Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES

Article
Illuminating the reaction pathways of viromimetic assembly
Hande E. Cingil, Emre B. Boz, Giovanni Biondaro, Renko de Vries, Martien
A. Cohen Stuart, Daniela J. Kraft, Paul van der Schoot, and Joris Sprakel
J. Am. Chem. Soc., Just Accepted Manuscript • Publication Date (Web): 22 Mar 2017
Downloaded from http://pubs.acs.org on March 23, 2017

Just Accepted

“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of the American Chemical Society is published by the American Chemical

Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.
Page 1 of 8 Journal of the American Chemical Society

1
2
3
4
5
6
7 Illuminating the reaction pathways of viromimetic assembly
8
9 Hande E. Cingil1, Emre B. Boz1, Giovanni Biondaro2, Renko de Vries1, Martien A. Cohen Stuart1,
10 Daniela J. Kraft2, Paul van der Schoot3,4 and Joris Sprakel1*
11
12 1. Physical Chemistry and Soft Matter, Wageningen University & Research, Stippeneng 4, 6708 WE, Wageningen, the
13 Netherlands
14 2. Soft Matter Physics, Huygens-Kamerling Onnes Laboratory, Leiden University, PO Box 9504, 2300 RA, Leiden, the
15 Netherlands
16 3. Theory of Polymers and Soft Matter, Eindhoven University of Technology, PO Box 513, 5600 MB, Eindhoven, the Neth-
17 erlands
18
19 4. Institute for Theoretical Physics, Utrecht University, Leuvenlaan 4, 3584 CE, Utrecht, the Netherlands
20
viruses, self-assembly, molecular sensors, pathway complexity, recombinant proteins, electrostatic complexation
21
22
23
24 ABSTRACT: The co-assembly of well-defined biological nanostructures relies on a delicate balance between attractive
25 and repulsive interactions between biomolecular building blocks. Viral capsids are a prototypical example, where coat
26 proteins exhibit not only self-interactions but also interact with the cargo they encapsulate. In Nature the balance be-
27 tween antagonistic and synergistic interactions is evolved to avoid kinetic trapping and polymorphism. To date it has re-
28 mained a major challenge to experimentally disentangle the complex kinetic reaction pathways that underlie successful
29 co-assembly of biomolecular building blocks in a non-invasive approach with high temporal resolution. Here we show
30 how macromolecular force sensors, acting as a genome proxy, allow us to probe the pathways through which a viromi-
31 metic protein forms capsids. We uncover the complex multistage process of capsid assembly, which involves recruitment
32 and complexation, followed by allosteric growth of the proteinaceous coat. Under certain conditions, the single-genome
33 particles condense into capsids containing multiple copies of the template. Finally, we derive a theoretical model that
34 quantitatively describes the kinetics of recruitment and growth. These results shed new light on the origins of the path-
35 way complexity in biomolecular co-assembly.
36
37
38
39 INTRODUCTION nate and why, and how they give rise to structural poly-
40 Supramolecular structures in Nature derive their func- morphism and kinetic trapping, is incomplete; not least
41 tionality from a precisely defined architecture, which in because experimental methods to probe them at the rele-
42 turn is formed by the assembly of biomolecular precur- vant length and time scales are scarce6. Radiation scatter-
43 sors1. Assembling these structures spontaneously, without ing methods, while offering excellent statistics, can be
intervention of the biochemical machinery of the cell, difficult to interpret due to the reciprocal space inversion
44
requires a finely-tuned balance of repulsive and attractive problem. By contrast, imaging methods, such as time-
45
interactions acting between multiple constituents. Con- resolved atomic force microscopy or electron microscopy,
46
trolling this delicate balance is often achieved through provide real-space insight into capsid formation but may
47
allostery, a highly cooperative process regulated through suffer from poor statistics and time resolution.
48
49 conformational switching2. Moreover, allosteric action
also controls the kinetic pathways of assembly; this is cru- Here, we present a new approach to resolve these chal-
50
cial to obtain well-defined structures with a high degree lenges and probe capsid assembly kinetics in detail. We
51
of fidelity. A case in point is the spontaneous assembly of employ morphology-sensitive luminescent polymers as
52
simple viruses, in which binding of the coat protein to its genome proxies, which act as optical sensors of their co-
53
genetic cargo sets in motion a cascade of events that leads assembly into linear virus-like particles with a recombi-
54
to successful co-assembly2c, 3. This route may involve a nant viromimetic protein. We find that capsid formation
55
multitude of competing pathways3b, 4, each of which in is initiated by the random binding of coat proteins onto
56
turn consists of a large number of elementary docking the template, after which a concerted capsid growth en-
57
and folding steps between individual molecules. This sues, caused by conformational switching of the protein.
58
gives rise to significant pathway complexity, a topic of Near conditions of charge compensation, the single-
59
intense study in the past decades, also in supramolecular genome assemblies condense into virus-like particles
60
chemistry5. Our understanding of what pathways domi- which carry multiple copies of the template. The binding
ACS Paragon Plus Environment
 Journal of the American Chemical Society Page 2 of 8

