Mechanistic Insights into Polymerization-Induced Self-Assembly Using Maleimide-Based Fluorophores

,


■ INTRODUCTION
−4 However, this evolving complexity is accompanied by significant challenges in characterizing the self-assembly pathways, limiting our overall understanding of ways of targeting specific nanoparticle morphologies through a rational experimental design.The ability to understand how changes in polymer design and assembly conditions impact the final nanoscale morphology by studying the self-assembly process in situ would provide critical insights toward achieving materials with tailored structures and functionalities.
We are especially interested in the study of polymerizationinduced self-assembly (PISA).This versatile and readily accessible method enables the production of nanoparticles of various morphologies in situ as the polymerization progresses. 5ISA involves chain-extension of a solvophilic macromolecular chain-transfer agent with monomers that are miscible with the continuous phase but form a solvophobic, immiscible polymer, which drives self-assembly.Amenable to a wide variety of reaction conditions and monomer chemistries, PISA has become an invaluable method for nanoparticle production across many research fields. 6,7However, challenges in studying the onset of self-assembly and the evolution of morphology in situ during PISA have critically limited our mechanistic understanding of the process to date.In particular, the final nanostructures produced during PISA can often be dynamic and kinetically trapped, and their assembly process involves passing through a range of initial transition states and spontaneous organization steps resulting from different intermolecular forces to achieve their final morphology. 8,9e note that in situ small-angle X-ray scattering (SAXS) has been demonstrated to sensitively monitor the morphological transitions during PISA; 10,11 however, this requires highly specialized equipment and expertise, which limits the practicality of this method.Overall, a complete understanding of the PISA process requires a readily accessible characterization methodology with high sensitivity, a fast response time, and the ability to be performed with particles in their native hydrated state.In this work, we aim to address this important research challenge.
Small-molecule fluorescent probes, which change their fluorescence properties in response to a binding event, chemical reaction, or change in the immediate environment, are among the most utilized spectroscopy tools. 12,13−23 More recently, fluorophores have also been employed in the study of macromolecular self-assembly events, for example, to probe the structures of porphyrins, 24,25 aggregation/assembly of peptides, 26,27 formation of hydrogel networks, 28,29 and more. 30,31−34 Small-molecule probes can be chemically modified to tune spectroscopic properties, and their small size enables incorporation into systems without compromising the overall structure or function, enabling the investigation of the true environment.−43 Focusing on fluorophores as reporters of polymer selfassembly, we have shown previously that the inclusion of a maleimide unit within an amphiphilic block copolymer enabled the formation of self-reporting polymer micelles. 44Photophysical characterization showed that the maleimide moiety could act as a reporter for micellization, whereby emission lifetime, anisotropy decay, and time-domain fluorescence lifetime imaging (FLIM) could rapidly resolve differences in the particle supramolecular state.Fluorescence lifetime is an intrinsic photophysical property that is independent of the local fluorophore concentration. 45,46As such, it is an absolute measurement and relatively unaffected by artifacts due to scattered light and photobleaching, making its fluorescence lifetime measurements particularly powerful in the study of self-assembled structures.
More recently, Rho et al. included a pyrene unit within a PISA system with the aim of monitoring assembly in situ, where they were able to demonstrate differences in fluorescence emission as assembly occurred on account of the changing fluorophore environment. 47However, only a fraction of polymer chain ends could be labeled with the pyrene unit using their synthetic strategy, and only fluorophore emission was studied, limiting mechanistic insights to the simple onset of assembly and final morphology.
Here, we rapidly accelerate progress in our understanding of the PISA process by using a maleimide-based fluorophore as an environmental probe.We hypothesized that in situ fluorescence lifetime measurements would provide quantitative evidence of the onset of particle formation, evolution of morphology, and, uniquely, any transition states during the assembly process in unprecedented detail.Furthermore, we expected that varied particle morphologies (spheres, worms, and vesicles) would provide distinctive lifetime decay curves on account of their different core environments.Finally, we set out to demonstrate the applicability of this technique for cell imaging using FLIM, with the anticipation that different particle morphologies and degradation behaviors could be readily deconvoluted when structures are imaged with the same fluorophore.
