Access to this full-text is provided by Springer Nature.
Content available from Nature Communications
This content is subject to copyright. Terms and conditions apply.
Article https://doi.org/10.1038/s41467-023-36060-8
Biomolecular condensates formed by
designer minimalistic peptides
Avigail Baruch Leshem
1
,SianSloan-Dennison
2
, Tlalit Massarano
1
,
Shavit Ben-David
1
,DuncanGraham
2
, Karen Faulds
2
,HugoE.Gottlieb
3
,
Jordan H. Chill
3
&AyalaLampel
1,4,5,6
Inspired by the role of intracellular liquid-liquid phase separation (LLPS) in
formation of membraneless organelles, there is great interest in developing
dynamic compartments formed by LLPS of intrinsically disordered proteins
(IDPs) or short peptides. However, the molecular mechanisms underlying the
formation of biomolecular condensates have not been fully elucidated, ren-
dering on-demand design of synthetic condensates with tailored physico-
chemical functionalities a significant challenge. To address this need, here we
design a library of LLPS-promoting peptide building blocks composed of
various assembly domains. We show that the LLPS propensity, dynamics, and
encapsulation efficiency of compartments can be tuned by changes to the
peptide composition. Specifically, with the aid of Raman and NMR spectro-
scopy, we show that interactions between arginine and aromatic amino acids
underlie droplet formation, and that both intra- and intermolecular interac-
tions dictate droplet dynamics. The resulting sequence-structure-function
correlation could support the future development of compartments for a
variety of applications.
The emerging field of liquid–liquid phase separation (LLPS), as the
basis of membraneless organelles formation1, has triggered a renewed
interest in intrinsically disordered proteins (IDPs) and the design of
materials based on their remarkable dynamic properties2.Membra-
neless organelles, or biomolecular condensates, are supramolecular
disordered compartments that include stress granules, nucleoli, and
Cajal bodies. The commonly suggested mechanism for the formation
of biomolecular condensates is based on LLPS of IDPs and other bio-
molecules (mainly nucleic acids)3,4,inwhichthecondensates’building
blocks are highly mobile and exchange rapidly with the surrounding
environment. While the exact functionalities of membraneless orga-
nelles are still being studied, a general function shared by different
biomolecular condensates is concentration, condensation, and sto-
rage of proteins, nucleic acids, enzymes and substrates and via this,
control of enzymatic reactions and protection of reaction products4.
Inspired by these remarkable functionalities, researchers have begun
to design dynamic compartments that are formed by LLPS2of IDPs or
polypeptides with disordered domains5–10 for delivery and encapsula-
tion of biomolecules by leveraging intra- and supramolecular order/
disorder11–20. Unlike thermodynamically stable compartments, these
dynamic assemblies7,21–26, can be designed to respond to specific
stimuli10,12,25,27,28, and they allow for control of various properties,
including polarity, rheology, and surface tension. Yet, the exact
molecular mechanisms underlying the formation of biomolecular
condensates have not been fully elucidated, although a number of
advances have been recently made29–31. Thus, the design of biomole-
cular condensates, or liquid droplets, with tunable chemical and phy-
sical functionalities ‘on demand’remains a challenge.
Received: 9 June 2022
Accepted: 13 January 2023
Check for updates
1
ShmunisSchool of Biomedicineand Cancer Research, George S. WiseFaculty of LifeSciences, TelAviv University, Tel Aviv 69978,Israel.
2
Department of Pure
and Applied Chemistry, Technology and Innovation Centre, University of Strathclyde, 99 George Street, Glasgow G1 1RD, UK.
3
Department of Chemistry,
Faculty of Exact Sciences, Bar Ilan University, Ramat Gan 52900, Israel.
4
Center for Nanoscience and Nanotechnology Tel Aviv University, Tel Aviv 69978,
Israel.
5
Sagol Center for Regenerative Biotechnology Tel Aviv University, Tel Aviv 69978, Israel.
6
Center for the Physics and Chemistry of Living Systems Tel
Aviv University, Tel Aviv 69978, Israel, Tel Aviv 69978, Israel. e-mail: Jordan.Chill@biu.ac.il;ayalalampel@tauex.tau.ac.il
Nature Communications | (2023) 14:421 1
1234567890():,;
1234567890():,;
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Experimental molecular-level studies of protein LLPS are per-
formed using recombinant IDPs, which suffer from several limitations.
In particular, IDPs have undetermined structure and their preparation
involves multistep expression in living cells and purification, which in
some cases produce limited yields and require stringent storage con-
ditions. Compared with protein production, peptide synthesis is
straightforward and does not require complex expression/purification
steps, yet IDPs undergo LLPS at lower, nano- or micromolar con-
centrations while peptides typically have higher critical LLPS con-
centrations. Importantly, unlike proteins, changes to composition of
peptides, even at the single-amino acid level, directly dictates the
supramolecular structure and material properties32–35,thereby
enabling to establish sequence-structure and structure-function
relationships.
To gain insights into the driving forces of biomolecular con-
densates formation, here we use systematic sequence variants of a
designer peptide as minimalistic building blocks of synthetic con-
densates. We design a library of LLPS-promoting peptides that self-
coacervate into liquid droplets with tunable chemical and material
properties. Since the analytical methods traditionally used to char-
acterize self-assembled peptides such as X-ray scattering techniques
are limited for solid-like assemblies36, we use complementary meth-
odologies including fluorescence recovery after photobleaching
(FRAP), Raman spectroscopy and nuclear magnetic resonance (NMR)
spectroscopy to shed light on the mechanism of droplet formation,
both at the material- and the molecular-level. Our findings show that
the peptide sequence controls the LLPS propensity and the material
properties of the resulting droplets including mobility and diffusion, as
well as the encapsulation efficiency of fluorescent payloads. Moreover,
these findings show that arginine (Arg) interactions with the side
chains of aromatic amino acids play a key role in LLPS.
Results
Peptide sequence controls LLPS propensity
We sought to design a library of LLPS-promoting peptide building
blocks which form synthetic biomolecular condensates with tunable
chemical composition and physical properties. We hypothesized that
minimalistic variants of protein low complexity domains (LCDs) will
self-assemble into liquid droplets through LLPS. Specifically, we envi-
sioned that in order to keep the sequence length relatively short, the
peptide composition should include high content of aromatic and
basic amino acids that can interact through π–πor cation–πinterac-
tions. To test this hypothesis, we designed a primary sequence that
contains various LLPS-promoting motifs. Inspired by LCDs of ribonu-
cleoproteins (RNPs) and IDPs that are rich in arginine-glycine (RG)
dyad or RGG triad repeats37, we incorporated three repeats of an RG
dyad (Fig. 1a), where glycine (Gly) provides flexibility and arginine
(Arg) promotes electrostatic interactions with the terminal carboxylic
group, cation- πor πinteractions with aromatic amino acids (Fig. 1a).
While both lysine (Lys) and Arg have basic side chain groups, the
guanidine group of Arg delocalizes the charge due to the πbonded
system and thus canpromote more versatile binding modescompared
to Lys, through cation–πand π–πinteractions38. Thus, considering
that the Arg side chain caninteract with side chains of aromatic amino
acids, we incorporated tryptophan (Trp) and tyrosine (Tyr) into the
peptide sequence at a 1:1 stoichiometry with Arg, with aromatic amino
acids positioned at both ends of the sequence to enhance π–πstacking
interactions21. Finally, we considered the elastin-like polypeptide (ELP)
repeating motif VPGXG. This pentapeptide sequence is a common
LLPS-promoting motif used in engineered ordered/disordered poly-
peptides, where X can be any amino acid except proline (Pro), and a
hydrophobic amino acid at this position promotes coacervation39,40.