and reorganization of proteins on the template is cap- tinct changes in its luminescence spectrum upon supra-
1 tured by a simple model in which aspecific unimolecular molecular stretching 12. The second sensor polymer 13
2 binding competes with cooperative multimolecular reor- (SP2) is used to evaluate if capsids contain a single copy of
3 ganization. These results shed new light on the origins of the template, or whether multiple copies become incor-
4 the pathway complexity that result from competing inter- porated at some stage along the assembly pathway.
5 actions, stoichiometry and the action of allostery in as- The proxy genome and sensor SP1 (Fig. 1a) exhib-
6 sembling biomolecular systems. its an optomechanical coupling between the confor-
7 mation of the polymeric backbone and its photolumines-
8 cence (PL). In a relaxed state, its PL spectrum features
RESULTS & DISCUSSION
9 three distinct vibronic bands: the 1-0 transition at 𝜆 = 418
Co-assembling species. We study the for-
10 nm dominates the luminescence, while minor transitions
mation of capsids from a recombinant coat protein in-
11 are visible as the lower energy 2-0 and 3-0 bands (Fig.1c).
spired by the structure of the tobacco mosaic virus
12 In this "naked" state, the conjugation length and delocal-
(TMV), whose design and production is described in de-
13 ized electronic structure along the backbone are limited
tail elsewhere 7 and in the SI. Coat proteins of the TMV
14 by rotations between the monomers and the conforma-
feature three distinct functionalities 8: I) a hydrophilic
15 tional flexibility of the chain 12-13. Upon application of a
domain that protects the capsid and its cargo against ag-
16 stretching force, e.g. by encapsulation in our virus-like
gregation, misfolding and enzymatic attack, II) a binding
17 particles, the conformational degrees of freedom are re-
domain with high affinity for the nucleic acids, and III) a
18 duced and the chain planarizes into a ribbon-like struc-
domain that, when folded, provides a specific attraction
19 ture12. This results in a distinct change in the vibronic
between neighboring capsid proteins. Our recombinant
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35 Figure 1. Chemical structure of the mechano-optical molecular sensors (a) SP1 (Mw = 16.7 kg/mol) and (b) SP2 (Mw = 12.0 kg/mol).
36 Encapsulation of the polymers in a capsid can be detected spectrally; in absence of encapsulating protein, the luminescence spec-
37 trum of a relaxed chain features three visible vibronic bands, in which the 1-0 dominates ([SP1] = 0.06 M, 𝑓+ = 0, c). Upon en-
38 capsulation in a capsid, the sensor polymer is stretched and planarized which changes the vibronic fine structure, with the 1-0
39 band decaying and the higher energy bands growing in intensity ([SP1] = 0.06 M, 𝑓+ = 0.50, d). Addition of a benzothiadiazole
40 acceptor-moieity within the chain leads to conformation-dependent Förster resonance energy transfer, introducing an optical
41 read-out to chain bundling and condensation ([SP2] = 0.08 M, 𝑓+ = 0.70, e).
42
43 protein (Mw = 45 kDa), produced biosynthetically in Pich- transitions: the highest energy band decays and vanishes
44 ia Pastoris hosts, features a cationic binding block B com- upon full planarization, while the intensity of lower ener-
45 posed of 12 lysine residues (domain II), a silk-inspired gy transitions grows (Fig.1d). Here, we exploit this opto-
46 association domain S10 (domain III)9, and a gelatin-like mechanical coupling to detect stress-induced conforma-
47 random coil motif C (domain I)10. These proteins have tional changes of the polymer during co-assembly as a
48 been show to form stable viromimetic capsids through function of time. The second sensor polymer SP2
49 templated co-assembly with DNA 7, 11; however, the kinetic (Fig.1b&e), that allows us to probe bundling of template
50 pathways through which these rod-shaped viral capsids chains 13a will be discussed below.
51 form remain elusive. Capsid assembly kinetics. We initiate the co-
To probe its assembly kinetics, we replace the assembly by mixing protein with the proxy genome. We
52
nucleic acid polymer with an anionic conjugated polymer express the mixing stoichiometry of the two species as
53
that acts simultaneously as a proxy for the templating 𝑓+ = [+]/([+] + [−]), with [+] and [−] the molar concen-
54
genome and as a molecular sensor for the assembly pro- trations of cationic charges on the oligolysine binding
55
cess. We use two different sensor polymers: To study the block and the anionic charges on the template, respec-
56
early stages of capsid formation, in which we expect sin- tively. Polyionic charge equality is reached if 𝑓+ = 0.50.
57
gle template chains to become planarised upon encapsu- Directly after mixing, we begin recording PL spectra every
58
lation, we use a sensor polymer (SP1) that undergoes dis- 6 minutes for ~3 days, during which the vibronic fine-
59
60
ACS Paragon Plus Environment
Page 3 of 8 Journal of the American Chemical Society

