Self-Assembly and Enzyme Responsiveness of Amphiphilic Linear-Dendritic Block Copolymers Based on Poly(N-vinylpyrrolidone) and Dendritic Phenylalanyl-lysine Dipeptides

In this study, we present the synthesis, self-assembly, and enzyme responsive nature of a unique class of well-defined amphiphilic linear-dendritic block copolymers (PNVP-b-dendr(Phe-Lys)n, n = 1–3) based on linear poly(N-vinylpyrrolidone) (PNVP) and dendritic phenylalanyl-lysine (Phe-Lys) dipeptides. The copolymers were prepared via a combination ofreversible addition-fragmentation chain transfer (RAFT)/xanthates (MADIX) polymerization of N-vinylpyrrolidone and stepwise peptide chemistry. The results of fluorescence spectroscopy, 1H NMR analyses, transmission electron microscopy (TEM), and particle size analysis demonstrated that the copolymers self-assemble in aqueous solution into micellar nanocontainers that can disassemble and release encapsulated anticancer drug doxorubicin or hydrophobic dye Nile red by trigger of a serine protease trypsin under physiological conditions. The disassembly of the formed micelles and release rates of the drug or dye can be adjusted by changing the generation of dendrons in PNVP-b-dendr(Phe-Lys)n. Furthermore, the cytocompatibility of the copolymers have been confirmed using human lung epithelial cells (BEAS-2B) and human liver cancer cells (SMMC-7721). Due to the fact of their enzyme responsive properties and good biocompatibility, the copolymers may have potential applicability in smart controlled release systems capable of site-specific response.


Introduction
Enzyme responsive polymers are highly promising carriers for targeted drug delivery because of their excellent biocompatibility and high degree of selectivity [1][2][3][4]. Some enzymes are frequently overexpressed in inflamed or tumor tissues [5,6]. By incorporating enzyme labile linkages, macromolecular carriers are able to be engineered to release the loaded drug on demand, which holds great potential in reducing the exposure to healthy cells and tissues [7]. According to the expression level of the enzyme in a specific region, the release rate of the drug can be controlled by interactive adjustment [8,9]. In addition, enzyme responsive polymers have shown high potential for long-term cycling and targeted therapy and diagnosis of disease in vivo [10]. However, although some enzyme responsive linear-linear block copolymers have been investigated, until recently, only a few studies related to enzyme responsive linear-dendritic block copolymers (LDBCs) have been reported [11,12].
Linear-dendritic block copolymers (LDBCs) combine the easy synthesis, availability, and processability of linear polymeric chains with the regular and well-defined structure of dendrons, which have an The 1 H NMR spectra were obtained on a Bruker DRX-500 (Rheinstetten, Germany) spectrometer in CDCl 3 , DMSO-d 6 or D 2 O. Tetramethylsilicone (TMS) was used as internal standards. Molecular weights and molecular weight distributions were measured by gel permeation chromatography (GPC; Waters, Milford, CT, USA) with a Waters 2690D separations module and a Waters 2414 refractive index detector (R). Styragel HR1 and HR4 columns were used at 40 • C with poly(methyl methacrylate) (PMMA) as standards and DMF (included 0.05 M LiBr) as a mobile phase at a flow rate of 0.3 mL/min.

Synthesis of PNVP
The PNVP was synthesized according to the previously published literature [44] in 86.1% yield. 1

Synthesis of PNVP-NH 2
The PNVP-NHBoc (5.6 g, 0.2 mmol) was dissolved in 60 mL of TFA. After stirring for 4 h at room temperature, the solution of pH was adjusted to 9 with 5 N NaOH aqueous solution. The resulting solution was dialyzed and subsequently freeze-dried, affording PNVP-NH 2 as a white solid in 91.2% yield (5.1 g). 1

Synthesis of PNVP-b-Phe-NH 2
The PNVP-b-Phe-NHFmoc (1.95 g, 0.093 mmol) was dissolved in 10 mL of mixed solution of 4-methyl piperidine and DMF (V/V = 1:4). After stirring for 2 h at room temperature, the reaction mixture was poured into cold ether to precipitate the product. The precipitate was redissolved in water, dialyzed, freeze-dried, affording PNVP-b-Phe-NH 2 as a white solid in 94.3% yield (1.84 g). 1

