4g/ml.
Imme sion eezing and DSC. DSC measu emen s we e ca ied ou on ∼ 10mg specimens in a Q2000 TA
Ins umen s, ope a ed in s anda d mode. Sealed aluminum pans we e used o all he samples. Fo he nuclea-
ion expe imen s, he samples we e cooled a 1K/min om 20 o − 33°C o de e mine he c ys alliza ion em-
pe a u e. A e annealing a − 33°C o 5min, he samples we e ehea ed a 10K/min o de e mine he mel ing
p ope ies. A helium low a e o 25mL/min was used h oughou .
(1)
h
ice =
P
P
ice
4
Vol:.(1234567890)
Scien i ic Repo s | (2022) 12:16512 | h ps://doi.o g/10.1038/s41598-022-20430-1
www.na u e.com/scien i ic epo s/
Deposi ion eezing and ESEM. We employed a Quan a 250 ESEM (FEI, Ne he lands), which p o ides
wa e apo p essu es o > 2000Pa and sample cooling by a Pel ie s age o < − 20°C. ESEM expe imen s we e
ca ied ou in low acuum mode (p essu e limi a ion 200Pa) a − 20°C. This allows he eco ding o la ge-scale
images (mm size) because a p essu e-limi ing small ape u e a he column cone is no equi ed.
On he Pel ie s age, a homemade coppe piece (cylind ical s ub) was ixed. Hea -conduc i e sil e pas e
was placed be ween he s age and he coppe piece o op imal hea ans e . The coppe su ace is he coldes
pa in con ac wi h he wa e apo , hence, condensa ion s a s he e. To achie e a highly s able empe a u e, i
is use ul o es ic his su ace a ea. To his end, we employed a poly(dime hylsiloxane) (PDMS) ilm wi h low
he mal conduc i i y (0.2 Wm−1 K−1). Fo dec easing he sample su ace a ea, oo, we co e ed i wi h ca bon ape,
moun ed abo e he PDMS ilm, lea ing a window o only se e al mm2. He e, a d ople o p o ein solu ion was
placed on he coppe su ace and e apo a ed in ai . This p ocedu e a oids ice c ys alliza ion on exposed sample
edges55. Figu e3a shows a c oss-sec ional scheme o he comple e assembly.
Be o e e e y new deposi ion expe imen , he coppe s ub was polished ( inal s age wi h a colloidal suspension
o 60nm SiO2 pa icles) o emo e any isible damage, and cleaned in an ul asound ba h, i s wi h ace one, hen
isop opanol, and inally wa e . The p o ein solu ion was deposi ed on he upwa d acing su ace o he coppe
piece by d op-cas ing, and d ied in ai (see abo e). Figu e3b shows he Pel ie s age eady o he deposi ion
expe imen s. The ed ci cle indica es he d y p o ein spo . I is impo an o highligh ha , whene e ice nuclea-
ion occu s, he coppe su ace should be cleaned, and a esh p o ein spo should be p epa ed. We ound ice
in ESEM samples e en a e exposu e o high acuum, sugges ing ha ob aining a “comple ely d y” sample is
ex emely challenging56.
Figu e3. (a) Schema ic o e iew and c oss-sec ion o Pel ie s age p epa ed o ESEM expe imen s on
SCW (no o scale). (b) Pho og aph o he Pel ie s age wi h he coppe s ub (subs a e su ace). The ed ci cle
indica es he p o ein spo . (c) Zoom o he phase diag am o wa e . G een a ows indica e he p o ocol o
deposi ion eezing expe imen s. P = 103 and 126Pa co espond o p essu e alues o each hice = 1 and hSCW = 1,
espec i ely, a − 20°C. The numbe s e e o he s eps o he deposi ion eezing expe imen s (see ex ). C ea ed
wi h O iginP o 9.0 (www. o igi nlab. com/) and Co elD aw (www. co el d aw. com/).