Synthesis and Characterization.Our essential design considerations were as follows: (1) a synthetic strategy whereby every polymer chain end could be reliably labeled with a maleimide unit without impacting the self-assembly process; (2) a polymer system amenable to particle formation in a pure aqueous environment to enable transfer into biological applications; and (3) a chemically tunable system to ensure wide applicability for future research endeavors.To achieve this, we synthesized a chain-transfer agent (CTA) for reversible addition−fragmentation chain transfer (RAFT) polymerization with a maleimide moiety coupled to the Z group (Scheme 1).An aminochloromaleimide (ACM) Scheme 1. (A) Reaction Schematic for the Synthesis of Maleimide-Functionalized Block Copolymers and (B) Cartoon Depicting the Self-Assembled Polymer Systems Targeted in This Work, Showing Spheres, Worms, and Vesicles with Maleimide Units Localized within the Hydrophobic Core Macromolecules derivative was chosen for this work as, thus far, this class is the brightest known of the maleimide dyes 35 and is widely reported and understood. 36We chose glycerol monomethacrylate (GMA) 48 and tetrahydrofurfuryl acrylate (THFA) as the corona and core monomers, respectively, in order to achieve sterically stable particles in an aqueous environment with a sufficiently hydrophobic core (monomer log P oct 1.31) 49 to assess subtle changes in the probe environment.
Polymerization of GMA yielded a water-soluble macroCTA with maleimide chain ends (see Scheme 1, Figure 1A, and Supporting Information for details).In this way, the probe would be situated within the growing hydrophobic polymer block, allowing for accurate environmental reporting during the polymerization.The polymerization was performed in methanol at 65 °C using 4,4′-azobis(4-cyanovaleric acid) (ACVA) as a thermal radical initiator and quenched at 90% conversion, achieving a final average degree of polymerization (DP) of 43.Size exclusion chromatography (SEC) analysis confirmed that the GMA macroCTA had a number-average molar mass (M n ) of 6.6 kDa and a dispersity (D̵ M ) of 1.1.Importantly, the SEC trace showed good overlap of the refractive index (RI) and UV detectors at 309 nm (trithiocarbonate group) and 380 nm (maleimide group), highlighting the excellent end group fidelity on the macroCTA  (Figure 1A,B).Subsequently, a series of PISA experiments were performed targeting a range of core DP values from 40 to 200 at 20 wt % solids in order to access a variety of nanoparticle morphologies.Each PISA reaction reached full monomer conversion within 8 h, as confirmed by 1 H NMR spectroscopic kinetic analysis.The final polymers were analyzed by SEC, which confirmed the quantitative blocking efficiency as well as the full retention of the macroCTA and maleimide UV signals on the polymer chains (Figure 1C).We were able to observe a range of morphologies ranging from spheres (DP = 40), large spheres (DP = 50), short worms (DP = 100), long worms (DP = 165), vesicles (DP = 180), and large vesicles (DP = 200), as confirmed by dry-state TEM analysis (Figure 1D, Table 1).These morphologies were further corroborated by collecting the small-angle X-ray scattering (SAXS) patterns of small spheres, long worms, and vesicles (Figure 1E) using a synchrotron SAXS beamline. 50he radially integrated patterns obtained for these PGMA 43 -PTHFA y particles were plotted as X-ray scattering intensity, I(q), vs the scattering vector, q, and demonstrated the expected gradients at low q of approximately zero for spheres, −1 for worms, and −2 for vesicles. 51Together, these data unequivocally confirm that the full range of nanoparticle morphologies was achieved.Pleasingly, the full range of higher-order morphologies could be obtained when solely using the maleimide-functionalized macroCTA, indicating that the presence of the small probe on the chain ends did not impact assembly.This was confirmed by targeting the same polymer DPs using a nonfluorescent control macroCTA, which showed similar morphologies at the same DP values (Figures S17−S19, Table S1).