Moreover, a previous work showed that substituting Val at the first
position with Trp induced coacervation of the 15-mer ELP (WPGVG)
3
41.
Building on these previous findings, we incorporated the sequence
WPGVG, thus obtaining the 14-mer peptide WGRGRGRGWPGVGY
termed WGR-1 (Figs. 1a, 2a). WGR-1 forms droplets by self-
coacervation at neutral pH in the presence of 0.2 M NaCl, which
reduces the electrostatic repulsion of the basic peptide (Fig. 1b). We
created a phase diagram of WGR-1 as a function of pH and peptide
concentration in tris buffer by gradually increasing the pH and mon-
itoring LLPS, as indicated by appearance of sample turbidity (Fig. 2b),
where all experiments performed at room temperature. As expected,
increasing peptide concentration decreased the pH in which visible
turbidity and droplets are observed, where the critical LLPS con-
centration is 8 mM. To confirm that the turbidity is a result of LLPS and
droplet formation rather than aggregation, we used bright field laser
scanning confocal microscopy (Fig. 2c).
To shed light on the role of each domainin LLPS,we designed five
additional sequence variants (Fig. 2a). Analyzing the LLPS propensity
of each sequence variant showed that omitting the Tyr at the
C-terminal position (WGR-2) completely arrests LLPS, as no droplets
were formed at peptide concentration 5–30 mM in the pH range of
3–12 (Fig. 2b). Substituting Tyr with phenylalanine (Phe) (WGR-3)
recovers droplet formation. To study the role of the ELP domain in
droplet formation, we omitted this motif from the peptide sequence
(WGR-4). To our surprise, removing the ELP domain does not inhibit
LLPS, but instead shifts the boundaries of the phase diagram, with
higherpH value at the critical LLPS concentration of 8 mM (pH = 9.5 for
WGR-4 vs. 8.5 for WGR-1). We attribute this to the higher charge den-
sity in WGR-4, with consequent stronger repulsion, when compared to
the other LLPS-forming peptides. Reducing the number of RG dyads
from three to two (WGR-5) shifts the phase diagram boundaries with
droplets formed at 5 mM (Fig. 2b), suggesting that removal of the b asic
Arg decreases the electrostatic repulsion between the peptide mole-
cules and as a result promotes intermolecular interactions and dro-
plets formation. By using solution NMR analysis (see “Experimental”
Section), we found that the pKa of the N-terminal amine group is in the
7.3–7.5 range (Supplementary Table 2), considerably lower than the
expected value in the 8.0–8.5 range. These results might explain the
changes to LLPS observed from the phase diagrams, where neu-
tralization of the terminal amine leads to reduced electrostatic repul-
sion between the peptide molecules, and in turn to LLPS. Strikingly,
substituting all three Arg with Lys completely arrested LLPS with no
turbidity or droplets observed, albeit some aggregates at low abun-
dancy (Fig. 2b, c). While both Lys and Arg can participate in cation-π
interactions, only Arg can form π–πinteractions due to the sp2nitro-
gen atoms. These results suggest that π-interactions between the
guanidium group of Arg side chain and the aromatic amino acid side
chains are critical for LLPS (Fig. 1c).
Since this analysis is highly sensitive to pH fluctuation, we also
created phase diagrams for all peptides using three different buffers
that are optimized for specific pH range: citrate buffer for pH 3–7, tris
buffer for pH 7–9 and ammonium bicarbonate for pH 9–12 (Supple-
mentary Fig. 1). While the trend in these phase diagrams is similar to
that obtained only in tris buffer, the critical pH for LLPS is lower in
citrate buffer for all LLPS-promoting peptides (Supplementary Fig. 1).
In addition, a slight difference is observed between WGR-1 and WGR-3
at 10 mM, where LLPS is observed at higher pH for WGR-3 (Supple-
mentary Fig. 1), suggesting that Tyr has a stronger contribution to the
intermolecular interactions which mediate LLPS than Phe. Indeed,
Wang et. al showed that Tyr–Arg interactions are more significant for
phase separation than Tyr–Lysinteractionsandevenmorethan
Phe–Arg42. Moreover,the critical LLPSconcentration of WGR-4is lower
in ammonium bicarbonate than that in tris buffer (5mM vs. 8 mM,
respectively). Thus, these results suggest that citrate and ammonium
bicarbonate promote LLPS. To shed light on this, we performed tur-
bidity assay of WGR-1 at 10mM using the same conditions used in the
phase diagram analysis. Higher turbidity is observed at pH 7 with
citrate and at pH 9–10 with ammonium bicarbonate compared to tris
Article https://doi.org/10.1038/s41467-023-36060-8
Nature Communications | (2023) 14:421 2
Content courtesy of Springer Nature, terms of use apply. Rights reserved
buffer (Supplementary Fig. 2), confirming that citrate and ammonium
bicarbonate induce peptide LLPS, presumably by reducing the elec-
trostatic repulsion between the peptide molecules provided by their
charge state at the respective pH range (−2/−3forcitrateand−1/−2for
ammonium bicarbonate). In contrast, the charge state of tris (+1/0 at
the respective pH range) is not expected to reduce this repulsion.
Following theseobservations, we sought to systematicallyanalyze
the effect of ions from the Hofmeister series on LLPS in our mini-
malistic system. For this, we performed turbidity analysis of WGR-1 at
10 mM in the presence of four different salts that are composed of
chaotropic and kosmotropic anions and cations: NaCl, KCl, Na
2
HPO
4
,
and K
2
HPO
4
. We measured sample turbidity at salt concentrations
between 10 mM and 200 mM and at three different pH values (6, 7, and
8). When HPO
4
2−is used as an anion, sample turbidity appears at pH 8
and no difference in turbidity is observed between K+and Na+at
concentrations up to 100 mM (Supplementary Fig. 3), which is
expected as Hofmeister cations have typically a smaller effect on LLPS
and K+and Na+are adjacent in the series. At 200 mM, lower sample
turbidity is observed for K
2
HPO
4
compared with Na
2
HPO
4
.Thisresult
correlates with previous reports on the stabilizing effect of K+on
proteins at high mM concentrations43. Interestingly, with Cl−as
the anion, Na+induces LLPS at pH 8 while K+does not, further showing
the stabilizing effect of K+on the peptide43. Moreover, at pH 6, high
turbidity is observed for both KCl and NaCl at low salt concentrations
(10 and 50mM). In a recent work, Knowles and co-workers proposed
that LLPS occurs at low pH is mediated by electrostatic interactions,
while thatoccurs at basic pH is mediated by hydrophobic interactions
through a salting-out process44. Similarly, our results suggest that at
pH 6, low Cl−concentration promotes LLPS by reducing the repulsion
between the basic peptide groups, whileat pH 8, high concentration of
NaCl (but not KCl) induces salting-out of the peptide molecules, where
the latter undergo LLPS through various modes of interactions,
including π–interactions.
Next, we studied whether the peptide sequence affects the
material properties of the resulting liquid droplets. For this, we per-
formed FRAP analysis using laser scanning confocal microscopy of
0.5% FITC-labeled peptides. At this concentration, the dye has a neg-
ligible effect on peptide LLPS (Supplementary Fig. 4). The apparent
diffusion coefficient for each peptide was calculated as
Dapp =r2
tð1Þ
where tis the recovery time. The calculated apparent diffusion coef-
ficients of the peptides were in the 1.7–5.5*10−14 m2s−1range (Supple-
mentary Table 1). Similar values were reported for condensates that
Fig. 1 | Designer minimalistic peptide droplets. a Chemical structure of WGR-1.