5
Vol.:(0123456789)
Scien i ic Repo s | (2022) 12:16512 | h ps://doi.o g/10.1038/s41598-022-20430-1
www.na u e.com/scien i ic epo s/
In Fig.3c, g een a ows indica e he p o ocol o deposi ion expe imen s in ESEM (no e ha he p essu e
is inc eased by admi ing wa e apo om a ese oi , which is connec ed o he chambe h ough a ca e ully
egula ed leak al e):
1. Chambe p essu e dec ease o high acuum (< 10−2Pa).
2. Tempe a u e dec ease o − 20°C.
3. Chambe p essu e ise o jus below hice = 1.
4. P essu e jump o jus below hSCW = 1.
5. P essu e jump o hSCW > 1.
Cha ac e iza ion o he coppe su ace. To cha ac e ize he coppe su ace (which we also call “sub-
s a e” in he ollowing, no o be con used wi h enzyme subs a es) in e ms o we abili y and oughness, s a ic
wa e con ac angle (WCA) and AFM measu emen s we e conduc ed. Figu e4a demons a es ha he coppe
subs a e was mo e hyd ophobic han an apo e i in spo , as seen om he signi ican ly highe wa e con ac
angles. Thus, he apo e i in-deco a ed su aces exhibi a be e we abili y han he coppe su ace, he con ac
angle on coppe is a leas 25° highe han ha on apo e i in. The dependencies o s a ic wa e con ac angle
on su ace exposu e ime in he ai , bo h on coppe and apo e i in, a e explained by ai bo ne con amina ion,
which gene ally ende s su aces mo e hyd ophobic o e ime57, compa ed o eshly p epa ed su aces. The
e ec is clea ly p esen on coppe and on apo e i in. The o he decisi e ea u e is he oughness. On he mm and
mic oscale, ou coppe su ace has a mi o inish. This is e i ied by AFM, which shows a oughness o < 20nm
(on a 3µm × 3µm a ea), Fig.4b. We can in e ha he e a e a he ew ypical condensa ion nuclei; once wa e
apo is condensing, ou su ace should hinde i s immedia e eezing. In passing, we no e ha he subs a e is
no pu e coppe ; p epa a ion in wa e and exposu e o ai ansla es in o he p esence o a hin laye o oxide(s).
Resul s and discussion
Wa e c ys alliza ion da a om imme sion eezing. This sec ion in es iga es he c ys alliza ion o
aqueous solu ions o apo e i in, Snomax, and myoglobin in imme sion eezing mode. Figu e5 shows he hea
low (HF) as a unc ion o empe a u e, o pu e wa e and o all he solu ions. The p o ein concen a ion was
in all cases cp = 3.4 10−4g/ml. Wa e c ys alliza ion is an exo he mic p ocess; consequen ly, hea is eleased o he
su ounding. The e o e, he onse o c ys alliza ion is iden i ied as he empe a u e whe e he hea low s a s o
inc ease.
As seen in Fig.5, he onse o c ys alliza ion o bulk wa e was (− 24.0 ± 1.0)°C, measu ed as an a e age o 30
independen measu emen s a a cooling a io o 1K/min. C ys alliza ion o pu e bulk wa e is a homogeneous
p ocess occu ing a ~ − 38°C11. Howe e , he impe ec ions o a DSC expe imen , such as he su aces o he
aluminum pans and possible wa e impu i ies, induce he e ogeneous c ys alliza ion. In such a case, he onse
o c ys alliza ion is shi ed o be ween − 25 and − 17°C, depending on he cooling a e26. A simila c ys alliza-
ion empe a u e, (− 20.6 ± 0.9)°C, was ound o he myoglobin solu ion, a p o ein ha does no show any ice
nuclea ion o ice binding p ope ies. This beha io is expec ed: An aqueous solu ion o myoglobin con ains small
globula molecules, which a e no agg ega ed. Thei su ace is no la bu highly cu ed: The adius o ~ 2nm
is jus abo e molecula adii, he shape is a om ha o a pe ec sphe e, and each su ace spo has i s own
speci ic beha io , de e mined by he local amino acid sequence. This is a signi ican di e ence o he p oposed
equi emen o low cu a u e o INPs53.