Steady-State Fluorescence and Fluorescence Lifetime Decay.We next sought to establish how the nanoparticle morphology influenced the fluorescent properties of the maleimide fluorophore.Steady-state excitation and emission spectra of the macroCTA in an aqueous solution showed an excitation maximum (λ Ex ) of 385 nm and emission maximum (λ Em ) of 535 nm.Upon assembly into nanoparticle morphologies, all samples showed a slight shift in the excitation maxima (λ Ex = 380 nm), with obvious blue-shifted emission  behavior to different extents depending on morphology (Figure 2A).Previous research into these probes has shown this same trend when moving from polar to nonpolar solvents, resulting from reduced dipole−dipole interactions and shielding the maleimide from electron-driven proton transfer quenching mechanisms. 52In this system, the fluorophore in the macroCTA was initially exposed to the quenching aqueous environment; however, upon self-assembly, it became progressively more shielded by the inclusion of the hydrophobic monomer units.The nanoparticle emission maxima values spanned from 510 nm for the small spheres down to 494 nm for the long worms, showing a trend of subtle changes in optical properties with the evolution of morphology.Solution-state time-correlated single-photon counting (TCSPC) was performed next to determine the emission decay profiles.Each nanoparticle solution and the macroCTA as control were assessed with a pulsed 405 nm diode laser (135 ps (full-width at half-maximum (fwhm)), and the resultant emission decays are presented in Figure 2B with exponential fitting details in Table 2. Every sample consistently displayed three components to decay after deconvolution, which were used to calculate average lifetime (τ av,A ) values for an initial comparison.These results presented a more remarkable trend, whereby obvious increases in lifetime were observed in comparison to that of the macroCTA (0.8 ns).Polymerization to a core DP of 40 yielded small spheres with a lifetime of 3.6 ns, whereas targeting a core DP of 50 gave slightly larger spheres with an increased τ av,A value of 4.7 ns.When targeting a core DP of 100, short worms were produced with a τ av,A of 5.9 ns, whereas vesicles (DP = 180) and large vesicles (DP = 200) further extended the lifetime out to 7.0 and 7.8 ns, respectively.Intriguingly, the biggest increase in lifetime was observed for the long worm sample (DP = 165), with a value of 12 ns.It is noteworthy that this morphology also showed a different trend in the excitation spectra, whereby sharp features can be observed in addition to the broad peak exhibited by all samples.In combination, these results indicate that the long worms facilitate a maleimide environment where different electronic (and possibly vibronic) transitions from the ground state to the excited states are favored.The prolonged lifetime displayed may be the result of intersystem crossing to the lowest excited triplet state; however, further experimentation is required to confirm this observation.
In order to rule out the possibility that changes in lifetime decay were simply due to an increased hydrophobic environment with increasing core DP, a control experiment was performed using a shorter macroCTA and varying the core DP over a limited range at the border between sphere and vesicle phases (GMA 33 -THFA 40 , spheres; GMA 33 -THFA 50 , spheres; GMA 33 -THFA 60 , spheres/vesicles) at 10 wt % to minimize phase transitions (Table S2).Pure sphere morphologies yielded steady-state emission maxima of 504 nm and average lifetime decay values of 4.8 ns, whereas the mixed sphere/ vesicle morphology showed a minute blue-shift in the emission maximum to 502 nm and a more dramatic shift of lifetime decay to 6.1 ns (Figures S24 and S25, Table S2).This confirmed that increasing the core block DP alone did not sufficiently explain the increase in lifetime decay values and that nanoparticle morphology plays a defining role.
To probe this further, each lifetime decay curve was subsequently analyzed with an exponential reconvolution fit (Table 2).For the macroCTA, the maleimide decays almost immediately, with two subnanosecond lifetimes comprising much of the average decay and a third slightly longer decay of 3.1 ns.The shortest decays are attributed to nonemissive aggregation and solvent-collision effects, and the longest lifetime is the intrinsic relaxation event (excited-state lifetime of the isolated probe) for the maleimide fluorophore. 44pon examining the deconvoluted lifetime decay profiles of the nanoparticles, it could be seen that regardless of morphology, each decay component was prolonged in comparison to the macroCTA.First, each τ 1 component, which we have attributed to the intrinsic lifetime of the maleimide in each particle, was increased above 1 ns and, except for the small spheres, comprised less than 30% of the total decay.This behavior confirms that upon self-assembly, the maleimide is more protected from the external aqueous environment, in agreement with the observations from the steady-state fluorescence data.Of particular interest is the analysis of the τ 2 and τ 3 components, which provide quantitative evidence that the environmental state of the maleimide changes with morphology.For example, increasing the number of spheres to the higher-order morphologies of worms and vesicles increased the value of τ 2 to above 5 ns, with this contributing more than 55% of the overall lifetime decay.Perhaps even more characteristic, when considering τ 3 , it can be clearly seen that this value slowly increases as morphology evolves from spheres to worms to vesicles as well as its contribution to the overall decay profile.