structure of the ensemble-averaged luminescence spectra

1 gradually shifts, signaling the progression of the co-
2 assembly process.
3 Initially, an intense peak at the 1-0 vibronic band
4 is observed which diminishes in time, while the second
5 and third band grow (Fig.2a-b). We previously confirmed
6 that electrostatic complexation of SP1 with a poly(lysine)
7 homopolymer, in absence of conformational changes in
8 the sensor polymer, does not give rise to these distinct
9 optical signatures 12. Thus, they are a direct result of the
10 changes in template conformation due to their encapsula-
11 tion in a proteinaceous coat. As time progresses, coat pro-
12 teins bind to the template and subsequently condense to
13 form a rigid, partially complete, capsid. The time evolu-
14 tion of the spectra reveals that within the capsid, the
15 template chain is forced into a planarized and stretched
16 conformation.
17 The DNA-templated growth of virus-like parti-
18 cles from the same biosynthetic protein has revealed that
19 the assembly of rigid capsids is strongly cooperative 7, 11.
20 Rather than forming a homogeneous coating that densi-
21 fies gradually, a phase-separated structure emerges on the Figure 2. Time-evolution of the photoluminescence spectra
22 template, where sections of naked template coexist with of molecular sensor SP1 ([SP1] = 0.06 M) during encapsula-
23 segments of condensed capsid7. Ultimately, this coopera- tion by the viromimetic protein C-S10-B at charge stoichi-
24 tive co-assembly leads to a population inversion, where ometries of 𝑓+ = 0.10 (a) and 0.50 (b), with time progressing
25 most genome proxies are completely covered and some from purple (𝑡 = 0) to dark red (𝑡 = 72h). Fraction 𝛼 of encap-
26 remain (almost) completely naked. This implies that in sulated template chain as a function of time at different
27 our experiments certain segments of the sensor polymer charge stoichiometries for [SP1] = 0.06 (c) and 0.6 M (d).
28 must be planarized, in a taut state, while the remainder
29 exhibits the emission of a polymer in its relaxed, slack, Our most sensitive measure of the encapsula-
30 state. The PL spectra we record are thus an ensemble- tion-induced planarization is the intensity ratio between
31 averaged convolution of both states. the 1-0 and 2-0 bands 𝑟12 = 𝐼 𝑖=1 /𝐼 𝑖=2 12-13. From this quan-
32 Assuming such a two-state scenario, the intensity tity, within the two-state approximation, we can directly
33 of the i-th vibronic band can be written as: 𝐼 𝑖 = obtain the value of 𝛼 from our experiments, as 𝛼 =
34 (1 − 𝛼)𝐼𝑛𝑖 + 𝛼𝐼𝑐𝑖 in which 𝐼𝑛𝑖 and 𝐼𝑐𝑖 are the normalized (𝑟12 𝐼𝑛2 − 𝐼𝑛1 ) ⁄ ((𝑟12 (𝐼𝑛2 − 𝐼𝑐2 ) − 𝐼𝑛1 + 𝐼𝑐1 )). The reference
35 intensity 𝐼𝑛1 (𝑖 = 1,2) is a constant determined from a
intensities of naked (slack) and coated (taut) states, re-
36 spectively. The quantity of interest is 𝛼, the fraction of spectrum for naked genome proxy chains in the absence
37 of protein. From previous experiments 12, in which we
template chain that is encapsulated and stretched. We
38 assume here that each capsid contains a single template stretched the sensor polymers to their contour length, we
39 chain; we will demonstrate below that is valid only for determine the reference intensity for the taut confor-
40 small values of 𝑓+ , the charge stoichiometry. mation 𝐼𝑐𝑖 .
41 The fraction of taut chains 𝛼 that we extract from
42 our experiments, allows us to quantify the co-assembly
43 kinetics. For a dilute solution of components, with [SP1] =
44 0.06 M, we observe a three-stage assembly process
45 (Fig.2c). Initially, the degree of encapsulation increases
46 weakly to a plateau at 𝛼 = 0.05 which persists for approx-
47 imately 30 hrs. After this time lag, the nucleation of dense
48 capsids commences and reaches completion over the
49 course of a few hours as marked by a steep growth in 𝛼 up
50 to a mixing-ratio dependent plateau. For 𝑓+ = 0.10 we
51 observe only partial encapsulation as insufficient protein
52 is available to neutralize the available charges on the
53 template chains in the solution (Fig.2c), which is con-
54 sistent with theoretical predictions11. At perfect stoichi-
55 ometry (𝑓+ = 0.50), almost complete coverage is achieved.
56 Interestingly, further increasing the protein content, de-
57 creases the encapsulation efficiency. We could hypothe-
size this to be caused by the self-assembly of empty shells
58
at high enough protein concentrations 7, but as we will
59
60
ACS Paragon Plus Environment
 Journal of the American Chemical Society Page 4 of 8