Aromaticamino acid side chains (Trp and Tyr) arecolored in blue, Arg side chain is
colored in turquoise, and non-polar amino acid side chains that are partof the ELP
domain (Pro-Gly-Val-Gly) are colored in orange. bSuggested mechanism of the
peptide liquid droplet formation and subsequent partitioning of fluorescent
payloads. cExpected intermolecular interactions underlying LLPS of WGR-1 into
liquiddroplets, including (from leftto right) Trp-Trp,Trp-Tyr, Arg-Trp,and Arg-Tyr
π-πstacking. Chemical structures of side chains are presented, color coded as
described in (a).
Article https://doi.org/10.1038/s41467-023-36060-8
Nature Communications | (2023) 14:421 3
Content courtesy of Springer Nature, terms of use apply. Rights reserved
are formed by complex coacervation of cationic peptide polymers and
nucleic acids38
Out of the four LLPS-promoting peptides, WGR-3 (substitution of
Tyr with Phe) has the largest apparent diffusion coefficient (D), more
than 5-fold larger than that of WGR-5 and slightly larger than that of
WGR-1 (Fig. 3and Supplementary Table. 1). The higher Dof WGR-3
compared to WGR-1 correlates with the LLPS propensity of the two
peptides (Supplementary Fig. 1), suggesting that the higher mobility of
WGR-3 is a result of weaker interactions between Phe and Arg com-
pared to those of Tyr and Arg42,45. WGR-4 has a lower diffusion coef-
ficient than WGR-1, suggestingthat the ELP domain interferes sterically
with the interactions between the aromatic amino acid side chains, or
between the aromatics and Arg, and thus, removing this domain might
increase the accessibility of the aromatics and Arg groups. Since the
ELP sequence increases the flexibility of the overall peptide, we
hypothesized that it could facilitate intra-molecular interactions, with
the presence of the bend-promoting Pro residue further enhancing
this tendency. Such intra-peptide contacts—if formed—might compete
with inter-peptide contacts necessary for LLPS and in turn affect dro-
plet dynamics. By using solution NMR, we followed changes in 13Cα
chemical shifts at low non-LLPS concentrations (3 mM) in which inter-
peptide contacts are less likely to occur, upon addition of 8 M urea,
expected to perturbintra-molecular contacts. Urea-induced 13Cαshifts
of residues Pro10 and Val12 are consistent with an increase in random
coil conformation and a decrease in turn conformation in WGR-1,
WGR-3, and WGR-5, but not in WGR-2, lacking the aromatic residue
required for intra-peptide interactions (SupplementaryTable 3). Shifts
of other residues (i.e., Trp1) do not exhibit this difference. Notably, one
of our findings distinguishes WGR-3 from all other peptides as striking
urea-induced 13Cshifts(inthe0.2–0.5 ppm range) for Trp9Cα,Cβand
aromatic Cγwere observed only in WGR-3 spectra. These findings
suggest that the ELP domain induces intramolecular interactions, yet
LLPS is obviously influenced by many different factors. The lowest
diffusion of WGR-5 indicates that decreasing the electrostatic repul-
sion by reducing the net charge of the peptide from +3 to +2 increases
the strength of intermolecular interactions between the peptide
building blocks and as a result, significantly lowers the mobility and
dynamics of the droplets (Fig. 3a–c). Moreover, these results suggest
that in the absence of the aliphatic ELP domain, cation–πor π–π
interactionsbetween Arg and the aromatic side chain are the dominant
driving force for droplet formation. As these interactions are short-
range, they can result in higher friction between the peptides mole-
cules, and in turn, reduced peptide diffusion and droplet dynamics37,38.
Peptide diffusion in the dilute phase, as calculated by solution NMR
analysis, is in the 1.9–2.4*10−10 m2s−1range (Supplementary Table 1).
Droplet encapsulation efficiency is influenced by peptide
hydrophobicity
Next, we studied how peptide composition affects the encapsulation
efficiency of the droplets by using GFP, rhodamine B, and fluorescein
as fluorescent payload model systems (Fig. 4). We analyzed the
encapsulation efficiency of the fluorescent payloads by using both
confocal microscopy and absorbance measurements of the payloads
in the dilute vs. the condensed phase (Fig. 4a–c). The encapsulation
efficiency (EE) of the fluorescent payloads ranges between 72 and 99%
(Fig. 4d), where the hydrophobic Phe-containing peptide WGR-3 has
thelowestEEofGFP,andthemostpolarpeptide,WGR-4,hasthe
highest EE of GFP and of rhodamine B, suggesting that the peptide
interacts with the dye either electrostatically, by π–π, or cation–π
interactions.WGR-1 has the highest EE of fluorescein. In addition to π-π
interactions, WGR-1 might form hydrogen bonding with the hydroxyl
groups or electrostatic interactions with the deprotonated carboxylic
acid of the dye. Thus, these results demonstrate that the encapsulation
efficiency of the compartments can be modulated by the chemical
composition of the peptide building blocks.
π-πinteractions and hydrogen bonding underlie peptide dro-
plet formation
To gain molecular-level insights into the network of intermolecular
interactions which underlie droplet formation, we performed
extensive Raman and NMR spectroscopy analyses. We envisioned
that employing these complementary techniques, which provide
information on molecular interactions of assemblies in the solid-
state (Raman spectroscopy) and in solution (NMR), might facilitate a
c
Notation Sequence
WGR-1 WGRGRGRGWPGVGY
WGR-2 WGRGRGRGWPGVG
WGR-3 WGRGRGRGWPGVGF
Notation Sequence
WGR-4 WGRGRGRGWY
WGR-5 WGRGRGWPGVGY
WGK WGKGKGKGWPGVGY
a
WGR-1
LLPS LLPS
No LLPS
WGR-2 WGR-3
LLPS
WGR-4
LLPS
WGR-5 WGK
No LLPS
b
Fig. 2 | Peptide sequence controls LLPS propensity and droplet formation.
aTable of the designed peptide sequences. bPhase diagram of the peptides as a
function of peptide concentration and pH, in tris buffer with 0.2M NaCl, at room
temperature. LLPS was not observed for WGR-2 and WGK. Data are presented as
mean values of n=3+/−SD. Source data are provided as a Source Data file.
cConfocal microscopy bright field images of peptide liquid droplets formed at a
concentration of 20mM in Tris buffer at pH 8 with 0.2 M NaCl. Scale bars =20 μm.
Inset: macroscopic images of the peptide solutions. ‘Source data are provided as a
Source Data file.
Article https://doi.org/10.1038/s41467-023-36060-8
Nature Communications | (2023) 14:421 4
Content courtesy of Springer Nature, terms of use apply. Rights reserved
more holistic understanding of the mechanism that governs droplet
formation.