Taken together, these data strongly suggest that the prolonged nanoparticle lifetimes compared to those of the macroCTA cannot be sufficiently explained by shielding from the aqueous environment alone.It is clear from the changes in fluorophore lifetime that the physical state of the fluorophore is changing within each morphology.However, understanding the mechanism for this is likely to be complex and beyond the scope of this study.We can hypothesize that there is a different packing of the maleimide moiety within the core of each morphology, enabled by the different core curvature as shown in Scheme 1B, as well as the restricted molecular movement that results from immobilization within the core environment. 43It is also possible that the exclusion of water from the core region is varied for each morphology.Of great interest, the long worm morphology displays the longest τ 3 value, which supports our observations from the absorbance data that different electronic transitions are occurring toward longer lifetime states.
Overall, we attribute the longer τ 2 and τ 3 values to better and different fluorophore protection when in the self-assembled state and propose that each component extraction provides unique fingerprinting for characterization of morphology using fluorescence.
Monitoring of PISA Kinetics during Self-Assembly.Next, we validated the ability for the maleimide fluorophore to give a detailed analysis of the PISA process as polymerization of the core block occurred.For this validation, we chose to investigate the highest DP (200, large vesicles) to enable the study of each transitional morphology that occurred during the polymerization.As described above, two PISA experiments were performed side-by-side, each targeting a final core DP = 200.One experiment was left untouched, while the second was sampled every 10 min for the first hour and then every hour until quantitative conversion was reached.Kinetics were compared across both experiments to ensure that sampling did not impact the polymerization (Figure S26).Each sample was then diluted 100 times, analyzed for both steady-state Macromolecules emission and lifetime decay behavior, and verified for the presence of self-assembly using dry-state TEM.It should be noted that due to the high weight percentages required for the PISA process, the polymerizations cannot be performed and analyzed within the fluorescence spectrometer as the high fluorophore content provides inaccurate results due to the oversaturation of the detector.
Fluorescence steady-state and lifetime measurements in conjunction with dry-state TEM imaging were able to monitor the evolution of the nanoparticle morphology as the polymerization progressed.Table S3 shows the blue-shifted emission maxima and the increase in average lifetime decay over time (Figure 3A, raw data shown in Figures S27−S29).Of interest is the time 0 value, which has a slightly longer τ av,A of 2.8 ns in comparison to 0.8 ns for the macroCTA alone in solution.We attribute this to the presence of the core monomer in the starting formulation, providing small pockets of hydrophobic environment for maleimide shielding (discussed further below).We first analyzed the data after the onset of assembly at 60 min, where a clear spherical morphology can be seen in the TEM images (Figure 3B).Beyond this point, the in situ emission and lifetime behavior very closely match those of the individually isolated morphologies.Specifically, at 2 h, short worms have formed with an λ Em of 500 nm and τ av,A of 5.8 ns, compared to the short worms from above with an λ Em of 502 nm and τ av,A of 5.9 ns.The vesicle morphology was observed in situ after 4 h, with an λ Em of 500 nm and τ av,A of 6.7 ns compared to the above vesicles with an λ Em of 498 nm and τ av,A of 7.0 ns.Finally, the large vesicles were obtained after 18 h of reaction, with an λ Em of 498 nm and τ av,A of 7.4 ns in agreement with the above data.This confirms that the data for the known nanoparticle morphologies can be used to rapidly,  and with accuracy, identify morphological evolution simply by using fluorescence analysis.Subsequently, we explored the data collected before the initial onset of assembly into spheres, with the hypothesis that this could provide insight into the mechanism of PISA in the early stages of polymerization.As above, we analyzed the collected data using fluorescence and TEM (raw data shown in Figures S30−S32), which revealed an unusual trend particularly in the average lifetime values (Figure 3C, Table 3).We observed a significant rapid increase in maleimide τ av,A and blue-shifting of λ Em within the first 10 min, from 2.8 to 5.7 ns and 540 to 525 nm, respectively.By 30 min, the lifetime had slightly decreased to 3.1 ns and the emission maximum had red-shifted to 530 nm.However, at 40 min, the lifetime once again increased to 4.9 ns and emission was blue-shifted to 510 nm.At 50 min, the emission maximum remained constant; however, the lifetime decreased back to 3.9 ns where it remained at 60 min, where the onset of initial assembly was complete.
