Kinetics of self-immolative degradation in a linear polymeric system: Demonstrating the effect of chain length

Article abstract of DOI:10.1021/ma4009753

We describe here a study demonstrating that the degradation time of self-immolative linear polymers is dependent on chain length. These materials are unique relative to most degradable polymers, in that they undergo end-to-end depolymerization in response to the cleavage of an end-cap. Although one of their cited attributes is a dependence of their degradation time on chain length, no conclusive study has been conducted to demonstrate and study this effect. In this work, using a linear self-immolative polymer backbone derived from alternating 4-hydroxybenzyl alcohol and N,N′-dimethylethylenediamine based spacers, we show that there is a proportional relationship between chain length and depolymerization time. This is first accomplished using a series of oligomers synthesized using a convergent, iterative route and then applied to the polydisperse case on a set of polymers displaying varying molecular weights. We also report the first development and validation of a self-immolative degradation model relating monomer kinetics to polymer degradation and show its application in explaining oligomeric and polymeric degradation profiles.

Full text of DOI:10.1021/ma4009753

Article
pubs.acs.org/Macromolecules
Kinetics of Self-Immolative Degradation in a Linear Polymeric
System: Demonstrating the Effect of Chain Length
Ryan A. McBride† and Elizabeth R. Gillies*,†,‡
‡
†Departments of Chemical and Biochemical Engineering and Chemistry, The University of Western Ontario, London, ON N6A
3K7, Canada
S
* Supporting Information
ABSTRACT: We describe here a study demonstrating that the
degradation time of self-immolative linear polymers is dependent on
chain length. These materials are unique relative to most degradable
polymers, in that they undergo end-to-end depolymerization in
response to the cleavage of an end-cap. Although one of their cited
attributes is a dependence of their degradation time on chain length,
no conclusive study has been conducted to demonstrate and study this
effect. In this work, using a linear self-immolative polymer backbone
derived from alternating 4-hydroxybenzyl alcohol and N,N′-dimethy-
lethylenediamine based spacers, we show that there is a proportional
relationship between chain length and depolymerization time. This is first accomplished using a series of oligomers synthesized
using a convergent, iterative route and then applied to the polydisperse case on a set of polymers displaying varying molecular
weights. We also report the first development and validation of a self-immolative degradation model relating monomer kinetics to
polymer degradation and show its application in explaining oligomeric and polymeric degradation profiles.
degradation process. Self-immolative polymers undergo end-to-
end depolymerization through a cascade of intramolecular
reactions upon the removal of a stimuli-responsive, stabilizing
end-cap (Figure 1a).18−20 This particular mechanism of
degradation possesses three main attributes: (i) it permits the
amplification of a given stimulus, in that only one triggering
event is required to achieve complete degradation of the
polymer backbone; (ii) the stimuli-responsive degradation of a
given polymer can be readily adjusted and tuned through the
incorporation of different end-capping agents; and (iii) the
degradation kinetics of these polymers are dictated and
controlled by the chemical composition of the polymer
backbone and the polymer length in a highly predictable
manner.
Following the seminal reports on the design and synthesis of
self-immolative dendrimers,21−23 a number of different self-
immolative polymer backbones and architectures have been
designed for signal amplification,24−27 drug delivery,28−31
microfluidics,32 protein labeling,33 self-healing,34 and shape-
memory applications.35 Our group, along with several others,
have focused predominantly on the design and synthesis of
linear self-immolative polymers with a special consideration
toward fine-tuning the degradation rate.26,33−41 Despite being
no longer monodisperse, linear self-immolative polymers
provide a high-degree of amplification in a limited number of
INTRODUCTION
■
Recent advances in the field of polymer science have
contributed to a paradigmatic shift in the use of biodegradable
polymers as environmentally friendly substitutes for traditional
commodity plastics and in biomedical applications including
drug delivery and tissue engineering. While there exists a large
variety of different biodegradable polymer backbones, the
predominant majority of such systems are polyester-based.