Critical Micelle Concentration (CMC) Measurement
Critical Micelle Concentration (CMC) were determined with a Hitich F-4500 luminescence spectrophotometer (Hitachi, Tokyo, Japan) using Nile red as a fluorescent probe [46]. Nile red stock solution (45µL, 0.88 mg/ml in ethanol) were added into 100 mL phosphate buffer solution (pH 7.4), and mixed to give a final concentration of 1.25 µM. A 1.0 mg/mL solution of each polymer (PNVP-b-dendr(Phe-Lys) n (n = 1-3) was prepared in diluent and sonicated for 15 min. This solution was repeatedly diluted by a factor of 2 with diluent. The solutions containing Nile red were kept at room temperature for 24 h before measurements. The fluorescence emission intensity was scanned for solution of each concentration. Maximum emission intensity was plotted versus polymer concentration in order to determine the CMC. All measurements were repeated three times.

Monitoring Trypsin-Induced Disassembly of the Nile Red-Loaded Micelles
Monitoring trypsin-induced disassembly of the Nile red-loaded micelles was performed using a Hitich F-4500 luminescence spectrophotometer (Hitachi, Tokyo, Japan). The entire experiment was carried out in a 37 • C water bath. The PNVP-b-dendr(Phe-Lys) n (n = 2, 3) was dissolved in PBS buffer (pH 7.4) to give a concentration of 1.0 mg/mL solution. The solution was sonicated for 15 minutes and then filtered through a 0.45 µm nylon syringe filter. This solution (2.0mL) was transferred to a quartz cuvette and 1.44 µL of Nile red stock solution (0.88 mg/mL in ethanol) was added to give a final concentration of 2.0 µM. A fluorescence emission scan was performed (t = 0) and then 10 µL of trypsin stock solution (16 U/µL in PBS buffer pH 7.4 at 37 • C) was added to give a final concentration of 80 U/mL. Repeating fluorescence scans were performed every 1 h for 48 h.

Particle Sizing and Transmission Electron Microscope (TEM) Measurements
The sizing of the micelles was determined by a ZataPALS particle size analyzer (Brookhaven Instruments Corp, New York, USA). A drop of micellar solution (1 mg/mL, or Nile red loaded micelles, or dealt with 80 U/mL of trypsin stock solution) was placed on a FORMVAR-coated copper grid and dried in air. The TEM images of the micelles were obtained using a JEM-2100 transmission electron microscope (JEOL, Akishima, Tokyo, Japan) at an acceleration voltage of 200 kV.

Loading and Enzyme-Triggered Release of Doxorubicin (DOX) into/from the LDBCs Micelles
The DOX-loaded micelles were prepared according to the literature [47]. Drug loading (DL) and encapsulation efficiency (EE) were calculated using the following equations: DL (w/w) = (amount of loaded drug)/(amount of polymer), EE (w %) = (actual amount of loaded drug)/(amount of added drug). In vitro release from the DOX-loaded micelles (1 mg/mL) was carried out using dialysis bags in PBS buffer (pH = 7.4) at 37 • C. The enzyme-triggered cargo release was achieved by adding the trypsin (80 U/mL) to the DOX-loaded solution. The amount of DOX released was determined by using fluorescence (Hitich F-4500) measurement (excitation at 480 nm).