Applying the deconvolution provided additional insight into this seemingly random trend of the initial 60 min.At time 0, the decays of <1.3 ns as identified for the macroCTA in solution can be observed.Until 30 min, these τ 1 and τ 2 values remain constant; however, their percentage contribution to the average lifetime changes.Specifically, the relative % of τ 1 decreases sharply and then slowly increases, while the relative % of τ 2 increases.During this time, both the values and relative % of τ 3 change, with an initial increase to 6.7 ns and 39%, respectively, followed by a decrease down to 4.9 ns and 15% by 30 min.This indicates that the maleimide is still primarily located within an aqueous environment as the <1.3 ns decay components are present; however, some form of longer-range aggregation, or premicellar assemblies, occurs.These appear to form at low conversion very rapidly as the macroCTA begins chain extending with the core monomer.We hypothesize that initially, the free THFA molecules in solution are associated with one another, stabilized by the growing macroCTA.This would cause the initial dramatic change in fluorophore photophysical properties.As the polymerization continues toward the formation of micelles, the aggregates will shrink in size and become more ordered, reflected in the decreasing trend in relative % for τ 1 .Around 40 min, we can observe that the onset of the initial sphere morphology starts to occur.At this point, τ 1 becomes difficult to measure [0.1 ns, which is below the fwhm of the diodes (135 ps)], with a shift in relative % to the longer decay events.By 50 min, both τ 1 and τ 2 most closely resemble those of the small spheres and show a less obvious change in the data at 60 min.Studying the TEM images in parallel (Figure 3F), we were able to observe the presence of small low-density spherical features at all time points from 10 min, with obvious spheres present at 60 min.Taken together, these data suggest that while defined spheres do not appear on the TEM analysis until 60 min, the actual onset of assembly occurs somewhere between 40 and 50 min where the τ 1 value increases from 0.1 to 0.9 ns.
We next analyzed two additional PISA formulations targeting spheres (DP = 40) and short worms (DP = 100) to understand whether this was a universal observation.During the first 60 min, the PISA experiment targeting spheres did not show this premicellar aggregation behavior, with no changes in lifetime or emission maxima before 60 min (Figure S33).Conversely, when targeting short worms, premicellar aggregation was observed albeit to a lesser extent (Figure S34).This suggests that this process occurs only when targeting higherorder morphologies, likely due to the higher concentration of core monomers in the reaction mixture.To the best of our knowledge, this is the first example of in situ characterization data showing the mechanism of particle formation in the preassembly state.Full characterization of this process using in situ small-angle X-ray scattering is underway and will be the focus of detailed future work in this area.
Computational Analysis Using Log P oct /SA.In eq 1, the octanol−water partition coefficients (log P oct ) quantify the partitioning of solute molecules between immiscible phases.Indeed, negative Log P values denote molecules, such as GMA, that prefer the aqueous phase.In contrast, positive values for THFA (log P oct = 1.31),CTA (log P oct = 3.56), and the maleimide fluorophore (log P oct = 3.97) indicate partition into the organic phase.
Normalizing a thermodynamic parameter (i.e., log P oct ) with a structural parameter (i.e., surface area, SA) facilitates comparison of architectural differences resulting from monomer size and functionality as well as oligomer length. 53,54−57 Here, we applied this analysis to help gain insight into the formation of vesicles during the first hour of polymerization.First, we constructed oligomeric molecular models with a ratio of GMA to THFA that matches the conversion of the growing copolymer (at T = 298 K).Then, log P oct values were normalized by SA after molecular dynamics (MD) simulations (see Supporting Information for a detailed model).In Figure 4a, the conversion and log P oct /SA were plotted against time.Before the polymerization of THFA, the macroCTA containing GMA units had a negative log P oct /SA value.Interestingly, as the polymerization of THFA approached 11% conversion by the first 10 min time point, the negative log P oct / SA value for the macroCTA became positive.As conversion continued, the hydrophobicity of the growing polymer chain increased before slowing down around 30 min.