To analyze the peptide droplets by Raman spectroscopy, WGR-1
droplets were drop-casted on a precoated glass substrate (see
“Experimental”section), and solution Raman spectra of the droplets
were collected. An average solution Raman spectrum of WGR-1 droplet
sample is shown in Fig. 5a. Upon analysis of the representative spec-
trum, we found that this and other significant peaks originated from
the Trp side chains. Most of the peaks are associated with C–Hbend-
ing, ring-stretching, and deformation in the indole ring (1551, 1010,
876, and 758 cm−1). The peaks at 1433 and 1615 cm−1are attributed to
the symmetric and asymmetric stretching of the COO−group and the
1573 cm−1,relatedtoNH
3
+vibrations46. These structural markers can be
used to assign the interactions which underlie droplet formation. It has
been previously shown that the Raman bands at 1551, 1358, and
1010 cm−1are strongest when the indole ring is hydrogen bonded47.
These three peaks are very prominent in the peptide droplet spectra
suggesting that hydrogen bonding is a critical interaction for droplet
formation. The 876 cm−1peak is an indole ring vibration mode asso-
ciated with a displacement of the N
1
H group nearly along the N
1
–H
bondwhichdecreasesuponhydrogenbonding
46.Inourspectra,this
peak is weak, further evidence for the involvement of hydrogen
bonding in droplet formation. Finally, slight shifting of the 1010 cm−1
band to the 1009–1010 cm−1range suggests a loss of van der Waals
interactions within the droplet. We also observe a doublet (850/830)
that originates from Tyr side chain. Conflicting explanations of the
850/830 ratio of peaks were previously reported48,49 and thus its
interpretation in our system is not obvious, yet it is clearly sensitive to
the hydrophobicity of the phenol environment49,50.
Next, we sought to analyze individual droplets by using Raman
mapping. Droplets in solution cannotbe mapped due to their mobility,
thus we dried the drop-casted droplets, mapped them, and created
false color 2D and 3D images (Fig. 5b, c) by plotting the intensity of the
758 cm−1peak throughout the imaged area, in the center and edge of
the droplet as well as in the surrounding phase. Similar spectra were
obtained from dried droplets (Supplementary Fig. 5) compared with
solution droplets (Fig. 5a), suggesting that drying the droplet did not
significantly alter the interactions that mediate droplet formation. Yet,
we did lose some information from the Tyr doublet, which is very weak
in the dried spectrum. Within the droplet itself, we see some differ-
ences in the Raman spectra throughout the mapped area. Notably,
spectral differences in the 700–800 cm−1and 1300–1500 cm−1regions
were observed between the center (white) and the edge (red) of the
droplet, as shown in the normalized Raman spectra (Fig. 5b, d). The
intensity ratio between the 1360/1340 cm−1peaksishigheratthe
droplet center when compared to its edge, an indicator of increased
hydrophobicity within the condensed phase. We attribute the weaker
spectrum at the droplet edge in the 1300–1400 cm−1region to focus-
ing. No peptide signal is observed in the spectrum of the surrounding
phase. The main difference in the 700–800 region is the shift in the
758 cm−1peak from the droplet center, to 765 cm−1at the edge. A
similar shift in the relative intensity of 759 cm−1was previously attrib-
uted to cation–πinteractions of the model compound, diaza crown
ether with two indole substituents51. Thus, the shift observed between
the center and the edge of the droplet indicates the involvement of Trp
in cation–πinteractions.
To further confirm the role of Trp in LLPS, we designed and stu-
died 4 additional sequence variants, where we omitted Trp at position
1 (WGR-6), omitted Trp at position 9 (WGR-7), substituted Trp at
position 9 with Ala (WGR-8), and both omitted Trp at position 1
and substituted Trp at position 9 with Ala (WGR-9). None of the pep-
tides undergoes LLPS at all conditions tested (concentrations up to
30 mM and at pH 3–11). Optical microscopy analysis of the peptides at
30 mM shows clear solutions and some amorphous aggregates
WGR-1
WGR-5
WGR-4
WGR-3
Pre-bleaching Bleaching 50 sec 100 sec
ab
d
c
Fig. 3 | Strength of intermolecular interactions affects peptide droplet
dynamics. a–cFRAP analysis of WGR-1, WGR-3, WGR-4, and WGR-5, performed
usinglaser scanningconfocal microscopy at 20 mMin Tris buffer at pH8 with 0.2 M
NaCl using 0.5% FITC-labeled peptides. aRepresentative confocal microscopy
images of FRAP for individualdroplets. Scale bars =5 μm. FRAP recovery plots (b),
apparent diffusion coefficient (c),and t
1/2
of the recovery (d). Data arepresented as
mean values +/−SD, n= 5 (WGR-1), 4 (WGR-3, WGR-4), and 5 (WGR-5). Source data
are provided as a Source Data file.
Article https://doi.org/10.1038/s41467-023-36060-8
Nature Communications | (2023) 14:421 5
Content courtesy of Springer Nature, terms of use apply. Rights reserved
(Supplementary Fig. 6). These results strengthen the findings from the
Raman spectroscopy and highlight the critical role of Trp side chains
in LLPS.
NMR points to the molecular mechanism of peptide droplet
formation
For a solution view of this system, we employed NMR, well-known for
its unique ability to provide information on the conformations of low-
complexity disordered peptides and their motions on a wide range of
timescales52–55. A combination of homonuclear 2D-1H,1H-COSY/TOCSY
and heteronuclear 2D-1H,13C-HMQC/HMBC spectra acquired for
20 mM WGR-1 at pH 6 without NaCl (non-LLPS conditions) afforded an
assignment of 1H/13C chemical shifts of this peptide. Using a similar
array of experiments, the vast majority of these assignments could
then be transferred and reassigned in peptides at various pH values
and NaCl concentrations. Observable chemical shifts for monomeric
WGR-1 under all conditions (see Supplementary Tables 4, 5 and Sup-
plementary Fig. 9) were consistent with a disordered peptide in ran-
dom coil conformation, in agreement with the results from the CD
analysis (Supplementary Fig. 7). Towards the peptide C-terminus side
chains exhibited a double (major/minor) set of shifts with a typical
intensity ratio of ~2.2:1, attributed to trans/cis isomers of the P10
pyrrolidine ring.
Although slowly tumbling peptides in the condensed phase are
typically inaccessible to high-resolution NMR,equilibrium between the
two phases results in chemical shift changes for the bulk phase,
reporting on molecular changes induced by droplet formation and
identifying intermolecular interactions contributing to this process. In
doing so we focused on the two known LLPS-inducing factors for this
system, salinity and pH. Since such changes are inherently small, we
based our analysis of shifts under LLPS-promoting conditions on a
comparison to non-LLPS-promoting conditions.
13C resonance frequencies were followed for 5 and 20 mM WGR-1
samples at pH 8 with increasing NaCl concentrations (Supplementary
Data 1). Whereas the 5 mM sample remains translucent throughout the
titration (non-LLPS conditions), the 20 mM sample exhibits
coacervation at higher concentrations (LLPS conditions). Thus, the
difference in NaCl-induced spectral changes between the two samples
(arbitrarily chosen at0 and 0.1 M) is an indication of the effects of LLPS
(Fig. 5e). The six Gly (lacking side chains) were omitted from this
analysis, and due to spectral overlap resonances of the three Arg side
chains were grouped together. While side chain-specific NaCl-induced
changes were in the 1.4–2.5 Hz range for the 5 mM sample (with the
exception of W1, due to slight changes in the pKa of the teminal NH
3
+
group), they were significantly larger, in the 5–7 Hz range, for the
20 mM sample. The differential change averaged 3.5 Hz and was rela-
tively uniform throughout the WGR-1 sequence, with the largest dif-
ference observed for residue P10 (~5.5 Hz). Since these changes reflect
the indirect effect of droplet environment upon the NMR-visible bulk
peptide, they confirm the Raman results regarding the LLPS-induced
environmental change and suggest a global effect upon the peptide.