show below it in fact signals the emergence of a 3rd stage Reaction pathway model. To demonstrate that
1 in the pathway towards forming complete capsids. the assembly process involves more pathways than just
2 Upon increasing the overall concentration of co- binding and cooperative condensation, we quantify these
3 assembling species, the nucleation lag time decreases sig- two assembly steps in a kinetic reaction model, illustrated
4 nificantly by approximately a factor of 10 (Fig.2d), as to be schematically in Figure 3. First, free proteins randomly
5 expected for a nucleation-limited process. Again we ob- and reversibly bind to the template following first-order
6 serve that in the presence of an excess of coat protein, full kinetics, which is described by a Langmuir adsorption
7 coating of the template chains is not accomplished. Note model14. Subsequently, the adsorbed proteins associate
8 that for a reaction equilibrium, increasing the overall and condense along the template to form a rigid capsid, a
9 amounts of reactants, should push the equilibrium to the process we model by cooperative 𝑛-th order Hill-type
10 right (product) side. If we compare the data for 𝑓+ = 0.50 kinetics15. We presume that cooperatively bound mole-
11 at low (Fig.2c) and high (Fig.2d) concentrations, we ob- cules can only leave the template by first transitioning to
12 serve exactly the opposite; the plateau value of 𝛼 decreas- the Langmuir state and following that are able to desorb.
13 es with the overall concentration. We could presume that Conversely, only adsorbed molecules can undergo the
14 this counter-intuitive behavior is due to the competing transition to the co-operatively bound Hill state. The fact
15 self-assembly process in the solution referred to above. that template binding is a required intermediate between
16 However, as we shall show below using a different proxy free proteins and the final capsids they form is the es-
17 genome, that in fact the observation signals the transfor- sence of our model. A complete derivation of our theory
18 mation of the single-genome particles formed initially for the reaction pathways of viral capsid assembly is pro-
19 into particles that contain multiple copies of the template vided in the SI. We note that the sequence of binding
20 that need not be taut to accommodate interactions with followed by lateral assembly has also been considered in
21 the coat proteins. For single-genome particles, stretching previous theoretical efforts 16.
22 the template chain increases the probability that anionic In brief, we consider a solution of template
23 and cationic charges can form a tight electrostatic com- chains, at mole fraction 𝑥𝑡 , and proteins with mole frac-
24 plex, and is thus driven by the Coulombic interactions tion 𝑥𝑎 . Each template chain has 𝑀 binding sites, with a
25 between genome-binding domain on the protein and total number of available binding site 𝑀𝑥𝑡 . A fraction 𝜂(𝑡)
26 template. Once the condensation occurs, each cationic of these sites is occupied by cooperatively associated pro-
27 charge can opt to bond to several templates, which thus teins in a capsid, in a Hill state, whereas 𝜃(𝑡) denotes the
28 reduces the necessity for the template stretching, which fraction of the remaining 1 − 𝜂(𝑡)sites filled by bound but
29 in itself is unfavorable as it decreases the conformational unassociated molecules, in a Langmuir state. At 𝑡 = 0 all
30 entropy of the templating polymer. Thus, in the con- proteins are free in solution, and thus 𝜂(𝑡) = 𝜃(𝑡) = 0
𝑓
31 densed capsid state, the coacervate core of the capsid al- and the fraction of free proteins 𝑥𝑎 (𝑡 = 0) = 𝑥𝑎 . During
32 lows the template to relax to some extent. 𝑓
protein binding, their total number $ 𝑥𝑎 = 𝑥𝑎 + (𝜂(𝑡) +
33 (1 − 𝜂(𝑡))𝜃(𝑡) )𝑀𝑥𝑡 is conserved. We describe the dy-
34 namics of aspecific binding by a Langmuir process 14:
35
36 𝑑𝜃 𝑑𝜂
37 = 𝐿+ [1 − 𝜃(𝑡)] − 𝐿− 𝜃(𝑡) − (1)
𝑑𝑡 𝑑𝑡
38
39 where 𝐿+ and 𝐿− are the time-dependent Langmuir ad-
40 sorption and desorption rates, respectively. In equilibri-
41 um we have:
42
43 𝑓
𝐾𝐿 𝑥𝑎 (𝑡)
44 lim 𝜃(𝑡) = lim 𝑓 (2)
𝑡→∞ 𝑡→∞ 1+𝐾𝐿 𝑥𝑎 (𝑡)
45
46 𝑓
with 𝐾𝐿 = 𝐿+ /𝐿− 𝑥𝑎 (𝑡) the dimensionless Langmuir bind-
47 ing constant. After binding, proteins can dock together by
48
forming -rolls along the template7, 9. We describe this
49
process of folding and cooperative self-assembly with the
50
Hill equation15:
51
52
𝑑𝜂
53 Figure 3.Schematic illustration of the reaction pathway mod- = 𝐻+ [1 − 𝜂(𝑡)] − 𝐻− 𝜂(𝑡) (3)
𝑑𝑡
54 el that describes capsid assembly as a combination of Lang-
55 muir adsorption of free proteins to a template, and the sub- in which the association and dissociation rates 𝐻+ and 𝐻−
56 sequent Hill-type co-operative reorganization of bound pro- of the Hill process are time-dependent. The Hill equation-
57 teins into a dense capsid of-state yields the fraction of coverage of dense and asso-
58 ciated capsids at equilibrium:
59
60
ACS Paragon Plus Environment
Page 5 of 8 Journal of the American Chemical Society

