|Home | About | Journals | Submit | Contact Us | Français|
We report herein thermally responsive elastin-like polypeptides (ELPs) in a linear AB diblock architecture with an N-terminal peptide ligand that self-assemble into spherical micelles when heated slightly above body temperature. A series of 10 ELP block copolymers (ELPBCs) with different molecular weights and hydrophilic-to-hydrophobic block ratios were genetically synthesized by recursive directional ligation. The self-assembly of these polymers from unimers into micelles was investigated by light scattering, fluorescence spectroscopy and cryo-TEM. These ELPBCs undergo two phase transitions as a function of solution temperature: a unimer to spherical micelle transition at an intermediate temperature, and a micelle to bulk aggregate transition at a higher temperature when the hydrophilic-to-hydrophobic block ratio is between 1:2 and 2:1. The critical micelle temperature is controlled by the length of the hydrophobic block and the size of the micelle is controlled by both the total ELPBC length and hydrophilic-to-hydrophobic block ratio. These polypeptide micelles display a critical micelle concentration in the range of 4-8 μM demonstrating high stability of these structures. These studies have also identified a subset of ELPBCs bearing terminal peptide ligands that are capable of forming multivalent spherical micelles that present multiple copies of the ligand on their corona in the clinically relevant temperature range of 37-42 °C and target cancer cells. These ELPBCs may be useful for drug targeting by thermally triggered multivalency. More broadly, the design rules uncovered by this study should be applicable to the design of other thermally reversible nanoparticles for diverse applications in medicine and biology.
Molecular self-assembly describes the spontaneous association of molecules under equilibrium conditions into well-defined stable aggregates joined by noncovalent bonds1. One of the simplest examples of self-assembly by polymers is demonstrated by linear AB diblock copolymers with a large solubility difference between the two blocks 2, 3. When an amphiphilic AB diblock copolymer is placed in a solvent that is only good for one block, an attractive and a repulsive force is generated that results in microphase separation4 and self-assembly of individual molecules into supramolecular structures. For amphiphilic AB diblock copolymers, self-assembly often results in the formation of a spherical micelle consisting of a core made up of an insoluble block that is shielded from the solvent by a hydrated corona composed of a more soluble block2. Other structures such as cylindrical micelles, toroids, lamellar sheets or vesicles can also be formed, and the formation of these structures depends on the length of the A and B segments relative to each other5. Bioinspired amphiphilic molecules that exhibit this behavior range from simple di- or tri-peptides6, 7 to more complex, naturally-occurring peptide motifs such as coiled-coils (i.e., “leucine zipper”)8-10, collagen, elastin-like polypeptides (ELPs) 11, 12, silk-like polypeptides13 and combinations thereof14, as well as conjugates of peptides with alkyl chains (i.e. peptide amphiphiles)15.
In contrast to the spontaneous self-assembly of block copolymers into supramolecular structures, we are interested in the design of “smart” –stimuli-responsive– peptide-based block copolymers11, 12, 14, 16-19 that can be triggered to self-assemble by a relatively small change in solution conditions. We are especially interested in the challenge of designing polymers that can be triggered to self-assemble in response to clinically relevant intrinsic stimuli such as the difference in pH between normal tissues (~7.4) and the tumor microenvironment (6.5-6.8)20, differences in calcium concentration between the extracellular and intracellular environment 21, 22 or extrinsic stimuli such as the elevated temperature employed in mild hyperthermia treatment of solid tumors23. This interest is motivated by the recognition that nanoparticles that can self-assemble in vivo by physiologically or clinically relevant triggers have enormous applications in the emerging field of nanomedicine15, 24-26.
In our first attempt at the design of triggered self-assembled nanoparticles, we have focused on the synthesis of a polypeptide that can be triggered to self-assemble into a spherical micelle by a small increase of temperature to between 37 and 42 °C, the temperature range approved for clinical application of hyperthermia23. We chose elastin-like polypeptides (ELPs) as the peptide building block of these AB diblock copolymers to explore the feasibility of this goal. ELPs are genetically encoded biopolymers of the pentameric repeat Val-Pro-Gly-Xaa-Gly where Xaa, termed the guest residue, is any amino acid besides proline27. The choice of ELPs as the building blocks for these systems was dictated by the following reasons. First, ELPs exhibit inverse temperature phase transition behavior (also called a lower critical solution temperature transition [LCST]); they are soluble at temperatures below their transition temperature (Tt) but become insoluble and aggregate at temperatures above their T 27-29t. Their Tt is tuned by adjusting the guest residue composition, molecular weight (MW) and ELP concentration11, 30, 31. This feature of ELPs is useful because it allows us to begin the design of temperature triggered self-assembling polymers using a repeat motif that is intrinsically responsive to its environment. Second, ELPs are not solely responsive to temperature; the phase transition of ELPs can also be triggered by ionic strength, pH and light29, 32. However, the use of these triggers in the context of self-assembly, as opposed to coacervation33, only has a few precedents 11, 34, 35. We emphasize that for such systems to move beyond proof-of-principle and be useful in the emerging field of nanomedicine, triggered self-assembly by physiologically relevant triggers is essential and, is thus, the driving force behind this study.