We then turned to examine the effects of pH increase as an
inducer of LLPS. As expected, higher pH-induced changes in chemical
shifts clustered around ionizable groups of the peptide at the amino
terminus (pKa ~ 7.5,Supplementary Table 2) and the Tyr14 phenolic ring
(pKa ~ 10). In contrast, smaller yet still significant differences observed
between shifts at pH 10 and concentrations of 5 (non-LLPS conditions)
and 20 mM (LLPS conditions, Fig. 6a, b and Supplementary Fig. 8) wer e
instructive in pointing to molecular changes accompanying droplet
formation. The most significant differences (between 5 and 20 mM
samples at pH 10) observed in the 13C NMR spectrum (>0.1 ppm) were
for the aromatic ring 13C nuclei of Tyr14 side chain (Fig. 6a). Specifically,
the changes seen for the γ-andε-13C (but not the δ-13C) suggest an
electron-donating effect of the anionic phenol group, a hypothesis
consistent with the tangible change in the ζ-13C resonance (Fig. 6a).
These findingsimplicate theTyr14 aromaticring as a key determinant of
coacervation. Raman-observed changes involving the Hε1–Nε1bond in
the Trp indole rings are undetectable by NMR under basic conditions,
and generally relatively small chemical shifts were observed for the Trp
carbons.
Reasoning that an intermolecular interaction is necessary for
coacervation to become favorable, a likely candidate for this
WGR-1 WGR-5WGR-4WGR-3
a
b
c
de
fg
Fig. 4 | Partitioning of fluorescent payloads within droplets depends on pep-
tide pola rity. Confocal microscopyimages of (a)fluorescein, (b)Rhodamine B and
(c) GFP partitioning within WGR-1, WGR-3, WGR-4, and WGR-5 peptide droplets.
Scale bar = 10 μm. Insets: Macroscopic imagesof the partitioning of the fluorescent
payload in peptide droplets samples before (left) and after (right) centrifugation
and droplet sedimentation. Encapsulation efficiency (EE) analysis of peptide dro-
plets calculatedfrom absorbancemeasurements of bulksolutions for (d)WGR-1(e)
WGR-3 (f)WGR-4(g) WGR-5. Data are presented as mean of n=3+/−SD. Source
data are provided as a Source Data file.
Article https://doi.org/10.1038/s41467-023-36060-8
Nature Communications | (2023) 14:421 6
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Fig. 5 | Molecular level analysis of droplet formation by using Raman and NMR
spectroscopy. a Raman spectrum obtained from averaging solution Raman map.
bNormalized Raman spectra taken from 3 different spots of the 2D false color
image: droplet center (black), droplet edge (red) and the droplet surrounding
(light grey). The large peak at 1100 cm−1originates from the glass background.
cFalse color 3D image showing whole peptide droplet and slice through the
center. dFalse color 2D image created usingthe 758 cm−1peak. . One-dimensional
13C spectrum of WGR-1 peptide with (red) or without (black) 100 mM NaCl, at
5 mM peptide (non-LLPS, lower spectra) and 20mM peptide (upper spectra) in
50 mM tris buffer pH 8 and 300 K, chemical shifts are assigned. Inset: averagedΔδ
in 13C spectra of the four samples for each amino acid at peptide concentration of
5 mM (left bar chart) and 20mM (right bar chart). Gly were excluded from the
analysis due to spectral overlap, and all three Arg were averaged due to partial
spectral overlap. Data are presented as mean values +/−SD, n= 8 (W1), 8 (R), 11
(20 mM W9), 10 (5 mM W9), 5 (20 mM P), 4 (5 mM P), 4 (V), 7 (Y). Source data are
provided as a Source Data file.
Article https://doi.org/10.1038/s41467-023-36060-8
Nature Communications | (2023) 14:421 7
Content courtesy of Springer Nature, terms of use apply. Rights reserved
intermolecular contact is an Arg sidechain whose positively charged π-
system is known for its ability to interact with electron-rich aromatic
rings, such as Tyr14. This would explain the loss of LLPS in the WGK
peptide (Fig. 2b), in which all three Arg at position 3, 5, and 7 were
replaced by Lys that aresimilarly positively charged yet unsuitable for
π–πstacking interactions. We assumed that such an interaction must
cause chemical shift perturbations in a second site along the peptide,
and to this end focused upon chemical shifts of Arg Hδnuclei, located
closest to the guanidino π-electron system and best detected using
their COSY cross-peak with the neighboring Hγprotons located in a
distinguishable spectral region. While Arg3and Arg5cross-peaks were
mostly unaffected by the concentration increase from 5 to20 mM, the
third—representing Arg7as determined from our assignment—exhib-
ited a concentration-induced change (Fig. 6c and Supplementary
Fig. 10).We conclude that the key intermolecular contact points in the
formation of droplets for WGR-1 are the sidechains of Arg7and Tyr14,
with effects upon all residues in this segment. A similarpattern ofArg7
chemical shift changes was observed for WGR-3, in which Tyr14 is
replaced by Phe14 (Supplementary Fig. 11). Thus, the two com-
plementary techniques allowed the detection of different inter-
molecular interactions that drive droplet formation, where the solid-
state analysis suggests the involvement of Trp hydrogen bonding and
π–πinteractions, and the solution-state analysis indicates the role of
Tyr/Phe and Arg π-interactions.
Discussion
We have developed a library of LLPS-promoting peptide building
blocks that form synthetic biomolecular condensates with varying
chemical composition and biophysical properties. Our findings show
that the peptide chemical composition directly affect LLPS propensity
and droplet formation, even at a single amino acid level. We show that
the material properties of the droplets canbe tuned by changes to the
peptide sequence, where electrostatic repulsion, steric hindrance, and
specific intra- and intermolecular interactions directly affects peptide
diffusion. Specifically, our findings suggest that intramolecular
contacts between Tyr/Phe and Arg side, induced by the ELP domain,
might compete with intermolecular interactions in the condensed
droplet phase, resulting in accelerated diffusivity. In turn, these
sequence changes can be applied to tune the encapsulation efficiency
of designed biomolecular condensates. To gain molecular level
understanding of the peptide-peptide interactions that underly dro-
plet formation, we combined solid-state analysis of droplets by Raman
spectroscopy and NMR solution-state analysis. We found that Trp side
chains participate in intermolecular interactions within the droplet
center (Raman spectroscopy), and that the interaction between Tyr/
Phe and Arg is crucial for droplet formation (NMR analysis). The latter
finding emphasizes previous evidence of the critical role of Arg (rather
than Lys) interactions with aromatic side chains in formation of
cellular42 and lab-based38 biomolecular condensates. To summarize,
this work demonstrates that minimalistic designer peptides are
attractive building blocks for biomolecular condensates with tuneable
material properties. This approach opens tremendous opportunities
to further develop tuneable and customizable peptide biomolecular
condensates as delivery and microreactor systems.
Methods
Materials
Peptides were custom synthesized, then purified by high performance
liquid chromatography to 95% and supplied as lyophilized powders by
Genscript, Hong Kong. Unless otherwise specified, all reagents were of
the highest available purity. Fluorescein, rhodamine B and ammonium
bicarbonatewere purchasedfrom Holland Moran. NaCl, NaOHand HCl
were purchased from BioLab, Trizma base was purchased from Sigma.