Although a comparison of log P oct /SA and lifetime values in Figure 4b did not reveal any obvious relationship, the largest changes in lifetime values occurred when values for log P oct /SA were positive.Since the maleimide probe also has a positive log P oct /SA value, we hypothesize that the fluorescence probe on the macroCTA initially starts in the aqueous phase but transfers to a hydrophobic environment after the conversion of THFA exceeds 11%.
In eq 2, log P oct is a free energy term. 58As such, molecules with positive log P oct values spontaneously transfer from the water phase to the hydrophobic phase.This suggests that by 10 min, the environment surrounding the macroCTA is sufficiently hydrophobic to sequester the THFA monomer into monomer aggregates as we have hypothesized above.We therefore suggest that this is the cause of the initial large jump in lifetime.Over the next 20 min, the lifetime decreases as these monomer units are converted from an aggregated form into the growing polymer chain.Beyond 30−40 min, the ordered morphology forms and lifetime concomitantly begins to increase again, this time with morphology rather than hydrophobicity.
Fluorescence Lifetime Imaging Microscopy.Our final aim was to determine the suitability of this PISA system for the characterization of cellular behavior using FLIM.We first established whether the results discussed above from the TCSPC laser attachment on the fluorimeter would correlate with the 405 nm pulsed diode laser attachment on the confocal microscope.For this, 0.1 wt % solutions of each morphology were sandwiched between glass slides and directly imaged using FLIM (Figure 5A).All obtained images were analyzed using the fast-FLIM method for lifetime decay spectra and histograms (Figure 5B,C).It is important to note that at this point, as the nanoparticles were intended for cellular uptake, we chose to analyze a medium-length worm sample rather than the long worms as described above and excluded the poorly defined vesicles.This is because long aspect ratio cylindrical particles and highly dispersed particles have difficulties crossing cell membranes, and therefore, we sought to prioritize morphologies with a higher chance of internalization. 59,60hile the changes in lifetime were subtler compared to that of the solution state, an increase in lifetime decay values could clearly be observed as morphology evolved into higher-order structures.This was accompanied by observable changes in the lifetime histogram.For the macroCTA, the lifetime decay was rapid and showed a narrow distribution in the histogram (Figure 5C), whereas the small and large spheres showed a broader peak with a low maximum value and tailing toward longer lifetimes.The short and medium worms had peaks significantly shifted to the right indicating a longer overall lifetime, a negligible presence of shorter lifetime events, and some tailing toward longer lifetimes.The vesicles showed the broadest histogram, with a maximum quite close to the worm morphologies, but with a larger proportion of shorter and longer lifetime events.Direct lifetime tail fittings of the entire field of view are presented in Table S6.As with the solutionstate measurements, the nanoparticles consistently showed a significantly larger contribution to the overall decay from longer-lived fluorophore states, which increased with higherorder morphology in comparison to that of the macroCTA alone.A visual comparison of these data shows quite clear qualitative differences in fluorophore behavior, even without formal emission fitting.This property enables ready analysis of fluorophore state to the nonexpert user, which facilitates the translation of this process into a wide range of pharmaceutical or industrial applications.
Finally, as a proof-of-concept, we demonstrated the potential of this maleimide PISA system as a cell imaging agent by studying the range of nanoparticle morphologies in a cellular environment.A549 lung cancer fibroblasts were chosen as the model cellular environment, and each nanoparticle morphology was first tested to confirm cytocompatibility at concentrations up to 0.1 wt % for 48 h (Figures S35 and  S36).A549 cells were then incubated with small spheres, short worms, medium worms, and vesicles for 30 min, after which internalization was evaluated using confocal microscopy followed by FLIM to gain quantitative photophysical values within the cellular environment (Figure 5D).Confocal microscopy confirmed that all particle morphologies could be internalized within the cell cytosol, showing a diffuse pattern indicative of nonspecific uptake. 61Of note, when solely analyzing the fluorescence images, all nanoparticle morphologies display the same qualitative emission data.In comparison, the images acquired using FLIM show obvious qualitative differences, in agreement with the images taken on glass slides.Owing to the lower resolution of FLIM (512 × 512 pixels), there was insufficient resolution to detect minute changes in the fluorophore state upon nanoparticle degradation events; however, future work is focused on extending the lifetime of the maleimide through chemistry design in order to improve on this.What was apparent, however, was that each lifetime histogram showed a general trend of shifting to the left, indicating that a greater proportion of shorter lifetime events occurred (Figure 5E).This is perhaps due to nanoparticle processing within the cellular environment causing the disassembly of the morphologies into single polymer chains; however, further experimentation is required to understand this observation.