We synthesized a series of ELP block copolymers (ELPBCs) in a linear AB diblock architecture by seamlessly fusing an N-terminal ELP gene with a high Tt (Tt > 90 °C, more hydrophilic Xaa) to a C-terminal ELP gene with a much lower Tt (Tt ~ 40 °C, more hydrophobic Xaa), as shown in Figure 1. These ELPBCs are highly soluble at a solution temperature below the Tt of both ELP blocks. However, upon an increase in solution temperature, we hypothesized that ELPBCs would self-assemble into a spherical micelle when the low Tt block undergoes its inverse temperature phase transition11. In this study, we synthesized 10 ELPBCs of various overall lengths (i.e. MWs) and hydrophilic-to-hydrophobic block MW ratios to (1) determine a fundamental set of design rules that results in the self-assembly of ELPBCs into defined structures; (2) isolate a construct that self-assembles into a multivalent spherical micelle within a clinically relevant hyperthermic temperature range of 37 to 42 °C; and (3) determine the physical properties of these self-assembled constructs.
The different ELP constructs are named using the notation ELP[XiYjZk-n]11. The bracketed capitol letters are the single amino acid code specifying the guest residues in the ELP sequence and the subscripts designate the number of Val-Pro-Gly-Xaa-Gly repeats for each corresponding guest residue in the monomer gene. n represents the number of pentapeptides. All ELPBCs herein are composed of an N-terminal ELP[V1A8G7-n] gene followed by a C-terminal ELP[V5-n] gene. The ELP[V1A8G7-n] gene has a monomer repeat unit of 16 pentapeptides and ELP[V5-n] has a monomer repeat unit of 5 pentapeptides. For semantic clarity, all ELPBC genes are named using the notation ELP-x/y where x is the number of pentapeptides for the ELP[V1A8G7-n] gene and y is the number of pentapeptides for the ELP[V5-n] gene. For example, ELP-64/90 is composed of an N-terminal ELP[V1A8G7-64] and C-terminal ELP[V5-90]. Because ELP[V1A8G7-n] and ELP[V5-n] have different lengths of their basic repeat unit, approximate hydrophilic-to-hydrophobic block ratios are rounded to the nearest consistent number. For example, ELP-64/90 has a reported hydrophilic-tohydrophobic block ratio of 2:3.
A complete materials and methods section may be found in supplemental information.
The goals of these studies were to: (1) discover the design rules for the self-assembly of ELPBCs into nano-mesoscale structures; (2) isolate a construct that self-assembles into a multivalent micelle within the clinically relevant hyperthermic temperature range of 37 to 42 °C; and (3) determine the physical properties of these micelles. To accomplish these goals, we synthesized a series of 10 ELPBCs with various MWs and hydrophilic-to-hydrophobic ratios and examined their self-assembly with DLS, SLS, fluorescence spectroscopy, and cryo-TEM. Finally, an ELPBC that formed stable micelles in the range of 37-42 °C was then modified at the gene level to present an RGD or NGR peptide, and these ligandfunctionalized ELPBCs were further characterized by light scattering to investigate their ability to self-assemble into multivalent nanoparticles in the desired range of 37-42 °C.
Single segment ELPs undergo a unimer-to-aggregate transition when heated above their Tt, producing a sharp (1-2 °C) increase in turbidity of an ELP solution11, 30. In contrast, some ELPBCs exhibited a two-step thermal response; the ELPBC solution was transparent at low temperatures (< 36 °C), but as the temperature was increased, the absorbance increased and remained elevated above baseline levels (range = 36 – 46 °C) (Figure 2). Upon further heating, the solution turbidity increased sharply and plateaued (range = 49 – 51 °C) at a constant value. We believe that the first small increase in absorbance was caused by the inverse temperature phase transition of the low Tt ELP block (Tt1). Upon heating above 49 °C, the high Tt ELP block underwent its inverse temperature phase transition (Tt2) which induced the formation of large aggregates indicated by the high OD value. Tt1 was defined as the temperature at which the OD first deviated from the baseline, and Tt2 was defined as the temperature at the maximum in the derivative of OD with respect to temperature (dOD/dT).