Citrate and citric acid were purchased from Tzamal. GFP (Abcam) was
purchased from Zotal as a solution of 1mg/ml in 0.316% Tris HCl, 10%
glycerol at pH 8 that was aliquoted and stored at −20 °C until use.
Phase diagrams
150 μl of 5 mM, 8 mM, 10 mM, 15 mM, 20 mM, and 30mM peptide
solutions were prepared in either 20 mM tris buffer or 20 mM of the
c
R3/R7
R5
a
ΗΟ
ζ
ε
δ
γ
Cζ
Cδ
Cγ
Cε
b
Fig. 6 | NMR determinesthe molecular mechanism of droplet formation. a The
one-dimensional13C spectrum for the WGR-1peptide at 20 (black, LLPS)and 5 (red,
non-LLPS) mM in 50 mM tr is buffer pH 10 and 300 K. Arrows indicate chemical shift
differences at the Y14 aromatic ring. bAromatic region of the 2D-1H,13C-HMBC
spectrum showinglong-rangeproton-carboncorrelationsallowing the detectionof
quaternary carbons. Y14 chemical shift changes are shown as before. cRegion of the
2D-1H, 1H-COSY spectrum showing the correlationbetween arginineHγ–Hδprotons.
Article https://doi.org/10.1038/s41467-023-36060-8
Nature Communications | (2023) 14:421 8
Content courtesy of Springer Nature, terms of use apply. Rights reserved
following three buffers: citrate buffer for pH 3–7trisbufferforpH7–9
and ammonium bicarbonate for pH 9–12,with0.2MNaCl.ThepHwas
increased gradually until a turbidity appeared and measured as
described below. All measurements were performed at room tem-
perature. Data points represent averages of three independent mea-
surements. Turbidity of 35 μl samples was estimated in triplicates fr om
sample absorbance at λ= 500 nm as described below.
Turbidity measurements
150 μl of 20 mM peptide solutions were prepared in 20 mM Tris buffer.
The pH was adjusted to the desired value of 6, 7, 8, 9, 10, and 11 then
the turbidity of 35 μl was measured in triplicates at λ= 500 nm using
a BioTek H1 synergy plate reader (purchased from Lumitron, Israel).
Secondary structure evaluation by Circular Dichroism (CD)
Samples solutions for CD were prepared at concentration of 1 mM in
20 mM Tris buffer solution with and without 0.2 M NaCl and were
placed in a 0.1 mm path length quartzcuvette at 25 °C, and the range of
190–260 nm was recorded on a Chirascan spectrometer. Background
(buffer with or without NaCl according to the sample) was subtracted
from the CD spectra.
Imaging
All samples were imaged in a 96-well Black Glass bottom plate, glass
1.5H (produced by Hangzhou Xinyou, and purchased from Danyel
Biotech) 10 min after preparation. The images were taken by Zeiss Zen
900 confocal microscope with ×20/0.8 NA Plan- Apochromat air
objective. Images were collected and processed using Zen software
(Zeiss). The light microscopy images were taken at PMT mode. PMT
imaging were taken with 561 nm laser and fluorescence imaging were
takenwith488,561,and405nmlasersforfluorescein, rhodamine B
and GFP, respectively.
Fluorescence recovery after photobleaching
FRAP experiments were performed by a Zeiss Zen 900 confocal
microscope with ×20/0.8 NA Plan- Apochromat air objective. For each
of the peptides, out of a total peptide concentration of 20 mM we used
0.5% FITC-labeled peptide, in tris buffer pH 8 with 0.2 M NaCl. A cir-
cular area with radius of 2.5 μm was bleached with a 488 nm laser 100%
intensity at 10 iterations; subsequent recovery of the bleached area
was recorded with a 488 nm laser. Monitoring of fluorescence recovery
in the condensates was analyzed using Zen Blue 3.2 software (Zeiss).
Photobleaching correction and recovery time were calculated using
OriginLab 9.95. The final FRAP recovery curve is the average of
recovery curves collected from n=4–6 separate droplets.
Encapsulation efficiency
Stock solutions (1 mM) of the fluorescein and rhodamine B dye mole-
cules were prepared in 20mM Tris buffer. Coacervate solutions of
20 mM peptide were prepared in 20 mM Tris + 0.2 M NaCl at pH 8.
From these, a volume of 148.5 µl was mixed with 1.5 µl of dye solution in
a 1.5 ml Eppendorf tube and pipetted. After 10 min the samples were
centrifuged at 15,000 × gfor 10 min. A volume of 120 μlfromthe
supernatant was collected and vortexed and then the absorbance of
35 μl triplicates was measured (at λ= 490 nm for fluorescein and
λ= 555 nm for RhB) in a 384 well black plate by Biotek H1 synergy plate
reader (purchased from Lumitron, Israel). For GFP, a 7.5µlof20mM
Tris buffer to a 7.5 μl GFP aliquot from the purchased stock solution.
The 15 µl of GFP solution was added to 135 µl of peptide solution and
was mixed in a 1.5 mlEppendorf tube and pipetted. The concentration
of GFP at the supernatant was measured via fluorescence. All experi-
ments were performed in triplicate. The concentration of the super-
natant solutions determinate by calibration curves. Imaging was made
to 30 μl of uncentrifuged samples. Efficiency of encapsulation
(%EE) was calculated using Eq. 2.
%EE = CTCsup
CT
ð2Þ
Raman spectroscopy
Samples of 10 mM WGR-1 in 20 mM Tris + 0.2 M NaCl peptide dro-
plets were prepared at pH 8 and analyzed using Raman spectro-
scopy in solution and when dried. To analyze the droplets in
solution, 5 μL of the solution was placed onto a glassslide which was
precoated with sigmacote. Average R aman spectra was obtained by
focusing on the surface of the liquid droplet using a 20× magnifi-
cation objective lens and Raman mapping an area with a step size of
20 μm, using a 532 nm laser excitation with 8 mW laser power and a
10 s integration time. To analyze a dried droplet, 1 μLofthesolution
was drop-casted onto an uncoated glass slide. Excess buffer was
removed by washing the dried droplet with water and again left to
dry. The droplet was focused on using a 100× magnification
objective and Raman mapped with a step size of 0.5 μm, using a
532 nm laser excitation with a laser power of 8 mW. All data was
collected and analyzed using WiRE 4.2 software. MATLAB_R2019b
wasusedformapprocessing.
NMR
NMR samples were prepared by dissolving peptides in ca.500µlof
20 mM Tris buffer and 5% 2H
2
O supplemented with the appropriate
concentration (0–0.2 M) of NaCl and adjusted to the desired pH
using dilute HCl or NaOH. Measurements were conducted on a
DRX700 Avance-III Bruker spectrometer using a cryogenic triple-
resonance TCI or RT-TXI probe-head equipped with z-axis pulsed
field gradients. All spectra were acquired at a field of 16.4 (700.45 and
176.12 MHz for 1Hand13C nuclei, respectively) and at 300 K unless
otherwise indicated. Chemical shift assignment was performed using
data from one- and two-dimensional NMR experiments run with
standard Bruker library files and acquisition parameters, including 1D
1H, 1D-13C, 2D-1H,1H-COSY, 2D-1H,1H-TOCSY, 2D-1H,1H ROESY, 2D-13C,1H
HMQC (set to 1Jcorrelations) and 2D-13C,1H HMBC (set to 2,3Jcorre-
lations) experiments. Typical mixing times for TOCSY and ROESY
experiments were 150 and 200 ms, respectively. Spectra were pro-
cessed and visualized using the Bruker TopSpin 3.6 software suite. To
obtain differential shift changes, following the assignment of peaks
for 5 and 20 mM WGR-1 at 0 and 100 mM NaCl (a total of four 1D-13C
spectra) Δδ values were calculated using the equation Δδ =δ
100
–δ
0
,
where δ
100
and δ
0
are the chemical shifts at 100 and 0 mM NaCl,
respectively. The average Δδ was then calculated by averaging Δδ
values for all non-overlapping 13C nuclei in each amino acid. Gly
residues were eliminated from the analysis due to overlaps in the
spectrum, and the three Arg residues were pooled together due to
partial overlaps in the spectrum. The W1α,β,andγcarbons were
excluded from the analysis due to thelocal NaCl-induced effect upon
the pKa of the N-terminal NH
3
+. For amino acids exhibiting two peaks
due to the P10 cis-trans equilibrium the aggregate Δδ was a 0.7:0.3
weighted average of values measured for the major and minor peaks.