Despite the proven utility of popular synthetic polyesters such
as poly(caprolactone), poly(lactic acid), and poly(glycolic
acid), the degradation of polyesters suffers from a lack of fine
control, as it initiated by a nonspecific triggering event and
relies on random hydrolytic cleavage of ester bonds throughout
the polymer backbone.1−3 In part, the limited control over the
degradation of polyesters has been addressed through the
development of stimuli-responsive, biodegradable polymers in
which polymer degradation is triggered4 by specific environ-
mental conditions such as changes in pH5−7 or redox
potential,8−10 biologically mediated events including enzymatic
cleavage11 and changes in chemical concentration,12,13 or by an
applied external stimulus such as UV/NIR light.14−17 Such
systems are an improvement over basic polyesters in that they
circumvent the issue of nonspecific degradation. However,
these systems still require many environmentally mediated
cleavage events to ensure complete degradation, thereby
restricting their usage to applications in which triggering events
are abundant.
Self-immolative polymers have been developed over the past
decade as an alternative to traditional biodegradable polymers
and offer new levels of control and amplification to the
Received: May 10, 2013
Revised: June 5, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ma4009753 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
Figure 1. Illustration of linear self-immolative degradation. (a) Generic overview depicting the end-to-end depolymerization of the polymer
backbone following the removal of the stabilizing end-cap. (b) Previously reported linear self-immolative polymer derived from alternating
cyclization and 1,6-elimination spacers and its mechanism of degradation.
synthetic steps while bypassing any restrictions associated with
steric constraints.
degradation model relating monomer kinetics to polymer
degradation and show its application in explaining oligomeric
and polymeric degradation profiles. Collectively, this work
offers the first quantitative evidence supporting the mixed zero-
and first-order degradation kinetics of linear self-immolative
polymers and proves the utility of chain length as an alternate
means to tune the degradation time in linear self-immolative
polymers.
The first reported linear self-immolative polymer was derived
entirely from monomers that underwent 1,6-elimination
reactions to degrade the polymer into fluorescent monomer
units.36 Our group subsequently developed linear polymer
backbones that were comprised of alternating cyclization and
1,6-elimination spacers or purely cyclization spacers and
showed that the degradation kinetics of such polymers could
be significantly slowed through the incorporation of cyclization
spacers.37,38 More recently, we have shown that the degradation
rate of such polymers can be effectively tuned by exploiting the
structure−property relationships of the monomer cyclization
spacers.39,41 A similar approach has also been utilized by
Phillips and co-workers for linear self-immolative oligomers
derived from 1,6-elimination spacers.42,43 While this approach
is a proven means to induce large changes in the
depolymerization rate, it should be possible to effectively
tune the time for complete degradation by simply changing the
length of the polymer. In particular, as the depolymerization of
linear self-immolative polymers involves a cascade of reactions
across a defined architecture, the degradation time of these
polymers should be related to the chain length. In the absence
of more complicated effects arising from alternative reaction
pathways or electronic communication between adjacent
monomer units, the degradation time of linear self-immolative
oligomers and polymers should be proportional to chain length.
While several studies have been performed to investigate the
effect of chain length on the overall degradation time of short
self-immolative oligomers, with varying results,44−47 the length
dependence for longer oligomers and polymers has not yet
been demonstrated, and no comprehensive model has been
developed to describe their behavior.
EXPERIMENTAL SECTION
■
General Procedures and Materials. All reagents were purchased
from commercial suppliers and used without further purification unless
otherwise noted. Anhydrous toluene was obtained from a solvent
purification system using aluminum oxide columns. Dichloromethane
(CH2Cl2), pyridine, triethylamine (NEt3), and N,N-diisopropylethyl-
amine (DIPEA) were distilled from CaH2 immediately prior to use.