We believe this two-step change in turbidity is caused by a monomer-micelle transition followed by a micelle-coacervate transition characterized by the formation of micron-sized aggregates. Previous experiments by our laboratory have demonstrated that a small increase in absorbance, as shown in Figure 2 from 36 – 46 °C, is indicative of the formation of nanoparticles (defined as a diameter < 100 nm)11. This finding motivated the screening of all ten ELPBCs with fixed-angle DLS, which indicated that six out of ten ELPBCs formed nanoparticles. These screening experiments also revealed that ELPBCs with a hydrophilic-to-hydrophobic ratio between 1:2 and 2:1 formed nanoparticles (1:2 ≤ ratio ≤ 2:1). This lower threshold of a 1:2 ratio is similar to an f value of 33%, which was shown by Discher et al36 to form vesicles. Increasing the hydrophilic-to-hydrophobic ratio above 1:2 also resulted in nanoparticle formation, analogous to Discher's findings36. However, ratios exceeding 2:1 did not form nanoparticles and instead behaved as a simple monoblock ELP, only exhibiting a single transition from unimer to micron-sized aggregates. This result is likely due to an insufficient solubility difference between the blocks to self-assemble into a nanoparticle.
The concentration dependence of Tt1 and Tt2 was determined for the six ELPBCs that formed nanoparticles (Figure 3). Both Tts were fit as a logarithmic function of concentration, shown as a solid line in Figure 3. Similar to simple monoblock ELPs, Tt1 had a decreasing logarithmic dependence on ELP concentration30. Furthermore, Tt1 appeared to scale with the Tt of the parent ELP[V5-n] block (see the dashed lines in Figure 3). For example, the average Tt1 for each ELPBC with an ELP[V5-60] block was consistently 4.2 °C higher than the parent ELP[V5-60], most likely due to the influence of the more hydrophilic ELP[V1A8G7-n] on ELP[V5-60]'s phase transition. This increase in Tt diminished to 1.5 and 1.2 °C for ELPBCs composed of ELP[V5-90] and ELP[V5-120], respectively. Therefore, these results suggested that Tt1 is controlled by the length (i.e., MW) of the low Tt block (ELP[V5-n]). In contrast, Tt2 appeared independent of both ELP[V1A8G7-n] and ELP[V5-n] block lengths and had a much weaker dependence on concentration (m = -0.55 to -1.48 for Tt2, and m = -1.59 to -2.60 for Tt1). This relative lack of length dependence and weaker concentration dependence for the Tt2 may be due to the close proximity of ELP[V1A8G7-n] in the corona of the micelle, effectively maximizing its local concentration. The high local ELP concentration in the corona causes its Tt2 to be independent of concentration, a trend also observed for monoblock ELPs at high concentrations29, 30.
The temperature dependent self-assembly of ELPBCs was next investigated by DLS performed at multiple angles to determine their size, CMT, polydispersity, and micelle formation temperature range. The ELPBC unimers (Rh range = 5.3 – 9.8 nm) formed a monodisperse nanoparticle (Rh range = 29.1 – 39.9 nm) at a temperature near Tt1, as shown in Figure 4. This aggregation temperature determined by DLS was defined as the CMT. In some instances, the monodisperse nanoparticle was stable for 5 to 8 °C until a larger polydisperse aggregate (>1 μm) was formed, representing bulk aggregation after the corona block underwent its phase transition (Figure 4A). However, in other cases the monodisperse nanoparticle transitioned into a more polydisperse nanoparticle that grew in size until bulk aggregation was observed (Figure 4B). This transition from a monodisperse to polydisperse nanoparticle may be due to simple growth of a spherical aggregate or due to a sphere-to-rod transition. The stable monodisperse nanoparticles were classified as type I nanoparticles (Figure 4A) and the less stable nanoparticles were classified as type II nanoparticles (Figure 4B). This independent stratification of ELPBCs based on their phase transition behavior was structurally illuminating because all stable type I nanoparticles were composed of ELP[V5-60] blocks and all type II nanoparticles were composed of ELP[V5-n] blocks with lengths greater than 60 pentapeptides. The thermal properties of the ELPBCs, including their CMT and classifications are summarized in Table 1. Consistent with previous UV-visible spectrophotometry data, the CMT occurred about 4 – 6 °C higher than the parent ELP[V5]'s Tt and this CMT was controlled by the length of the ELP[V5] block.