Uncertainties were determined by the spectral resolution and
adjusted to account for ambiguous assignments in the case of
Trp1/Trp9.
Statistics & reproducibility
No statistical method was used to predetermine sample size. No data
were excluded from the analyses. The experiments were not rando-
mized. The investigators were not blinded to allocation during
experiments and outcome assessment. Microscopy images represent
at least three independent analyses.
Article https://doi.org/10.1038/s41467-023-36060-8
Nature Communications | (2023) 14:421 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Reporting summary
Further information on research design is available in the Nature
Portfolio Reporting Summary linked to this article.
Data availability
All data generated or analysed during this study are included in this
published article (and its supplementary information files). Source
data are provided with this paper.
References
1. Shin,Y.&Brangwynne,C.P.Liquidphasecondensationincell
physiology and disease. Science 357, eaaf4382 (2017).
2. Bracha, D., Walls, M. T. & Brangwynne, C. P. Probing and engi-
neering liquid-phase organelles. Nat. Biotechnol. 37,
1435–1445 (2019).
3. Priftis, D., Laugel, N. & Tirrell, M. Thermodynamic characterization of
polypeptide complex coacervation. Langmuir 28,
15947–15957 (2012).
4. Banani, S. F., Lee, H. O., Hyman, A. A. & Rosen, M. K. Biomolecular
condensates: organizers of cellular biochemistry. Nat. Rev. Mol.
Cell Biol. 18,285–298 (2017).
5. Blocher, W. C. & Perry, S. L. Complex coacervate-based materials
for biomedicine. Wiley Interdisciplinary Rev.-Nanomed. Nanobio-
technol. 9,e1442(2017).
6. Aumiller, W. M. & Keating, C. D. Phosphorylation-mediated RNA/
peptide complex coacervation as a model for intracellular liquid
organelles. Nat. Chem. 8,129–137 (2016).
7. Faltova, L., Kuffner, A. M., Hondele, M., Weis, K. & Arosio, P. Multi-
functional protein materials and microreactors using low com-
plexity domains as molecular adhesives. ACS Nano 12,
9991–9999 (2018).
8. Abbas, M., Lipiński, W. P., Wang, J. & Spruijt, E. Peptide-based
coacervates as biomimetic protocells. Chem.Soc.Rev.50,
3690–3705 (2021).
9. Roberts, S. et al. Injectable tissue integrating networks from
recombinant polypeptides with tunable order. Nat. Mater. 17,
1154 (2018).
10. Mohammadi, P. et al. Controllable coacervation of recombinantly
produced spider silk protein using kosmotropic salts. J. Colloid
Interface Sci. 560,149–160 (2020).
11. Srinivasan, N. & Kumar, S. Ordered and disordered proteins as
nanomaterial building blocks. Wiley Interdiscip. Rev.: Nanomed.
Nanobiotechnol. 4,204–218 (2012).
12. Lampel, A. Biology-inspired supramolecular peptide systems.
Chem 6, 1222–1236 (2020).
13. Simon,J.R.,Carroll,N.J.,Rubinstein,M.,Chilkoti,A.&López,G.P.
Programming molecular self-assembly of intrinsically disordered
proteins containing sequences of low complexity. Nat. Chem. 9,
509 (2017).
14. Dzuricky, M., Rogers, B. A., Shahid, A., Cremer, P. S. & Chilkoti, A. De
novo engineering of intracellular condensates using artificial dis-
ordered proteins. Nat. Chem. 12, 814 (2020).
15. Black, K. A. et al. Protein encapsulation via polypeptide complex
coacervation. ACS Macro Lett. 3,1088–1091 (2014).
16. Aumiller, W. M. Jr & Keating, C. D. Phosphorylation-mediated RNA/
peptide complex coacervation as a model for intracellular liquid
organelles. Nat. Chem. 8,129(2016).
17. Tang,T.-Y.D.,Antognozzi,M.,Vicary,J.A.,Perriman,A.W.&Mann,
S. Small-molecule uptake in membrane-free peptide/nucleotide
protocells. Soft Matter 9, 7647–7656 (2013).
18. Martin, N., Li, M. & Mann, S. Selective uptake and refolding of
globular proteins in coacervate microdroplets. Langmuir 32,
5881–5889 (2016).
19. Sun, Y. et al. Phase-separating peptides for direct cytosolic delivery
and redox-activated release of macromolecular therapeutics. Nat.
Chem. 14,274–283 (2022).
20. Capasso Palmiero, U. et al. Programmable zwitterionic droplets as
biomolecular sorters and model of membraneless organelles. Adv.
Mater. 34, 2104837 (2022).
21. Lim, J. et al. Liquid–liquid phase separation of short histidine- and
tyrosine-rich peptides: sequence specificity and molecular topol-
ogy. J. Phys. Chem. B 125,6776–6790 (2021).
22. Abbas,M.,Lipinski,W.P.,Nakashima,K.K.,Huck,W.T.S.&Spruijt,
E. A short peptide synthon for liquid-liquid phase separation. Nat.
Chem. 13, 1046 (2021).
23. Yuan, C. et al. Nucleation and growth of amino acid and peptide
supramolecular polymers through liquid–liquid phase separation.
Angew. Chem. 131, 18284–18291 (2019).
24. Shen, Y. et al. Biomolecular condensates undergo a generic shear-
mediated liquid-to-solid transition. Nat. Nanotechnol. 15,
841–847 (2020).
25. Donau, C. et al. Active coacervate droplets as a model for mem-
braneless organelles and protocells. Nat. Commun. 11, 5167 (2020).
26. Shapiro,D.M.,Ney,M.,Eghtesadi,S.A.&Chilkoti,A.Proteinphase
separation arising from intrinsic disorder: first-principles to bespoke
applications. J. Phys. Chem. B 125,6740–6759 (2021).
27. Smits,F.C.M.,Buddingh,B.C.,vanEldijk,M.B.&vanHest,J.C.M.
Elastin-like polypeptide based nanoparticles: design rationale
toward nanomedicine. Macromol. Biosci. 15,36–51 (2015).
28. Reed, E. H., Schuster, B. S., Good, M. C. & Hammer, D. A. SPLIT:
stable protein coacervation using a light induced transition. ACS
Synth. Biol. 9,500–507 (2020).
29. Hazra,M.K.&Levy,Y.Biophysicsofphaseseparationofdisordered
proteins is governed by balance between short- and long-range
interactions. J. Phys. Chem. B 125, 2202–2211 (2021).