Synthesis of Amphiphilic LDBCs
The synthetic route of three generations of linear-dendritic block copolymers PNVP-b-dendr (Phe-Lys) n (n = 1-3) was shown in Scheme 1. Firstly, well-defined PNVP with an O-ethylxanthyl end-group was synthesized by reversible addition-fragmentation chain transfer (RAFT)/xanthates (MADIX) polymerization of NVP with a [1-(O-ethylxanthyl)ethyl]benzene as a RAFT/MADIX agent. The O-ethylxanthyl end-group of PNVP subsequently was reduced via aminolysis, and the resulting thiol was trapped with a Michael acceptor, tert-butyl (2-acrylamide ethylene) carbamate in situ to yield a BocNH-terminated PNVP (PNVP-NHBoc). Then, the protecting Boc group from amino group of PNVP-NHBoc was removed with trifluoroacetic acid to afford amino-terminated PNVP (PNVP-NH 2 ). The PNVP-NH 2 was used as the polymeric supporter, and then PNVP-b-dendr(Phe-Lys) n (n = 1-3) were synthesized via stepwise peptide chemistry. Each synthesis step was monitored by a ninhydrin test, 1 H NMR spectra and GPC.   The GPC analysis of the copolymers revealed monomodal peaks with narrow molecular weight distributions (M w /M n < 1.35; Figure 2, Table 1). The molecular weights M n,GPC of the copolymers increased in the order of PNVP < PNVP-b-dendr(Phe-Lys) < PNVP-b-dendr(Phe-Lys) 2 < PNVP-b-dendr(Phe-Lys) 3 . The M n,NMR values of the LDBCs were determined by comparing the integration ratio of Ar protons of Fmoc in dendritic blocks at 7.28-7.95 ppm (z) with CHN protons of the PNVP backbone at 3.45-4.00 ppm (e). The M n,NMR values of LDBCs increased with increasing the generation of the dendritic dipeptides, indicating three generations of copolymers were synthesized. signals of the PNVP repeat-units in the 1 H NMR spectra of PNVP-b-dendr(Phe-Lys) (Figure 1a) can be expressed in the peaks marked with letters from d to h, d at 1.74-1.42 ppm (m, CH2), e at 3.98-3.46 ppm (m, NCH), f at 3.31-2.85 ppm (m, NCH2), g at 1.99-1.75 ppm (m, CH2), h at 2.40-1.99 ppm (m, COCH2), and the resonance signals of dendron protons of Fmoc group and phenyl group at 7.95-7.13 ppm, α-methine group of Lys, α-methine group of Phe, methine group and methine group of Fmoc group at 4.43-4.16 ppm. The GPC analysis of the copolymers revealed monomodal peaks with narrow molecular weight distributions (Mw/Mn < 1.35; Figure 2, Table 1). The molecular weights Mn,GPC of the copolymers increased in the order of PNVP < PNVP-b-dendr(Phe-Lys) < PNVP-b-dendr(Phe-Lys)2 < PNVP-bdendr(Phe-Lys)3. The Mn,NMR values of the LDBCs were determined by comparing the integration ratio of Ar protons of Fmoc in dendritic blocks at 7.28-7.95 ppm (z) with CHN protons of the PNVP backbone at 3.45-4.00 ppm (e). The Mn,NMR values of LDBCs increased with increasing the generation of the dendritic dipeptides, indicating three generations of copolymers were synthesized.

Self-Assembly of Amphiphilic LDBCs
The amphiphilic PNVP-b-dendr(Phe-Lys)n were expected to be able to self-assemble in aqueous solution to form micelles with a hydrophobic dendritic Phe-Lys with peripheral Fmoc groups' cores surrounded by a hydrophilic PNVP shell. To test this, the self-assembled micelles with core-shell structures of PNVP-b-dendr(Phe-Lys)n were first examined by comparing the 1 H NMR spectra of the LDBCs in DMSO-d6 with in D2O (Figures 1, S1, and S2). All protons of the LDBCs can clearly be seen in 1 H NMR spectra in DMSO-d6, because DMSO-d6 is a good solvent for the amphiphilic LDBCs with both types of blocks. However, in the 1 H NMR spectra of PNVP-b-dendr(Phe-Lys)n in D2O, the peaks of the Fmoc group disappeared at 7.28-7.95 ppm and the benzene proton peaks at 7.12-7.30 ppm (Figures 1b, S1b, and S2b) originating from Phe are highly suppressed, indicating that the mobility of Fmoc and Phe-Lys groups was limited by the formation of core-shell micellar structures with