To further elucidate the structure and polydispersity of these ELPBC nanoparticles, DLS and SLS at multiple angles were performed at a temperature within the stable nanoparticle temperature range. The Rg, MWapp, and Z of each ELPBC were determined with SLS. The structure of the ELPBC nanoparticle was characterized further by combining the DLS and SLS data together to calculate the ρ ratio (ρ = Rg/Rh). The ratio equals 0.775 for a homogeneous hard sphere (e.g., spherical micelle), 1.0 for a thin hard spherical shell (e.g., vesicle), 1.504 for a random monodisperse coil in a θ solvent and 1.73 for a random monodisperse coil in a good solvent (it is important to note that polydispersity will increase the ratio because large molecules of a broad distribution will contribute more to Rg than Rh provided that internal modes of motion are absent)37, 38.
An example of DLS and SLS analysis is shown in Figure 5 for ELP-64/90. In general, nanoparticles in this size range are considered to be a monodisperse sphere if their Dapp does not vary by more than 10% (positive sloped line) of the mean Dapp as the form factor (P(q)) is highly dependent upon scattering angle39. This ELPBC nanoparticle (Rh = 30 nm) was a monodisperse hard sphere as indicated by the lack of angular dependency of Dapp in the DLS data (Figure 5A)39-41. The Rg (19.0 nm), MWapp (6.9 × 106), and Z (110) of this ELPBC nanoparticle were determined with SLS as shown in Figure 5B. The ratio for ELP-64/90 was 0.633, indicating that the nanoparticle was a hard sphere and therefore a spherical micelle. A summary of the properties of each nanoparticle-forming ELPBC including Rh, Rg, MWapp, ρ ratio, and Z is shown in Table 2.
These data in Table 2 prove that ELPBCs with a hydrophilic-to-hydrophobic ratio between 1:2 and 2:1 form a monodisperse spherical micelle at a temperature between Tt1 and Tt2. In addition, the size of the micelle was controlled by both the total length of the ELPBC and the hydrophilic-to-hydrophobic block ratio supported by the models of Israelachvili et al.42 and Discher et al36. For example, increasing the total length while maintaining a constant block ratio increased the micelle size from 29.1 nm for ELP-64/60 (length = 120) to 36.6 nm for ELP-96/90 (length = 180). Furthermore, increasing the hydrophobic block length relative to the hydrophilic block increased the micelle size for an ELPBC with a constant length due to the increased hydrophobic volume fraction. Although the ratios (range = 0.633 to 0.753) were slightly lower than the theoretical value of 0.775, they further confirm the formation of a hard spherical micelle structure. These lower values may be due to inhomogeneous packing within the micelle in the radial direction. The high aggregation numbers (Z = 56 – 133) suggest that a sufficient number of targeting ligands may be presented in the micelle's corona to enhance the binding avidity.
The CMC was determined using fluorescence spectroscopy and pyrene, a fluorescent probe sensitive to the polarity of its microenvironment. The ratio of the first and third peaks (I1/I3) of the monomer emission spectrum decreases as the probe partitions into a less polar region such as a micelle's hydrophobic core. For an ELP-64/90 solution, this ratio remained relatively constant at temperatures below its CMT (T = 20 °C – 35 °C) as shown in Figure 6A. At temperatures between Tt1 and Tt2, the ratio monotonically decreased until it reached a minimum at Tt2 (T = 50 °C) indicating that the overall polarity of the micelle decreased throughout the micelle temperature range. There was a slight increase in average polarity of the solution at temperatures above the bulk aggregation temperature (T > 50 °C) suggesting that pyrene partitioned into a more aqueous environment. Similar findings by Chung et al. were interpreted as release of the entrapped molecule from thermo-responsive micelles43. The CMC was determined by plotting the minimum of the I1/I3 ratio from 35 °C – 45 °C (see Figure 6A) as a function of ELPBC concentration (Figure 6B). The inflection point of a sigmoid fit to data in this temperature range was defined as the CMC. The CMC was determined for all six ELPBCs capable of forming nanoparticles (Table 2). The I1/I3 ratio remained constant at low concentrations of ELP-64/90 but decreased to a lower value as the concentration was increased. These data demonstrate that ELP-64/90 micelles formed at a low concentration of 8.1 μM. Furthermore, all six of the nanoparticle-forming ELPBCs have a CMC < 10 μM, implying that ELPBC micelles are quite stable structures44.