On he con a y, he c ys alliza ion o wa e wi h apo e i in occu ed a a highe empe a u e (− 5.4 ± 0.1°C),
and Snomax solu ions (con aining InaZ) showed (− 4.4 ± 0.1)°C. These empe a u es a e much highe han hose
obse ed in pu e wa e , o in wa e wi h myoglobin. This inc ease in he c ys alliza ion empe a u e is due o he
ice nuclea ion abili y o bo h p o eins, as p e iously epo ed12. O he many in e p e a ions, we he e poin ou
one o he mos simple ones, based on a pu ely geome ic e ec 58, which links ice nuclea ion o la molecula
Figu e4. (a) Wa e con ac angle as a unc ion o exposu e o ai , measu ed on he coppe s ub su ace (blue
disks, h ee independen measu emen se ies) and on an apo e i in d ople d ied on coppe ( ed squa es, wo
independen measu emen se ies); (b) su ace opog aphy measu ed by AFM. C ea ed wi h WSXM (wsxm.eu/)
and Co elD aw (www. co el d aw. com/).

6
Vol:.(1234567890)
Scien i ic Repo s | (2022) 12:16512 | h ps://doi.o g/10.1038/s41598-022-20430-1
www.na u e.com/scien i ic epo s/
su aces. The cu a u es o he la ge p o ein InaZ, wi h i s ex ended s acked β-shee s, and o he la ge sup a-
molecula su ace o apo e i in (6nm adius, almos pe ec ly sphe ical) a e ela i ely small. Mo eo e , bo h
apo e i in and InaZ eadily o m agg ega es16, esul ing in e en lowe cu a u es. In con as , he cu a u e o
a small globula p o ein such as myoglobin (~ 2nm adius) is e y high. Apo e i in showed se e al addi ional
c ys alliza ion o e-c ys alliza ion p ocesses a lowe empe a u es. The eleased hea was e y small, so hese
p ocesses migh no ha e occu ed in all pa s o he sample.
In he ollowing, we in es iga e whe he he abili y o apo e i in and Snomax o bind ice inside a d ople
o supe cooled wa e (i.e., imme sion eezing mode) ex ends o p o eins in con ac wi h supe sa u a ed apo
(i.e., deposi ion mode).
Wa e c ys alliza ion du ing deposi ion eezing on Snomax and myoglobin. We s a by ana-
lyzing how he c ys alliza ion o wa e is p oduced on a spo o solid Snomax, which con ains InaZ. This p o ein
is conside ed an excellen ice binding and ice nuclea ing ma e ial, and is known o es ic supe cooling o a me e
4.4K. The e o e, we had expec ed as ice condensa ion on he InaZ spo (we e e o his as a "posi i e con ol").
Figu e6a shows he Snomax spo (da k con as ) on he coppe subs a e unde "no ice" condi ions (P = 102Pa
and T = − 20°C). Ice g ow h was igge ed when he p essu e was inc eased o 119Pa (Fig.6b). Subsequen ly, ice
g ew exclusi ely inside he spo (Fig.6c,d).The expe imen shows ha ice g ows om he im o he spo (see
Fig.SI1 in he supplemen a y in o ma ion), and apidly co e s all he subs a e. This also p o es ha ice g ows
exclusi ely om he Snomax spo , as no ice nuclea ion was igge ed on he coppe ou side he p o ein spo .
Hence, o he InaZ p o ein (in Snomax), wa e c ys allizes di ec ly om he apo phase, p e en ing he o -
ma ion o SCW. We conclude ha InaZ is an excellen INP since i shows an ex ao dina y abili y o a oid SCW in
bo h imme sion and deposi ion modes. This esul also con i ms ou choice o InaZ/Snomax as a posi i e con ol.