Self-Assembly of Amphiphilic LDBCs
The amphiphilic PNVP-b-dendr(Phe-Lys) n were expected to be able to self-assemble in aqueous solution to form micelles with a hydrophobic dendritic Phe-Lys with peripheral Fmoc groups' cores surrounded by a hydrophilic PNVP shell. To test this, the self-assembled micelles with core-shell structures of PNVP-b-dendr(Phe-Lys) n were first examined by comparing the 1 H NMR spectra of the LDBCs in DMSO-d 6 with in D 2 O (Figure 1, Figures S1 and S2). All protons of the LDBCs can clearly be seen in 1 H NMR spectra in DMSO-d 6 , because DMSO-d 6 is a good solvent for the amphiphilic LDBCs with both types of blocks. However, in the 1 H NMR spectra of PNVP-b-dendr(Phe-Lys) n in D 2 O, the peaks of the Fmoc group disappeared at 7.28-7.95 ppm and the benzene proton peaks at 7.12-7.30 ppm (Figure 1b, Figures S1b and S2b) originating from Phe are highly suppressed, indicating that the mobility of Fmoc and Phe-Lys groups was limited by the formation of core-shell micellar structures with hydrophobic inner cores and hydrophilic outer shells. The CMC of the LDBCs was determined by encapsulating Nile red as a hydrophobic fluorescent probe. As shown in Figures S3 and  S4, the CMC of PNVP-b-dendr(Phe-Lys) 2 and PNVP-b-dendr(Phe-Lys) 3 were 0.071 and 0.056 mg/mL, respectively, which depends on the hydrophobicity of the dendritic block. When the hydrophobicity of the dendritic block increased, the CMC values of the LDBCs decreased dramatically. It should be noted that the first-generation LDBC, PNVP-b-dendr(Phe-Lys) does not form a valid envelope for Nile red due to its fewer hydrophobic Ph and Fmoc groups.
The micellar diameters were determined using particle sizing measurements. The result showed that PNVP-b-dendr(Phe-Lys) 2 and PNVP-b-dendr(Phe-Lys) 3 formed micelles with sizes of about 195 nm and 224 nm, and a slight increase in the diameter of the micelle after Nile red was loaded (Figure 3). This may arise from the micelles swelling after the inclusion of hydrophobic molecules. The formation of the spherical structures of LDBC micelles can be observed by TEM and the sizes were in agreement with particle sizing measurements ( Figure 4). As expected, both particle sizing measurements and TEM show that PNVP-b-dendr(Phe-Lys) 3 had a larger micelle diameter compared with PNVP-b-dendr(Phe-Lys) 2 . hydrophobic inner cores and hydrophilic outer shells. The CMC of the LDBCs was determined by encapsulating Nile red as a hydrophobic fluorescent probe. As shown in Figures S3 and S4, the CMC of PNVP-b-dendr(Phe-Lys)2 and PNVP-b-dendr(Phe-Lys)3 were 0.071 and 0.056 mg/mL, respectively, which depends on the hydrophobicity of the dendritic block. When the hydrophobicity of the dendritic block increased, the CMC values of the LDBCs decreased dramatically. It should be noted that the first-generation LDBC, PNVP-b-dendr(Phe-Lys) does not form a valid envelope for Nile red due to its fewer hydrophobic Ph and Fmoc groups. The micellar diameters were determined using particle sizing measurements. The result showed that PNVP-b-dendr(Phe-Lys)2 and PNVP-b-dendr(Phe-Lys)3 formed micelles with sizes of about 195 nm and 224 nm, and a slight increase in the diameter of the micelle after Nile red was loaded ( Figure  3). This may arise from the micelles swelling after the inclusion of hydrophobic molecules. The formation of the spherical structures of LDBC micelles can be observed by TEM and the sizes were in agreement with particle sizing measurements ( Figure 4). As expected, both particle sizing measurements and TEM show that PNVP-b-dendr(Phe-Lys)3 had a larger micelle diameter compared with PNVP-b-dendr(Phe-Lys)2.   hydrophobic inner cores and hydrophilic outer shells. The CMC of the LDBCs was determined by encapsulating Nile red as a hydrophobic fluorescent probe. As shown in Figures S3 and S4, the CMC of PNVP-b-dendr(Phe-Lys)2 and PNVP-b-dendr(Phe-Lys)3 were 0.071 and 0.056 mg/mL, respectively, which depends on the hydrophobicity of the dendritic block. When the hydrophobicity of the dendritic block increased, the CMC values of the LDBCs decreased dramatically. It should be noted that the first-generation LDBC, PNVP-b-dendr(Phe-Lys) does not form a valid envelope for Nile red due to its fewer hydrophobic Ph and Fmoc groups. The micellar diameters were determined using particle sizing measurements. The result showed that PNVP-b-dendr(Phe-Lys)2 and PNVP-b-dendr(Phe-Lys)3 formed micelles with sizes of about 195 nm and 224 nm, and a slight increase in the diameter of the micelle after Nile red was loaded ( Figure  3). This may arise from the micelles swelling after the inclusion of hydrophobic molecules. The formation of the spherical structures of LDBC micelles can be observed by TEM and the sizes were in agreement with particle sizing measurements ( Figure 4). As expected, both particle sizing measurements and TEM show that PNVP-b-dendr(Phe-Lys)3 had a larger micelle diameter compared with PNVP-b-dendr(Phe-Lys)2.