The microviscosity of the micelles was further investigated with fluorescence spectroscopy using a fluorescent PC3P probe45, 46. The emission spectrum of this probe has two characteristic peaks representing the contribution from monomer and excimer forms of the probe. The ratio of these two peaks (IE/IM), is proportional to the microviscosity of PC3P's environment; PC3P molecules with restricted motion will form intramolecular excimers while freely-rotating PC3P molecules will remain in monomer form. At solution temperatures below the first Tt, the microviscosity of an ELP-64/90 solution is similar to that of a control ELP[V5A2G3-150], as shown in Figure 7 (T = 30 – 40 °C). The microviscosity increased within the micelle temperature range (T = 40 – 50 °C) as indicated by the higher IE/IM of an ELP-64/90 solution. The control ELP[V5A2G3-150] exhibited a lower microviscosity than the micelle after it underwent its inverse temperature phase transition (Tt = 42 °C) into a bulk aggregate. Furthermore, the ELP-64/90 solution demonstrated a similar microviscosity to the control ELP[V5A2G3-150] after it underwent bulk aggregation at 50 °C. These data suggest that the core of an ELPBC micelle had a greater microviscosity than both a soluble ELP solution and an ELP bulk aggregate. Also, a reduction of microviscosity as the ELPBC goes through its micelle-to-aggregate transition suggests that the micelles reorganize their structure during their transition to a bulk aggregate.
The micellar structures of these ELPBCs were further validated by cryo-TEM obtained using a vitrified ELPBC sample in the nanoparticle phase. A selected set of ELPBC micelle images are shown in Figure 8. The homogeneous intensity throughout the interior of the nanoparticles is indicative of a micellar structure.
ELPBCs that formed stable, type I spherical micelles between 37-42 °C (ELP-64/60 and ELP96/60) were selected as candidates to create temperature triggered multivalent nanoparticles. We chose two different peptide motifs, the RGD and NGR tripeptides, as ligands to be presented in the corona following micelle self-assembly. These peptide sequences were selected because they are known targeting ligands for angiogenic tumor vasculature47, 48 and multivalent presentation of these ligands can improve tumor targeting and therapy49, 50.
The self-assembly of ligand-targeted ELPBCs were compared to their parent ELPBCs (i.e., same ELPBC without an N-terminal ligand) using DLS as shown in Figure 9. Importantly, DLS verified that ELPBCs with targeting ligands incorporated into their N-terminus also formed micelles of a similar size and in a similar temperature range as ELPBCs without targeting ligands, which suggest their utility for in vivo targeting. Furthermore, the fact that neither peptide ligand perturbed the temperature triggered self-assembly behavior of the ELPBC suggests that this diblock copolymer system is highly tolerant to the incorporation of other peptide motifs for multivalent presentation. These results are notable because they suggest the feasibility of modular design of multivalent nanoparticles simply by swapping different peptide motifs, as well as the intriguing possibility of synthesizing multifunctional, multivalent nanoparticles by mixing several peptide-functionalized ELPBCs prior to thermally triggered self-assembly.
Cellular uptake of NGR functionalized ELPBCs above and below body temperature was investigated with laser scanning confocal microscopy, shown in Figure 10. HT-1080 cells were used to examine internalization of NGR-functionalized ELPBCs because this cell line was positive for NGR's receptor, aminopeptidase N (APN)51 (data not shown). Cells treated with ELP-96/60-Alexa488 and NGR-ELP-96/60-Alexa488 exhibited modest internalization and punctuate fluorescence at temperatures below body temperature (See Figure 10A & C), as has been shown previously52. Above body temperature, ELP-96/60-Alexa488 revealed a similar pattern of uptake as low temperature controls, but ligand-targeted NGR-ELP-96/60-Alexa488 demonstrated enhanced cellular uptake (see Figure 10B). The number of overlapping red-green pixels in the projection of NGR-ELP-96/60-Alexa488 was approximately 19-fold greater than for unimers of NGR-ELP-96/60-Alexa488 (monovalent control) and 3.5-fold greater than the micellar ELP-96/60-Alexa488 that did not present a NGR ligand (micelle control). These images suggest that micellar NGR-ELP-96/60-Alexa488 accumulated in cells to a greater extent than either unimeric NGR-ELP-96/60-Alexa488 or micellar ELP-96/60-Alexa488 controls.