[𝐾𝐻 𝜃(𝑡)]𝑛 We find quantitative agreement between model

lim 𝜂(𝑡) = lim (4)
1 𝑡→∞ 𝑡→∞ 1+[𝐾𝐻 𝜃(𝑡)]𝑛 and experiment under stoichiometric ratios sufficiently
2 where 𝐾𝐻𝑛 = 𝐻+ /𝐻− 𝜃(𝑡)𝑛 is the dimensionless Hill con- far from charge compensation (𝑓+ =0.25 and 0.40 in
3 stant associated with cooperative binding and 𝑛 the Hill Fig.4a). Decomposing the signal 𝐹𝑤 (𝑡) into its separate
4 coefficient that controls the degree of cooperativity of the contributions illustrates how free proteins first bind onto
5 system 15. We solve these equations numerically using the the template, followed by their cooperative reorganisa-
6 Runge-Kutta approach 17 (see SI) and compare the results tion into a dense capsid (Fig.4b). The fraction of Lang-
7 to the experimental data for [SP1] = 0.6 M. Each site on muir sites initially increases until it reaches a maximum
8 the template can exist in three states, i) naked, ii)a loosely after which it decreases again. The maximum is located at
9 bound and iii) a strongly bound capsid state. The fraction the end of the nucleation lag time, where the Hill process
10 of the template sites coated with loosely bound proteins is takes over. Interestingly, the Hill coefficient 𝑛 = 5 that we
11 𝜃(𝑡) and the fraction of the remaining sites encapsulated need to describe our data, indicates that five coat proteins
12 in a dense capsid is 𝜂(𝑡). Hence, the overall fraction of are required to form a critical capsid nucleus. The agree-
13 occupied sites on the template is ment between model and experiment shows that our pre-
14 𝐹(𝑡) ≡ 𝜂(𝑡) + (1 − 𝜂(𝑡))𝜃(𝑡) (5) sumption of the two step aggregation process holds at
15 Adsorbed proteins engaged in Langmuir- and Hill-type least under certain conditions. Our experiments feature a
16 adsorption produce different signals in a measurement distinct nucleation lag time. This effect is much weaker in
17 that probes the occupied fraction of sites. Thus, we our theoretical model. This is because the model does not
18 weight them with a weighting factor 𝑤: 𝐹𝑤 (𝑡) ∼ 𝛼 = feature an energy barrier for forming a nucleus, while this
19 𝑤𝜂(𝑡) + (1 − 𝑤)(1 − 𝜂(𝑡))𝜃(𝑡). The experimental data is most likely the case in the experiments. As such, the
20 shows how initial Langmuir binding is followed by a low experiments will feature thermally-activated nucleation,
21 plateau value of 𝛼 = 0.05, after which a second rise in- while our model does not account for this. Since intro-
22 creases 𝛼 to its final equilibrium value. We interpret this ducing this feature would increase the number of adjust-
23 initial plateau at 𝛼 = 0.05 as the system existing solely in able parameters, we choose here, for the sake of simplicity
24 the Langmuir state, where proteins are randomly bound not to introduce this additional term.
25 to the template but not yet folded into the beta-sheet At higher values of 𝑓+ , i.e. 0.50 and 0.70 in Fig.4a,
26 structure required for capsid formation. Thus, we take the as charge compensation is approached, the predicted de-
27 value of w = 0.05 as an approximate measure for the signal gree of encapsulation exceeds the value determined ex-
28 intensity resulting from pure Langmuir bound proteins. perimentally. Clearly, the co-assembly pathways are more
29 complex under stoichiometric and super-stoichiometric
30 circumstances. A clue to what is happening is provided
31 by earlier experiments on electrostatic complexes of syn-
32 thetic polymers in the context of what is known as com-
33 plex coacervation. Indeed, we have very recently observed
34 a phase transition from small and soluble complexes un-
35 der sub-stoichiometric conditions, to multi-molecular
36 liquid-like structures upon approaching charge neutrali-
37 zation at 𝑓+ = 0.513a. If this translates to our system, then
38 one would expect a similar transition between a capsid
39 containing a single, stretched, genome proxy to one
40 where the capsid contains multiple copies of the tem-
41 plate, that no longer need to stretch to accommodate the
42 coat proteins and interact electrostatically. We speculate
43 that the gain in configurational entropy of the template
44 offsets the Coulombic penalty of bringing multiple tem-
45 plates into close proximity. Moreover, the latter are
46 screened anyway by the presence of the positively charged
47 tails of the coat proteins.
48
49
50