Enzyme-Triggered Disassembly of the LDBC Micelles
The dendritic blocks of the LDBCs consist of dipeptide Phe-Lys which renders them a suitable substrate for trypsin. It is reasonable to expect that the disassembly of LBDC micelles due to the enzymatic degradation of the dendritic blocks. To test this, an 80 U/mL solution of trypsin was added to the Nile red-encapsulated micelles of PNVP-b-dendr(Phe-Lys) 2 and PNVP-b-dendr(Phe-Lys) 3 . It was observed that time-dependence decreased in fluorescence intensity, indicating that the Nile red molecules were released from the hydrophobic cores of the micelles as they disassembled upon trypsin treatment ( Figure 5). Nile red micelles without enzyme were used as a control under the same test conditions, where we found only a small amount of Nile red was released. In addition, the disassembly of Nile red was depended on the generation of dendrons in LDBCs. The higher the generation, the slower the disassembly rate (Figure 5c). When the length of hydrophilic linear chain PNVP was the same, the third-generation LDBC with more the number of hydrophobic Phe in its dendron and peripheral Fmoc groups had a slower enzymatic hydrolysis rate than the second-generation LDBC. Because the enzyme is not able to penetrate the PNVP shell into the micellar core to degrade the dendritic block, conversely, enzymatic degradation occurs at a "monomeric" form of amphiphilic LDBC, which is in equilibrium with the micelle aggregate [11] (Figure 6).

Enzyme-Triggered Disassembly of the LDBC Micelles
The dendritic blocks of the LDBCs consist of dipeptide Phe-Lys which renders them a suitable substrate for trypsin. It is reasonable to expect that the disassembly of LBDC micelles due to the enzymatic degradation of the dendritic blocks. To test this, an 80 U/mL solution of trypsin was added to the Nile red-encapsulated micelles of PNVP-b-dendr(Phe-Lys)2 and PNVP-b-dendr(Phe-Lys)3. It was observed that time-dependence decreased in fluorescence intensity, indicating that the Nile red molecules were released from the hydrophobic cores of the micelles as they disassembled upon trypsin treatment ( Figure 5). Nile red micelles without enzyme were used as a control under the same test conditions, where we found only a small amount of Nile red was released. In addition, the disassembly of Nile red was depended on the generation of dendrons in LDBCs. The higher the generation, the slower the disassembly rate (Figure 5c). When the length of hydrophilic linear chain PNVP was the same, the third-generation LDBC with more the number of hydrophobic Phe in its dendron and peripheral Fmoc groups had a slower enzymatic hydrolysis rate than the secondgeneration LDBC. Because the enzyme is not able to penetrate the PNVP shell into the micellar core to degrade the dendritic block, conversely, enzymatic degradation occurs at a "monomeric" form of amphiphilic LDBC, which is in equilibrium with the micelle aggregate [11] (Figure 6).  Particle sizing measurement was used to confirm the enzyme triggered disassembly of the LDBC micelles. Figure 3 shows that the particle size of copolymers micelles loaded with Nile red