The results presented here demonstrate that AB diblock polypeptides can be triggered to self-assemble into a spherical micelle in a narrow, clinically relevant temperature range of 37-42 °C. We envision the use of these polymers as multivalent carriers of drugs that self-assemble only in a defined region in vivo, such as a tumor heated with externally applied hyperthermia, to increase the drug carrier's binding affinity by converting it from a monovalent targeting agent to a multivalent nanoparticle. The increased affinity of the multivalent micelles relative to the unimer will selectively target the micelles to a specific tissue in vivo through high affinity interactions while minimizing background uptake of the unimer by normal tissues. We call this novel technique of controlling valency through triggered self-assembly affinity modulated targeting.
Affinity modulated targeting has three potential advantages over commonly used affinity targeted strategies53, especially within the context of cancer therapy: (1) the higher avidity of the nanoparticle to a heated tumor should localize more anticancer agents to that tumor; (2) the lower affinity in systemic circulation will reduce affinity to normal tissue, reducing systemic exposure; and (3) modulated affinity allows the targeting of more widely expressed tumor associated antigens. Tumor accumulation of an affinity targeted drug carrier appears to require a minimum threshold affinity of 10-7 M (KD) and a maximum affinity (~10-9 M) above which tumor accumulation is not improved54. Affinity modulated targeting hence requires a low affinity ligand for a tumor receptor (KD ≥ 10-7 M) when it is in its monovalent form. However, when an ELPBC functionalized with this ligand self-assembles into a micelle that presents multiple copies of a ligand in a hyperthermic tumor, the avidity of the multivalent nanoparticle should theoretically increase between 103- and 108-fold from its monovalent affinity55 and consequently result in maximal tumor accumulation (KD ≤ 10-9 M). The low affinity in systemic circulation should reduce normal tissue exposure if the receptor is expressed in tissues other than the tumor. In addition, the selectively higher avidity within the tumor facilitates the use of tumor associated antigens that are ubiquitously expressed on many clinically relevant cancers, thus making this strategy potentially more widely applicable than other affinity targeting techniques incorporating a single tumor specific antigen.
A series of 10 ELPBCs with a range of MWs and hydrophilic-to-hydrophobic ratios were recombinantly synthesized by recursive directional ligation of a monomer gene and plasmid-borne expression of the oligomerized gene in E. coli. ELPBCs with a hydrophilic-to-hydrophobic ratio between 1:2 and 2:1 formed monodisperse spherical micelles. The CMT was controlled by the length of the low Tt block and scaled with the parent ELP block's Tt. The size of the micelle was controlled by both the total ELPBC length and hydrophilic-to-hydrophobic block ratio. These polypeptide micelles displayed a CMC in the range of 4-8 μM, indicating that these structures are highly stable. Furthermore, the microviscosity of the ELPBCs increased as they self-assembled into a micelle from unimers. The microviscosity of the micelles was greater than the bulk aggregate, suggesting that the micellar structure was altered during reorganization into the micron size aggregates of the coacervate that occured at the second thermal transition. The major findings of this study are that: (1) it provides the first set of design rules for the self-assembly of AB diblock ELP copolymers into spherical micelles; (2) it has identified a set of ELPBCs that exhibit thermally triggered self-assembly into spherical micelles in the desired range of 37 – 42 °C; and (3) it is the first demonstration of temperature triggered multivalency– that ELPBCs presenting biologically relevant peptide ligands at their termini self-assembled into spherical micelles at a clinically-relevant temperature and target tumor cells. More broadly, the design rules uncovered by this study should be applicable to the design of other thermally reversible nanoparticles for diverse applications in medicine and biology.
This work was supported by grants NIH EB–000188 (A.C.) and Sonderforschungsbereich 625 (M.S.). Matthew Dreher would like to acknowledge the Intramural Research Program of the National Institutes of Health for support. We thank Daniel Meyer for synthesis of some of the ELP genes and Michael Stepputat for assistance with the cryo-TEM.
Supporting Information Available: Experimental procedures can be found in the supporting information. This material is available free of charge via the Internet at http://pubs.acs.org.