51
52
53 Figure 4. a) comparison of the experimentally determined
54 value of the coating fraction 𝛼 (symbols are experimental
data with a legend as in Fig.2d) and predictions by the model
55
as described in the text (solid lines) for 𝑓+ = 0.25, 0.40, 0.50
56
and 0.70 (from bottom to top). b) Time-evolution of the frac-
57 𝑓
tion of free 𝑥𝑎 , Langmuir 𝜃 and Hill proteins 𝜂 for 𝑓+ = 0.50.
58
59
60
ACS Paragon Plus Environment
 Journal of the American Chemical Society Page 6 of 8

same conditions. We attribute this small change to a re-

1 duction in linear charge density upon introducing un-
2 charged BT moieties at the expense of dicarboxylated flu-
3 orene units.
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24 Figure 5. Time-evolution of emission spectra, normalized to
25 the total emitted intensity, of molecular sensor SP2 ([SP2] =
26 0.08 M) during co-assembly with C-S10-B at 𝑓+ = 0.3 (a) and
27 0.5 (b). Insets show a close-up around the emission peak of
28 the energy transfer acceptor benzothiadiazole (BT).
29 Capsid condensation. To probe if such bun- Figure 6. The fraction of encapsulated template chains 𝛼 and
30 dling indeed occurs and explains the observed deviations corresponding energy-transfer ratio 𝜖 between donor (fluo-
31 from our theoretical predictions, we employ a second rene) and acceptor (BT) for 𝑓+ = 0.10 (a), 0.30 (b), 0.50 (c)
32 molecular sensor, coded SP2. This genome proxy is ran- and 0.70 (d).
33 domly doped with a small fraction of benzothiadiazole
34 (BT), an acceptor for Förster resonance energy transfer As the point of charge compensation is ap-
35 (Fig.1b)13a, 18. In isolation, the fluorescence spectra of this proached and exceeded, we observe a significant increase
36 polymer exhibits the same response as its homopolymer in the intensity of the energy transfer band (inset Fig.5b
37 equivalent SP1 with no appreciable energy transfer be- and Fig.6c-d); this signals the electrostatic condensation
38 tween the fluorene donors and BT acceptors due to its transition of the co-assembled objects into multi-
39 low doping degree. However, when these sensor polymers template capsids 13a. We determine the energy transfer
40 condense and bundle or fold, a significant increase in lu- efficiency 𝜖 as the ratio of the BT acceptor intensity to
41 minescence at 550 - 600 nm is observed due to intermo- that of the fluorene donor. We observe a significant bun-
42 lecular energy transfer between the fluorene and BT dling-induced energy transfer as multiple template chains
43 units13a, 18. Encapsulation of SP2 with our coat protein in- become encapsulated in a single capsid, which becomes
44 deed shows an energy transfer signal emerging under (su- increasingly pronounced as the mixing ratio of the tem-
45 per-)stoichiometric conditions (Fig.1e). plate and the protein is increased (Fig.6a-d).
46 In order to test our coacervation hypothesis, we These data suggest that the void of the viral cap-
47 repeat the capsid assembly studies with the molecular sid is occupied by more than one template chain at
48 sensor SP2. The normalised PL spectra show the same charge-neutral stoichiometry, while each virus-like parti-
49 change in vibronic structure as observed for SP1 but with cle contains a single template at low values of 𝑓+ . Interest-
50 an additional energy transfer band due to the BT accep- ingly, the spectroscopic signal for template condensation
51 tor, as shown in Figure 5; this is particularly pronounced does not appear until the final kinetic stages of capsid
52 for high charge stoichiometries, as can be seen in the in- formation. This can be most clearly seen in Fig.6c, where
53 set in Fig.5b. condensation does not occur until about half of the tem-
54 From the ratio of the vibronic bands 1-0 and 2-0, as dis- plate chains have been encapsulated. This indicates that a
55 cussed above for SP1, we find the highest fraction of coat- third kinetic phase emerges in assembly pathways, during
56 ed templates is reached at 𝛼 ∼ 0.7, as shown for different which nucleated capsids around a single template under-
57 stoichiometries in Figure 6. This is slightly lower than go a condensation transition to multi-genome objects.
58 what we find for the homopolymer sensor SP1 under the
59
60
ACS Paragon Plus Environment
Page 7 of 8 Journal of the American Chemical Society