Nex , we es ed myoglobin, which we had designa ed as a nega i e con ol because he DSC esul s had
shown ha i is a ela i ely poo ice nuclea o . Figu e7a shows he myoglobin spo a hSCW = 1. Ne e heless, we
de ec ed nei he ice no SCW. We inc eased he p essu e well abo e Pwa e , o each hSCW > 1. When hSCW > 1 and
when c ys alliza ion was no igge ed (see Fig.7b,c abo e he dashed line), we obse ed mic oscale d ople s
(da k disks) on he subs a e and on he myoglobin spo . We in e p e hese d ople s as liquid SCW, bu due o
Figu e5. Hea Flow (each cu e se o o cla i y) as a unc ion o he empe a u e, o pu e wa e , and
o aqueous solu ions o apo e i in, Snomax, and myoglobin (all p o ein concen a ions 3.4 10−4g/ml).
The c ys alliza ion empe a u es a e indica ed in he plo . Apo e i in shows addi ional c ys alliza ion o
e-c ys alliza ion p ocesses a lowe empe a u es. C ea ed wi h O iginP o 9.0 (www. o igi nlab. com/).
Figu e6. (a) Snomax spo on he coppe subs a e a he condi ion P = 102Pa and T = − 20°C (hice < 1) whe e
no ice condensa ion is obse ed. (b–d) Time se ies o obse ing a Snomax spo . Ice g ow h is igge ed on he
spo a he condi ion P = 119Pa, hice = 1.05 and hSCW = 0.94. Figu es b–d a e images cap u ed du ing ice g ow h a
(b) = 0, c) = 20s, and d) = 67s.
7
Vol.:(0123456789)
Scien i ic Repo s | (2022) 12:16512 | h ps://doi.o g/10.1038/s41598-022-20430-1
www.na u e.com/scien i ic epo s/
he ins abili y o his condi ion, ou empo al esolu ion was insu icien o analyze he liquid na u e o he SCW
d ople s, e.g., mobili y and a changing h ee-phase line, which we will desc ibe below o he case o apo e i in.
Figu e7c shows ha , du ing he scan, SCW c ys allized (see he dashed line). The eezing on p opaga ed
e y as inside he p o ein spo and c ys allized all neighbo ing d ople s o SCW, p obably by ice b idging59.
Howe e , ice b idges a e no gua an eed o connec o SCW. I an SCW liquid d ople is su icien ly dis an om
he eezing on , as in he case o he ed ec angle in Fig.7b,d, he SCW liquid d ople can e apo a e be o e
an ice b idge can connec 59. In addi ion, we ha e obse ed a second on om he le o he images (yellow
ellipse). This new on was no in con ac wi h he p o ein spo , so i s o igin should ha e been ice c ys alliza ion
gene a ed on he coppe subs a e ou side o ou obse a ion window.
In conclusion, myoglobin is a poo ice nuclea o in bo h imme sion and deposi ion. No ably, in deposi ion
eezing, wa e c ys alliza ion occu s by he i s mechanism, i.e., wa e is condensed in he o m o SCW and
c ys allizes. This is expec ed, and we can employ ou esul as a nega i e con ol.
Supe cooled wa e and wa e c ys alliza ion du ing deposi ion eezing on apo e i in. Fig-
u e8 shows an apo e i in spo unde he same condi ions as in Fig.6b. Howe e , we ound no ice in his case.
By inc easing he p essu e ela i ely as om 119 o 131Pa (hSCW > 1), we we e able o c ea e d ople s (Fig.8a),
which we in e p e as SCW. Upon aising he p essu e u he , om 131 o 134Pa, he amoun o SCW inc eased
(see Fig.8b). When we dec eased he p essu e o 129Pa, wi h he p ima y goal o s ay close o hSCW = 1, he
SCW d ople s condensa ed h oughou he comple e a ea o he apo e i in spo .