Unless otherwise stated, all reactions were performed under an Ar
atmosphere using flame-dried glassware. Column chromatography was
performed using silica gel (0.040−0.063 mm particle size, 230−430
mesh). Thin layer chromatography (TLC) was carried out using EMD
silica gel 60 F254 plates (20 cm × 20 cm, 250 μm). Deionized water
was purified using a Millipore purification system. Dialyses were
performed using Spectra/Por regenerated cellulose membranes with
either a 6000−8000 or 25 000 g/mol molecular weight cutoff
(MWCO).
1H NMR spectra were obtained at 400 or 600 MHz using a Varian
INOVA spectrometer, while 13C NMR spectra were obtained at 100
MHz using a Varian INOVA spectrometer. Chemical shifts are
reported in ppm versus tetramethylsilane and referenced to residual
solvent signals of CDCl3 (δ 7.26, 77.2), (CD3)2CO (δ 2.05), or D2O
(δ 4.79). Coupling constants are expressed in hertz. High-resolution
mass spectrometry (HRMS) was performed on either a Finnigan MAT
8400 or a PE-Sciex API 365 mass spectrometer using electron impact
(EI) or electrospray (ESI) ionization, respectively. Fourier transform
infrared spectra were obtained using a Bruker Tensor 27 from CH2Cl2
films on KBr plates. Size exclusion chromatography (SEC) was carried
out at a flow rate of 1 mL/min in N,N-dimethylformamide (DMF)
with 10 mM LiBr and 1% (v/v) NEt3 at 85 °C using a Waters 2695
separations module equipped with a Waters 2414 differential
refractometer and two PLgel 5 μm mixed-D (300 mm × 7.5 mm)
columns from Polymer Laboratories connected in series. SEC
calibrations were performed using poly(methyl methacrylate)
(PMMA) standards. Reverse phase high-pressure liquid chromatog-
raphy (RP-HPLC) was performed at a flow rate of 1 mL/min on a
Waters 2695 separations module equipped with a Waters 2998
photodiode array detector at a wavelength of 225 nm and a Nova-Pak
Using our previously reported linear polycarbamate back-
bone derived from alternating 1,6-elimination and cyclization
spacers (Figure 1b),37 we herein report the first conclusive
study demonstrating the effect of chain length on the kinetics of
linear self-immolative degradation. This is accomplished using a
set of monodisperse oligomers synthesized through a new
convergent iterative route and a series of polymers optimized to
display varying molecular weights. We also report the
development and validation of a new linear self-immolative
B
dx.doi.org/10.1021/ma4009753 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
C-18 analytical column (3.9 mm × 150 mm, 4 μm). The HPLC
gradient was linear from 60/40 MeCN/H2O to 100/0 MeCN/H2O
over 15 min. Solvent mixtures for all chromatographic analyses
contained 0.1% TFA.
h) as determined by TLC. The reaction mixture was then diluted with
CH2Cl2 and washed with 1 M HCl followed by saturated NaHCO3.
The organic layer was dried over MgSO4, filtered, and concentrated in
vacuo. The resulting oil was purified by silica gel flash chromatography
(1:4 EtOAc:CH2Cl2, then 2:3 EtOAc:CH2Cl2) to afford 2c as a white
Synthesis. Synthesis of Compound 1a. Unactivated monomer
1b37 (1.34 g, 4.0 mmol, 1.0 equiv) was dissolved in 20 mL of
anhydrous CH2Cl2. Acetyl chloride (0.42 mL, 5.9 mmol, 1.5 equiv),
pyridine (1.0 mL, 12.0 mmol, 3 equiv), and 4-(dimethylamino)-
pyridine (DMAP) (0.05 g, 0.4 mmol, 0.1 equiv) were successively
added to the reaction flask, and the solution was stirred at room
temperature until completion (∼1.25 h) as determined by TLC. The
reaction mixture was then diluted with CH2Cl2 and washed with 1 M
HCl. The organic layer was dried over MgSO4, filtered, and
concentrated in vacuo. The resulting oil was purified by silica gel
flash chromatography (1:19 EtOAc:CH2Cl2, gradient to 3:7
1
solid (0.51 g, 79%). H NMR (CDCl3, 400 MHz): δ 8.23 (d, J = 9.0,
2H), 7.46−7.27 (m, 6H), 7.14−7.01 (m, 4H), 5.24 (s, 2H), 5.13−5.03
(m, 2H), 3.66−3.35 (m, 8H), 3.15−2.82 (m, 8H), 1.50−1.37 (m, 9H,
rotamers). FT-IR (νmax/cm−1): 2972, 2934, 1767, 1720, 1697, 1614,
1593, 1526. HRMS: calcd [M + Na]+ (C37H45N5O13Na): 790.2912.