CONCLUSIONS Stonehouse, N. J.; Homans, S. W.; Ashcroft, A. E., J.

1 Our results reveal the pathway complexity of the co- Mol. Biol. 2007, 369, 541; (c) Zlotnick, A., J. Mol. Biol.
2 assembly of a viromimetic protein and a genome proxy, 1994, 241, 59; (d) Zlotnick, A.; Mukhopadhyay, S.,
3 hinting that similar pathway complexity underlies the Trends Microbiol. 2011, 19, 14.
4 formation of natural viruses. Our data and theoretical 3. (a) Bruinsma, R. F.; Gelbart, W. M.; Reguera,
5 analysis suggests that for our experimental model for lin- D.; Rudnick, J.; Zandi, R., Phys.Rev.Lett. 2003, 90,
6 ear viruses, capsid assembly follows a distinct sequence of 248101; (b) Misra, N.; Lees, D.; Zhang, T.; Schwartz, R.,
7 steps: i) coat proteins bind randomly to the template; ii)
8 Comput Math Methods Med. 2008, 9, 277; (c) Schwartz,
Bound proteins reorganize and fold until a multiprotein
9 R.; Shor, P. W.; Prevelige, P. E.; Berger, B., Biophys. J.
nucleus is formed and a dense capsid grows. Interestingly,
10 1998, 75, 2626; (d) Zandi, R.; van der Schoot, P.;
a similar sequence of events has been evidenced in-vitro
11 Reguera, D.; Kegel, W.; Reiss, H., Biophys. J. 2006, 90,
for spherical viruses 19. iii) Under conditions of (super-
12 )stoichiometry, a third stage emerges in which single-
1939.
13 genome virus-like particles condense to form objects that 4. Hagan, M. F.; Chandler, D., Biophys. J. 2006,
14 contain multiple copies of the template. Whether this is 91, 42.
15 relevant in the biological context is unclear but it would 5. (a) Korevaar, P. A.; George, S. J.; Markvoort, A.
16 be interesting to test this using the experimental strategy J.; Smulders, M. M.; Hilbers, P. A.; Schenning, A. P.; De
17 we follow. Our approach relies on the planarization of the Greef, T. F.; Meijer, E., Nature 2012, 481, 492; (b)
18 sensor molecule in a linear capsid cavity. Engineering the Korevaar, P. A.; de Greef, T. F.; Meijer, E., Chem. Mater.
19 molecular design of the sensor polymer, e.g. by introduc- 2013, 26, 576; (c) Besenius, P., J. Polym. Sci., Part A-1:
20 ing donor-acceptor moieties or biomolecule-specific bind- Polym. Chem. 2017, 55, 34.
21 ing sites, could open up possibilities to explore the as- 6. (a) Uetrecht, C.; Barbu, I. M.; Shoemaker, G.
22 sembly pathways of more complex and naturally- K.; Van Duijn, E.; Heck, A. J., Nat. Chem. 2011, 3, 126;
23 occurring viral structures such as the tobacco mosaic vi- (b) Zlotnick, A.; Johnson, J. M.; Wingfield, P. W.; Stahl,
24 rus. S. J.; Endres, D., Biochemistry 1999, 38, 14644.
25 7. Hernandez-Garcia, A.; Kraft, D. J.; Janssen, A.
26 ASSOCIATED CONTENT
F.; Bomans, P. H.; Sommerdijk, N. A.; Thies-Weesie, D.
27 Supporting Information. The supporting information is M.; Favretto, M. E.; Brock, R.; De Wolf, F. A.; Werten,