The cha ac e is ics o he SCW d ople s obse ed in Fig.8 can be explained by conside ing he na u e o
he su aces. Since he apo e i in spo ep esen s a hyd ophilic su ace, SCW liquid d ople s sp ead ou and
occupy mo e su ace a ea han ou side he spo (on he coppe su ace). In addi ion, he shape o hese d ople s
is i egula . Fa away om he apo e i in spo we encoun e a clean (albei oxidized) coppe su ace, which is
Figu e7. Time- esol ed ESEM o he condensa ion o wa e apo , caused by a p essu e excu sion, on a
myoglobin spo on coppe . (a) = 0s and P = 126Pa. No ice o SCW a e de ec ed. (b) = 3s and P = 135Pa. SCW
s a s o appea on he p o ein spo and on he coppe subs a e. The ed ec angle indica es a de ail o SCW
on he ba e coppe su ace, and he yellow ellipse ma ks a coppe zone ee o SCW. (c) = 6s and P = 128Pa.
Du ing he scan, SCW e apo a es in he ed zone. A he same ime, ice g ow h is igge ed inside he p o ein
spo (see he do ed whi e line) and causes c ys alliza ion o neighbo ing SCW d ople s by ice b idging. (d)
= 9s and P = 129Pa. Ice now co e s he comple e p o ein spo . I has also appea ed in he yellow a ea, which
is no in con ac wi h he spo . SCW con inues e apo a ing in he ed zone. Modi ied wi h Powe Poin (www.
mic o so . com/ es- es/ mic o so - 365/ powe poin ).
8
Vol:.(1234567890)
Scien i ic Repo s | (2022) 12:16512 | h ps://doi.o g/10.1038/s41598-022-20430-1
www.na u e.com/scien i ic epo s/
mo e hyd ophobic (Fig.4), hence, liquid d ople s ha e a ound shape and a e smalle in his a ea. SCW liquid
d ople s we e p e iously de ec ed on hyd ophobic su aces o simila cha ac e is ics30,58,60. Howe e , his is o ou
knowledge he i s ime ha an ESEM expe imen has de ec ed SCW liquid d ople s on a hyd ophilic su ace
(i.e. on ou p o ein).
Figu es8d– p o ide u he p oo o he liquid na u e o he mic oscale d ople s: They mo e and coalesce
while eezing con inues. The mo emen is seen as a changing h ee-phase line (ci cum e ence o a d ople )
apo –liquid–solid. The coalescence o g owing ice c ys als esul s in ough (and necessa ily solid) su aces,
which we ound se e al seconds la e . The ideo in he supplemen a y in o ma ion p o ides a good isualiza ion
o he liquid na u e o SCW, and o he ice nuclea ion p ocess.
Simila o myoglobin, SCW c ys alliza ion was igge ed on he apo e i in spo du ing he scan, co e ing
i comple ely (dashed line in Fig.8e, he as ESEM scan di ec ion is ho izon al). The eezing on p opaga ed
Figu e8. (a–d) SCW d ople s on an apo e i in spo and on he coppe subs a e. (a) = 0s and P = 131Pa. (b)
= 2s and P = 134Pa. (c) = 5s and P = 129Pa. (d) = 8s and P = 129Pa. (d– ) C ys alliza ion p ocess o SCW.
The yellow iangle shows SCW d ople s ha emain liquid du ing he c ys alliza ion p ocess. The blue disk
ma ks an ice b idging e en du ing eezing on mo emen . (e) = 9s and P = 129Pa. Du ing he scan, SCW
c ys allizes apidly (dashed line) on he apo e i in spo and on some pa s o he coppe subs a e. ( ) = 10s
and P = 129Pa. Modi ied wi h Powe Poin (www. mic o so . com/ es- es/ mic o so - 365/ powe poin ).