Overall, the above results demonstrate the potential of this system to assist with imaging different particles when the same fluorophore is present due to the ability of the FLIM experiment to differentiate between the different morphologies.Further to this, we have presented a very promising concentration-independent fluorescence reporter system to enable the quantitative microscopy analysis of drug delivery platforms.We envision this becoming a powerful tool for the analysis of nanoparticle states within complex biological environments, where the maleimide moiety can provide unambiguous tracking of particle uptake and degradation.More broadly, this work has demonstrated a highly accurate and sensitive technique for the study of PISA mechanisms, while importantly maintaining accessibility and simplicity for other researchers in the field.

■ CONCLUSIONS
Here, we present the design and synthesis of a modular polymer platform for studying polymerization-induced selfassembly by using steady-state fluorescence and lifetime measurements.By incorporating an ACM moiety within a RAFT CTA, we were able to demonstrate that the PISA process can be studied in detail using the photophysical properties of the reporter.Importantly, inclusion of the maleimide unit did not impact polymer self-assembly on account of its small size and enabled unique and unambiguous characterization of nanoparticle morphologies using a combination of emission and lifetime decay analysis.In addition, the analysis of deconvoluted lifetime decay components enabled us to study the mechanism of PISA in the preonset of the assembly time frame, yielding information about premicellar aggregates that form in the early stages when targeting higherorder morphologies such as worms and vesicles.We further analyzed the hydrophobicity of the growing macroCTA via computational log P oct /SA values to support the experimental observations during the PISA process.The applicability of this maleimide system to biological applications was demonstrated through confocal microscopy and FLIM, where the unique lifetime behavior of each was distinguishable within a cellular environment.We believe that the strategy presented herein will pave the way for further investigations into the use of polymeric materials as cell imaging agents and, more generally, for mechanistic studies of PISA in an accessible way.

Figure 1 .
Figure 1.(A) UV−vis spectra confirming the presence of trithiocarbonate and maleimide absorbance peaks, (B) SEC trace of the macroCTA showing RI and UV detection at 309 and 380 nm, (C) SEC traces from the RI detections of PISA polymers, (D) dry-state TEM images of PISA morphologies confirming (i) small spheres, (ii) large spheres, (iii) short worms, (iv) long worms, (v) vesicles, and (vi) large vesicles, and (E) background-subtracted small-angle X-ray scattering profiles of small spheres, long worms, and vesicles confirming observed morphologies from drystate TEM (N.B. the two upper patterns are offset by arbitrary factors for visual clarity and dashed lines indicate low q gradients of 0, −1, and −2 as a guide to the eye, indicating the presence of spheres, worms, and vesicles, respectively).

Figure 2 .
Figure 2. (A) Steady-state excitation, (B) emission spectra, and (C) fluorescence decay curves for the macroCTA and PISA morphologies in water.

Figure 4 .
Figure 4. Plots demonstrating relationships between (A) THFA conversion and log P oct /SA versus time, (B) fluorescence lifetime and log P oct /SA versus time, and (C) THFA conversion and fluorescence lifetime versus time.

Table 1 . Characterization Data for the Fluorescent PGMA 43 macroCTA and PGMA 43 -PTHFA x Diblock Copolymers
a Determined by SEC analysis in DMF containing 5 mM NH 4 BF 4 .b D̵ M = M w /M n .c Determined by DLS at 0.01 wt % in Milli-Q H 2 O. d Determined by TEM analysis.

Table 3 .
Photophysical Decay and Exponential Fitting on the Large Vesicles and Steady-State Emission Maxima , experimental procedures, and supplementary characterization data (PDF) Authors Amanda K. Pearce − School of Chemistry, University of Birmingham, Edgbaston, Birmingham B15 2TT, U.K.; Present Address: Department of Chemistry, Loughborough University, Loughborough, LE11 3TU, United Kingdom Macromolecules