28 available free of charge on the ACS Publications website at
M. W., Nat. Nanotechnol. 2014, 9, 698.
29 DOI: xxxxxxx. Fig.S1 details synthesis protocols and charac-
tersiation of the materials, Figure S2-S3 additional excitation-
8. (a) Caspar, D. L.; Namba, K., Adv. Biophys.
30
31 emission spectra of the sensor polymers and lifetime meas- 1990, 26, 157; (b) Klug, A., Philos. Trans. R. Soc., B 1999,
32 urements, Figures S4-S5 give the protein sequence and 354, 531; (c) Namba, K.; Pattanayek, R.; Stubbs, G., J.
33 MALDI-TOF measurements (PDF link). Mol. Biol. 1989, 208, 307.
34 9. Zhao, B.; Stuart, M. A. C.; Hall, C. K., Soft
35 AUTHOR INFORMATION matter 2016, 12, 3721.
36 Corresponding Author 10. Werten, M. W.; Wisselink, W. H.; Jansen-van
37 * email: joris.sprakel@wur.nl den Bosch, T. J.; de Bruin, E. C.; de Wolf, F. A., Protein
38 Eng. 2001, 14, 447.
39 Author Contributions 11. Punter, M. T.; Hernandez-Garcia, A.; Kraft, D.
40 The manuscript was written through contributions of all J.; de Vries, R.; van der Schoot, P., J. Phys. Chem. B
41 authors and all authors have given approval to the final ver- 2016, 120, 6286.
42 sion of the manuscript. 12. Cingil, H. E.; Storm, I. M.; Yorulmaz, Y.; te
43 Brake, D. W.; de Vries, R.; Cohen Stuart, M. A.;
44 Sprakel, J., J. Am. Chem. Soc. 2015, 137, 9800.
45 ACKNOWLEDGMENT 13. (a) Cingil, H. E.; Boz, E. B.; Wang, J.; Stuart, M.
46 A. C.; Sprakel, J., Adv. Funct. Mater. 2016, 1420; (b)
47 This work was financially supported by a European Research Zhang, Y.; Liu, B.; Cao, Y., Chem. - Asian J. 2008, 3, 739.
48 Council Advanced Grant (ERC-267254) and the Netherlands 14. Langmuir, I., J. Am. Chem. Soc. 1918, 40, 1361.
49 Organisation for Scientific Research (NWO) through projects
15. Hill, T. L., J. Chem. Phys. 1946, 14, 263.
50 712.012.007 and 680-47-431.
16. Hu, T.; Shklovskii, B. I., Phys.Rev.E 2007, 75,
51
REFERENCES 051901.
52
17. Butcher, J. C., J. Austral. Math. Soc. 1963, 3,
53 1. (a) Philp, D.; Stoddart, J. F., Angew. Chem., Int.
54 185.
Ed. Engl. 1996, 35, 1154; (b) Whitesides, G. M.;
55 18. Liu, B.; Bazan, G. C., Chem. Mater. 2004, 16,
Grzybowski, B., Science 2002, 295, 2418.
56 4467.
2. (a) Packianathan, C.; Katen, S. P.; Dann, C. E.;
57 19. Garmann, R. F.; Comas-Garcia, M.; Gopal, A.;
Zlotnick, A., J. Virol. 2010, 84, 1607; (b) Stockley, P. G.;
58 Knobler, C. M.; Gelbart, W. M., J. Mol. Biol. 2014, 426,
Rolfsson, O.; Thompson, G. S.; Basnak, G.; Francese, S.;
59 1050.
60
ACS Paragon Plus Environment
 Journal of the American Chemical Society Page 8 of 8

1
2
3
4
5
6
7
8
9
10
11
12
13 For Table of Contents only
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58 8
59
60
ACS Paragon Plus Environment