9
Vol.:(0123456789)
Scien i ic Repo s | (2022) 12:16512 | h ps://doi.o g/10.1038/s41598-022-20430-1
www.na u e.com/scien i ic epo s/
e y as by con ac eezing. The wa e c ys alliza ion has u he mo e p opaga ed om he apo e i in spo o
he emaining ba e coppe su ace by ice b idging. Ice b idging migh be p esen on coppe , oo, see Fig.8d,e
(compa e he blue a eas). Al oge he , he eezing on p opaga ion is he e no ins an aneous because he SCW
d ople s a e mo e isola ed. In Fig.8 , all SCW d ople s ha e c ys allized on he apo e i in spo , and almos
all on he coppe subs a e, while in he co ne o he igu e (yellow iangle) he SCW d ople s emain liquid,
indica ing ha he eezing on had s a ed inside he p o ein spo .
Conclusions
This wo k combines calo ime y and eal- ime en i onmen al scanning elec on mic oscopy (ESEM) measu e-
men s o compa e ice nuclea ion in he liquid phase (imme sion nuclea ion) and om he gas phase (deposi ion
nuclea ion), o h ee p o eins. In addi ion, i p esen s elec on mic oscopy o (liquid) supe cooled wa e (SCW).
Myoglobin was es ed as a well-known s anda d p o ein, o which no ice nuclea ion ac ion is documen ed. As
expec ed, his is e lec ed in he e y low nuclea ion empe a u e (− 20°C) in he imme sion es . The deposi ion
nuclea ion om wa e apo had no been epo ed be o e. We demons a e he p esence o some SCW on he
solid p o ein o a sho ime (sub-seconds), bu in gene al, he nuclea ion appea s o p oceed di ec ly om apo
o solid. Hence, we pos ula e wo c ys alliza ion ou es: F om apo o solid and om apo o SCW o solid.
We chose he comme cial "snowmake " Snomax, which con ains he ice nuclea ion p o ein (INP) InaZ, as es
o a well-cha ac e ized INP. We ep oduced he e y high imme sion nuclea ion empe a u e o − 4.4°C, which
is he base o snow p oduc ion om SCW. We u he demons a ed ha InaZ is also an excellen deposi ion
nuclea o , esul ing in he as deposi ion o ice om wa e apo on o he solid p o ein.
Ou hi d p o ein is he sup amolecula cage apo e i in, which is a su p isingly good INP in he imme sion
mode (− 5.4°C). We ound, howe e , ha he deposi ion mode in ol es he o ma ion o SCW om wa e apo ,
and only a e wa ds c ys alliza ion. Hence, apo e i in is a ela i ely poo deposi ion mode nuclea o .
In compa ison, apo e i in and InaZ di e p ima ily in he deposi ion mode: Apo e i in is a ela i ely poo
INP o deposi ion. I de elops SCW as an in e media e be ween gas and solid phase, i.e., i s , SCW condensa es
om apo , and, wi h some delay, i eezes. This would be a slow analog o he as SCW eezing on myoglobin.
Ano he in e p e a ion is ha he o ma ion o SCW d ople s canno be a oided on apo e i in; he e o e, his
p o ein, al hough being a good INP, canno ollow he di ec apo ou e o nuclea ion o solid ice. Hence, he
wo c ys alliza ion modes a e no co ela ed.
The ESEM obse a ion o SCW also en ailed imaging o SCW d ople s, whene e hey we e p ese ed o
su icien ime (minu es). We p esen a ep oducible me hod o p oduce SCW, based on a (hyd ophobic) pol-
ished and sol en -cleaned coppe su ace, and a wa e apo p essu e jump a a low empe a u e. As expec ed,
SCW d ople s on ou hyd ophilic p o eins show a e y di e en beha io : They ha e an i egula shape, co e
mo e a ea, and ea u e smalle gaps be ween d ople s, i.e., hey nicely we he su ace o a p o ein spo . Once he
c ys alliza ion o SCW on p o eins is igge ed, i is e y as and co e s he comple e mm-sized p o ein spo in
seconds. The ice p opaga es om he im o he spo , whe e i b idges he p o ein spo wi h he coppe su ace.