Enzyme-Triggered Disassembly of the LDBC Micelles
The dendritic blocks of the LDBCs consist of dipeptide Phe-Lys which renders them a suitable substrate for trypsin. It is reasonable to expect that the disassembly of LBDC micelles due to the enzymatic degradation of the dendritic blocks. To test this, an 80 U/mL solution of trypsin was added to the Nile red-encapsulated micelles of PNVP-b-dendr(Phe-Lys)2 and PNVP-b-dendr(Phe-Lys)3. It was observed that time-dependence decreased in fluorescence intensity, indicating that the Nile red molecules were released from the hydrophobic cores of the micelles as they disassembled upon trypsin treatment ( Figure 5). Nile red micelles without enzyme were used as a control under the same test conditions, where we found only a small amount of Nile red was released. In addition, the disassembly of Nile red was depended on the generation of dendrons in LDBCs. The higher the generation, the slower the disassembly rate (Figure 5c). When the length of hydrophilic linear chain PNVP was the same, the third-generation LDBC with more the number of hydrophobic Phe in its dendron and peripheral Fmoc groups had a slower enzymatic hydrolysis rate than the secondgeneration LDBC. Because the enzyme is not able to penetrate the PNVP shell into the micellar core to degrade the dendritic block, conversely, enzymatic degradation occurs at a "monomeric" form of amphiphilic LDBC, which is in equilibrium with the micelle aggregate [11] (Figure 6).  Particle sizing measurement was used to confirm the enzyme triggered disassembly of the LDBC micelles. Figure 3 shows that the particle size of copolymers micelles loaded with Nile red Particle sizing measurement was used to confirm the enzyme triggered disassembly of the LDBC micelles. Figure 3 shows that the particle size of copolymers micelles loaded with Nile red dramatically increased, from below 270 nm to more than 4000 nm, upon the addition of an 80 U/mL solution of trypsin for 48 h. This is in good agreement with the enzyme responsive feature observed with the naked eye (Figure 7). The encapsulation of Nile red by micelles made the micellar solution appear pink. After adding trypsin for 48 h, the color of the solution gradually became lighter and generated precipitates. The producing precipitate of PNVP-b-dendr(Phe-Lys) 3 after the incubation 48 h with trypsin was isolated through centrifugation and subsequently analyzed by 1 H NMR spectroscopy. Compared with 1 H NMR spectrum of PNVP-b-dendr(Phe-Lys) 3 , the resonance signals of the protons of PNVP in 1 H NMR spectrum of the precipitate almost disappeared, which demonstrated the precipitate consists mainly of (Fmoc)Lys(Fmoc)-OH and Phe-OH ( Figure S5). In order to further examine if the disassembly occurs only due to the enzymatic hydrolysis of dendritic Phe-Lys, a solution of the same trypsin concentration was added to PNVP-NH 2 , the linear chain of the LDBCs, where we do not observe any precipitates. It suggests the enzyme-specific disassembly of the LDBC micelles. The enzyme-induced disassembly of the LDBC micelles was also supported by TEM studied. As shown in Figure 4c,f, we observed unstructured and large aggregates instead of spherical structures in the aqueous solution of Nile red-loaded micelles after treatment with trypsin.
Polymers 2019, 11, x FOR PEER REVIEW 10 of 14 dramatically increased, from below 270 nm to more than 4000 nm, upon the addition of an 80 U/mL solution of trypsin for 48 h. This is in good agreement with the enzyme responsive feature observed with the naked eye ( Figure 7). The encapsulation of Nile red by micelles made the micellar solution appear pink. After adding trypsin for 48 h, the color of the solution gradually became lighter and generated precipitates. The producing precipitate of PNVP-b-dendr(Phe-Lys)3 after the incubation 48 h with trypsin was isolated through centrifugation and subsequently analyzed by 1 H NMR spectroscopy. Compared with 1 H NMR spectrum of PNVP-b-dendr(Phe-Lys)3, the resonance signals of the protons of PNVP in 1 H NMR spectrum of the precipitate almost disappeared, which demonstrated the precipitate consists mainly of (Fmoc)Lys(Fmoc)-OH and Phe-OH ( Figure S5). In order to further examine if the disassembly occurs only due to the enzymatic hydrolysis of dendritic Phe-Lys, a solution of the same trypsin concentration was added to PNVP-NH2, the linear chain of the LDBCs, where we do not observe any precipitates. It suggests the enzyme-specific disassembly of the LDBC micelles. The enzyme-induced disassembly of the LDBC micelles was also supported by TEM studied. As shown in Figure 4c,f, we observed unstructured and large aggregates instead of spherical structures in the aqueous solution of Nile red-loaded micelles after treatment with trypsin.