30. Alberti, S. & Hyman, A. A. Biomolecular condensates at the nexus of
cellular stress, protein aggregation disease, and ageing. Nat. Rev.
Mol. Cell Biol. 22,196–213 (2021).
31. Bremer, A. et al. Deciphering how naturally occurring sequence
features impact the phase behaviours of disordered prion-like
domains. Nat. Chem. 14,196–207 (2022).
32. Saiani, A. et al. Self-assembly and gelation properties of α-helix
versus β-sheet forming peptides. Soft Matter 5,193–202 (2009).
33. Pashuck, E. T., Cui, H. & Stupp, S. I. Tuning supramolecular rigidity
of peptide fibers through molecular structure. J. Am. Chem. Soc.
132,6041–6046 (2010).
34. Lampel,A.,Ulijn,R.&Tuttle,T.Guidingprinciplesforpeptide
nanotechnology through directed discovery. Chem.Soc.Rev.47,
3737–3758 (2018).
35. Lampel, A. et al. Polymeric peptide pigments with sequence-
encoded properties. Science 356,1064–1068 (2017).
36. Alberti,S.,Gladfelter,A.&Mittag,T.Considerationsandchallenges
in studying liquid–liquid phase separation and biomolecular con-
densates. Cell 176,419–434 (2019).
37. Alshareedah, I. et al. Interplay between short-range attraction and
long-range repulsion controls reentrant liquid condensation of
ribonucleoprotein-RNA complexes. J. Am. Chem. Soc. 141,
14593–14602 (2019).
38. Fisher, R. S. & Elbaum-Garfinkle, S. Tunable multiphase dynamics of
arginine and lysine liquid condensates. Nat. Commun. 11,1–10 (2020).
39. Inostroza-Brito, K. E. et al. Co-assembly, spatiotemporal control and
morphogenesis of a hybrid protein-peptide system. Nat. Chem. 7,
897–904 (2015).
40. Roberts, S., Dzuricky, M. & Chilkoti, A. Elastin-like polypeptides as
models of intrinsically disordered proteins. FEBS Lett. 589,
2477–2486 (2015).
Article https://doi.org/10.1038/s41467-023-36060-8
Nature Communications | (2023) 14:421 10
Content courtesy of Springer Nature, terms of use apply. Rights reserved
41. Taniguchi, S., Watanabe, N., Nose, T. & Maeda, I. Development of
short and highly potent self-assembling elastin-derived pentapep-
tide repeats containing aromatic amino acid residues. J. Pept. Sci.
22,36–42 (2016).
42. Wang, J. et al. A molecular grammar governing the driving forces
for phase separation of prion-like RNA binding proteins. Cell 174,
688 (2018).
43. Yang, Z. Hofmeister effects: an explanation for the impact of ionic
liquids on biocatalysis. J. Biotechnol. 144,12–22 (2009).
44. Krainer, G. et al. Reentrant liquid condensate phase of proteins is
stabilized by hydrophobic and non-ionic interactions. Nat. Com-
mun. 12,1–14 (2021).
45. Li, W. & Jiang, H. Nuclear protein condensates and their properties
in regulation of gene expression. J. Mol. Biol. 434, 167151 (2021).
46. Lai, C. W., Schwab, M., Hill, S. C., Santarpia, J. & Pan, Y. L. Raman
scattering and red fluorescence in the photochemical transforma-
tion of dry tryptophan particles. Opt. Express 24,1654–1667 (2016).
47. Takeuchi, H. Raman structural markers of tryptophan and histidine
side chains in proteins. Biopolymers 72,305–317 (2003).
48. Siamwiza, M. N. et al. Interpretation of the doublet at 850 and 830
cm−1in the Raman spectra of tyrosyl residues in proteins and certain
model compounds. Biochemistry 14,4870–4876 (1975).
49. Hernández,B.,Coïc,Y.M.,Pflüger, F., Kruglik, S. G. & Ghomi, M. All
characteristic Raman markers of tyrosine and tyrosinate originate
from phenol ring fundamental vibrations. J. Raman Spectrosc. 47,
210–220 (2016).
50. Sloan‐Dennison, S. et al. Elucidation of the structure of supramo-
lecular polymorphs in peptide nanofibres using Raman spectro-
scopy. J. Raman Spectrosc. 52,1108–1114 (2021).
51. Schlamadinger,D.E.,Daschbach,M.M.,Gokel,G.W.&Kim,J.E.UV
resonance Raman study of cation-pi interactions in an indole crown
ether. J. Raman Spectrosc. 42,633–638 (2011).
52. Yuwen, T. et al. Measuring solvent hydrogen exchange rates by
multifrequency excitation N-15 CEST: application to protein phase
separation. J. Phys. Chem. B 122, 11206–11217 (2018).
53. Ambadipudi,S.,Reddy,J.G.,Biernat,J.,Mandelkow,E.&Zweck-
stetter, M. Residue-specific identification of phase separation hot
spots of Alzheimer’s-related protein tau. Chem. Sci. 10,
6503–6507 (2019).
54. Fawzi,N.L.,Parekh,S.H.&Mittal,J.Biophysicalstudiesofphase
separation integrating experimental and computational methods.
Curr. Opin. Struct. Biol. 70,78–86 (2021).
55. Gabryelczyk, B. et al. Hydrogen bond guidance and aromatic
stacking drive liquid–liquid phase separation of intrinsically dis-
ordered histidine-rich peptides. Nat. Commun. 10, 5465 (2019).
Acknowledgements
This research was supported by the Israel Science Foundation, Grant No.
2589/21 (A.L.). We thank Dr. M. Afri for help with NMR experiments and
analysis. T.M. thanks the ADAMA Center for Novel Delivery Systems in
Crop Protection for the PhD ADAMA Fellowship and the Marian Gertner
Institute for Medical Nano systems for the Gertner scholarship.
Author contributions
A.B.L. and A.L. conceived and designed the experiments. Peptide LLPS
experiments and analysis were conducted by A.B.L. and S.B.D. Raman
experiments and analyses were performed by S.S.D., K.F., and D.G. NMR
experiments were conducted by H.G. and A.B.L. NMR analyses were
conducted by J.C., H.G., and T.M. A.L., A.B.L., and J.C. wrote and edited
the manuscript. All authors discussed and commented on the
manuscript.
Competing interests
The authors declare no competing interests.
Additional information
Supplementary information The online version contains
supplementary material available at
https://doi.org/10.1038/s41467-023-36060-8.
Correspondence and requests for materials should be addressed to
Jordan H. Chill or Ayala Lampel.
Peer review information Nature Communications thanks the other,
anonymous, reviewer(s) for their contribution to the peer review of this
work. Peer reviewer reports are available.
Reprints and permissions information is available at
http://www.nature.com/reprints
Publisher’s note Springer Nature remains neutral with regard to jur-
isdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as
long as you give appropriate credit to the original author(s) and the
source, provide a link to the Creative Commons license, and indicate if
changes were made. The images or other third party material in this
article are included in the article’s Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not
included in the article’s Creative Commons license and your intended
use is not permitted by statutory regulation or exceeds the permitted
use, you will need to obtain permission directly from the copyright
holder. To view a copy of this license, visit http://creativecommons.org/
licenses/by/4.0/.
© The Author(s) 2023
Article https://doi.org/10.1038/s41467-023-36060-8
Nature Communications | (2023) 14:421 11
Content courtesy of Springer Nature, terms of use apply. Rights reserved
1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com
Available via license: CC BY 4.0
Content may be subject to copyright.