Ou esul s demons a e ha ESEM is help ul o expe imen s in a b oad ange o wa e /ice- ela ed opics,
speci ically o in es iga ing eezing and sublima ion on he mic oscale. This scale should be expanded o he
nanoscale, hus p o iding new insigh s in o he p ocesses leading o ice cloud o ma ion on small nuclei. In addi-
ion, he ice g ow h on ai plane wings, a c i ically dange ous consequence o SCW condensa ion on la su aces,
migh equi e in es iga ing submic on su ace ea u es. Un o una ely, ex ending he s udies o li ing cells and
biological an i eeze p ocesses, i.e., wo king wi h li ing ma e , is no s aigh o wa d61. Once could ci cum en
his p oblem by employing complex biological su aces such as he cell wall o Pseudomonas sy ingae43, om
which InaZ is isola ed.
Da a a ailabili y
The au ho s decla e ha all da a suppo ing he indings o his s udy a e a ailable wi hin he pape and i s sup-
plemen a y in o ma ion iles.
Recei ed: 21 June 2022; Accep ed: 13 Sep embe 2022
Re e ences
1. Mu phy, D. M. & Koop, T. Re iew o he apou p essu es o ice and supe cooled wa e o a mosphe ic applica ions. Q. J. R.
Me eo ol. Soc. 131, 1539–1565. h ps:// doi. o g/ 10. 1256/ qj. 04. 94 (2005).
2. Fan, J., Leung, L. R., Rosen eld, D. & DeMo , P. J. E ec s o cloud condensa ion nuclei and ice nuclea ing pa icles on p ecipi a ion
p ocesses and supe cooled liquid in mixed-phase o og aphic clouds. A mos. Chem. Phys. 17, 1017–1035. h ps:// doi. o g/ 10. 5194/
acp- 17- 1017- 2017 (2017).
3. K ede , M. J., Al a enga, J., Kim, P. & Aizenbe g, J. Design o an i-icing su aces: Smoo h, ex u ed o slippe y?. Na . Re . Ma e .
1, 15003. h ps:// doi. o g/ 10. 1038/ na e ma s. 2015.3 (2016).
4. Geidoble , R. & Win e , G. Con olled ice nuclea ion in he ield o eeze-d ying: Fundamen als and echnology e iew. Eu . J.
Pha m. Biopha m. 85, 214–222. h ps:// doi. o g/ 10. 1016/j. ejpb. 2013. 04. 014 (2013).
5. Zhuo, Y., Xiao, S., Håkonsen, V., He, J. & Zhang, Z. An i-icing ionogel su aces: Inhibi ing ice nuclea ion, g ow h, and adhesion.
ACS Ma e ials Le e s 2, 616–623. h ps:// doi. o g/ 10. 1021/ acsma e ia lsle . 0c000 94 (2020).
6. Lede le, C., Sa ig, M. & Vogel, M. E ec s o pa ial c ys alliza ion on he dynamics o wa e in mesopo ous silica. J. Phys. Chem.
C 122, 15427–15434. h ps:// doi. o g/ 10. 1021/ acs. jpcc. 8b038 15 (2018).
7. Ce eny, S., Mallamace, F., Swenson, J., Vogel, M. & Xu, L. Con ined wa e as model o supe cooled wa e . Chem. Re . 116,
7608–7625. h ps:// doi. o g/ 10. 1021/ acs. chem e . 5b006 09 (2016).
8. Ce eny, S., Comba o-Palacios, I. & Swenson, J. E idence o coupling be ween he mo ions o wa e and pep ides. J. Phys. Chem.
Le . 7, 4093–4098. h ps:// doi. o g/ 10. 1021/ acs. jpcle . 6b018 64 (2016).