In Vitro Drug Release from DOX-Loaded Micelles
We investigated the trypsin-triggered release of doxorubicin (DOX) from the LBDCs micelles in PBS buffer (pH = 7.4) at 37 °C to simulate the physiological environment. The encapsulation efficiency (EE) of DOX in PNVP-b-dendr(Phe-Lys)2 and PNVP-b-dendr(Phe-Lys)3 micelles was 71.2% and 77.6%, respectively, when the feeding amount of DOX/ LBDCs (wt/wt) was <5%. The release of DOX in micelles was very low about 30% at 48 h in the absence of trypsin. However, DOX-loaded micelles showed significant release of DOX with about 15% release at 2 h and about 65% release at 48 h after the addition of an 80 U/mL solution of trypsin ( Figure 8). The cumulative release of DOX-loaded micelles was not very accurate because the copolymers produced insoluble precipitate upon enzymatic activation and was mixed with DOX. The mixture was stirred in PBS buffer (pH = 5) to fully dissolve the DOX and then the fluorescence intensity of the resulting solution was measured. We found that the cumulative release of the drug reached more than 80% at 48 h upon trypsin treatment. The reason for the remarkably increase of DOX release may arise from the enzymeinduced disassembly of DOX-loaded micelles at physiological pH condition.

In Vitro Drug Release from DOX-Loaded Micelles
We investigated the trypsin-triggered release of doxorubicin (DOX) from the LBDCs micelles in PBS buffer (pH = 7.4) at 37 • C to simulate the physiological environment. The encapsulation efficiency (EE) of DOX in PNVP-b-dendr(Phe-Lys) 2 and PNVP-b-dendr(Phe-Lys) 3 micelles was 71.2% and 77.6%, respectively, when the feeding amount of DOX/ LBDCs (wt/wt) was <5%. The release of DOX in micelles was very low about 30% at 48 h in the absence of trypsin. However, DOX-loaded micelles showed significant release of DOX with about 15% release at 2 h and about 65% release at 48 h after the addition of an 80 U/mL solution of trypsin ( Figure 8). The cumulative release of DOX-loaded micelles was not very accurate because the copolymers produced insoluble precipitate upon enzymatic activation and was mixed with DOX. The mixture was stirred in PBS buffer (pH = 5) to fully dissolve the DOX and then the fluorescence intensity of the resulting solution was measured. We found that the cumulative release of the drug reached more than 80% at 48 h upon trypsin treatment. The reason for the remarkably increase of DOX release may arise from the enzyme-induced disassembly of DOX-loaded micelles at physiological pH condition.

In Vitro Cytotoxicity of the LDBCs
To apply the LDBCs in drug delivery field, the biocompatibility of PNVP-b-dendr(Phe-Lys)2 and PNVP-b-dendr(Phe-Lys)3 was investigated by cytotoxicity assessments. The LDBCs at varying concentrations were incubated with human lung epithelial cells (BEAS-2B) or human liver cancer cells (SMMC-7721) for 48 hours. The viability of BEAS-2B and SMMC-7721 cells after incubation with the copolymers was determined via MTS assay. As shown in Figure 9, no apparent toxicity effect was found on BEAS-2B and SMMC-7721 by the LDBCs when the concentration of the LDBCs increased from 0.05 to 0.8 mg/mL. These results demonstrate that PNVP-b-dendr(Phe-Lys)2 and PNVP-bdendr(Phe-Lys)3 have a good biocompatibility within limit concentration and are suitable for drug delivery.

Conclusions
In summary, we reported the design and synthesis of a novel class of enzyme responsive amphiphilic LDBCs (PNVP-b-dendr(Phe-Lys)n, n = 1-3) consisted of a hydrophilic linear PNVP and hydrophobic dendritic Phe-Lys dipeptides. The amphiphiles were synthesized through a combination of RAFT/MADIX polymerization of N-vinylpyrrolidone and by stepwise condensation approach of peptide chemistry. Specifically, we have demonstrated the second and the third generation LDBCs self-assemble to form spherical micelles under aqueous condition. The micelles were able to disassemble and release the encapsulated anticancer drug doxorubicin or hydrophobic dye Nile red upon trypsin treatment. Unlike classical amino acid-based LDBCs which contain just

In Vitro Cytotoxicity of the LDBCs
To apply the LDBCs in drug delivery field, the biocompatibility of PNVP-b-dendr(Phe-Lys) 2 and PNVP-b-dendr(Phe-Lys) 3 was investigated by cytotoxicity assessments. The LDBCs at varying concentrations were incubated with human lung epithelial cells (BEAS-2B) or human liver cancer cells (SMMC-7721) for 48 hours. The viability of BEAS-2B and SMMC-7721 cells after incubation with the copolymers was determined via MTS assay. As shown in Figure 9, no apparent toxicity effect was found on BEAS-2B and SMMC-7721 by the LDBCs when the concentration of the LDBCs increased from 0.05 to 0.8 mg/mL. These results demonstrate that PNVP-b-dendr(Phe-Lys) 2 and PNVP-b-dendr(Phe-Lys) 3 have a good biocompatibility within limit concentration and are suitable for drug delivery.

In Vitro Cytotoxicity of the LDBCs
To apply the LDBCs in drug delivery field, the biocompatibility of PNVP-b-dendr(Phe-Lys)2 and PNVP-b-dendr(Phe-Lys)3 was investigated by cytotoxicity assessments. The LDBCs at varying concentrations were incubated with human lung epithelial cells (BEAS-2B) or human liver cancer cells (SMMC-7721) for 48 hours. The viability of BEAS-2B and SMMC-7721 cells after incubation with the copolymers was determined via MTS assay. As shown in Figure 9, no apparent toxicity effect was found on BEAS-2B and SMMC-7721 by the LDBCs when the concentration of the LDBCs increased from 0.05 to 0.8 mg/mL. These results demonstrate that PNVP-b-dendr(Phe-Lys)2 and PNVP-bdendr(Phe-Lys)3 have a good biocompatibility within limit concentration and are suitable for drug delivery.

Conclusions
In summary, we reported the design and synthesis of a novel class of enzyme responsive amphiphilic LDBCs (PNVP-b-dendr(Phe-Lys)n, n = 1-3) consisted of a hydrophilic linear PNVP and hydrophobic dendritic Phe-Lys dipeptides. The amphiphiles were synthesized through a combination of RAFT/MADIX polymerization of N-vinylpyrrolidone and by stepwise condensation approach of peptide chemistry. Specifically, we have demonstrated the second and the third generation LDBCs self-assemble to form spherical micelles under aqueous condition. The micelles were able to disassemble and release the encapsulated anticancer drug doxorubicin or hydrophobic dye Nile red upon trypsin treatment. Unlike classical amino acid-based LDBCs which contain just

Conclusions
In summary, we reported the design and synthesis of a novel class of enzyme responsive amphiphilic LDBCs (PNVP-b-dendr(Phe-Lys) n , n = 1-3) consisted of a hydrophilic linear PNVP and hydrophobic dendritic Phe-Lys dipeptides. The amphiphiles were synthesized through a combination of RAFT/MADIX polymerization of N-vinylpyrrolidone and by stepwise condensation approach of peptide chemistry. Specifically, we have demonstrated the second and the third generation LDBCs self-assemble to form spherical micelles under aqueous condition. The micelles were able to disassemble and release the encapsulated anticancer drug doxorubicin or hydrophobic dye Nile red upon trypsin treatment. Unlike classical amino acid-based LDBCs which contain just one kind of amino acid such as L-lysine, the dipeptide Phe-Lys are used as branch units, which could not only increase hydrophobicity which is essential for the formation of stable micelles but also be cleaved by protease trypsin. The disassembly of the micelles and release rates of the drug or dye can be adjusted by changing the generation of dendrons in PNVP-b-dendr(Phe-Lys) n . The higher the generation, the more stable the micelles and the slower the disassembly rate. MTS assays indicate that the LDBCs are well tolerated by different cell lines. This design and simple synthetic method will hopefully stimulate the manufacture of more enzyme-responsive polymer micelles in the future field of drug delivery.