Found (ESI) 790.2898. RP-HPLC: tR = 7.2 min (purity 99%).
Synthesis of Compounds 3a, 3b, 3c, and 4a. Acetylated tetramer
3a and octamer 4a were prepared using similar conditions to those
described for compound 2a. Deprotected tetramer 3b and 4-
nitrophenyl activated tetramer 3c were prepared using similar
conditions to 2b and 2c, respectively. Please refer to the Supporting
Information for procedural details and complete characterization data.
Synthesis of Polymer 6. Activated monomer 1c37 (0.28 g, 0.55
mmol, 1 equiv) was dissolved in 3 mL of 1:1 TFA:CH2Cl2 and stirred
at room temperature for 2 h. The solvent was then removed under a
stream of nitrogen in the fume hood, prior to subjecting the reaction
mixture three times to a repeat cycle of dilution with CH2Cl2 followed
by concentration under reduced pressure to remove residual TFA and
provide the deprotected monomer 5.37 End-cap 1c37 (0.014 g, 0.027
mmol, 0.05 equiv) was added, and the resulting mixture was dissolved
in 3 mL of anhydrous toluene. DIPEA (0.48 mL, 2.75 mmol, 5.0
equiv) and DMAP (0.016 g, 0.13 mmol, 0.24 equiv) were sequentially
added, and the solution was stirred at room temperature for 6 h. The
solvent was then evaporated under reduced pressure, and the crude
polymer was dissolved in 2 mL of DMF and dialyzed against DMF for
24 h (200 mL, 1 solvent change) using a regenerated cellulose
membrane (6000−8000 g/mol MWCO). The contents of the dialysis
membrane were then concentrated in vacuo and lyophilized to afford
polymer 2.6 (0.063 g, 42%). 1H NMR indicated a degree of
polymerization of ∼25 by integrating the benzylic peak against the
1
EtOAc:CH2Cl2) to afford 1a as a pale yellow oil (1.31 g, 87%). H
NMR (CDCl3, 400 MHz): δ 7.31 (d, J = 7.8, 2H), 7.11−7.03 (m, 2H),
5.04 (s, 2H), 3.61−3.49 and 3.48−3.37 (2 m, 4H total, rotamers), 3.09
and 3.00 (2 s, 3H total, rotamers), 2.93−2.82 (m, 3H, rotamers), 2.05
(s, 3H), 1.48−1.39 (m, 9H, rotamers). 13C NMR (CDCl3, 150 MHz):
δ 170.7, (155.9, 155.7 and 155.5, rotamers), (154.7, 154.5, and 154.4,
rotamers), (151.4 and 151.2, rotamers), (133.0 and 132.9, rotamers),
129.4, (122.0 and 121.8, rotamers), (79.8, 79.7, 79.6, and 79.4,
rotamers), 69.7, (47.3, 47.1, 46.9, 46.7, 46.5, 46.4, and 45.6,
methylenes, rotamers), (35.4, 35.2, 35.1, 34.7, and 34.5, methyls,
rotamers), 28.4, 21.0. FT-IR (νmax/cm−1): 2976, 2934, 1724, 1693,
1515. HRMS: calcd [M]+ (C19H28N2O6): 380.1947. Found (EI)
380.1948. RP-HPLC: tR = 3.8 min (purity >99%).
Synthesis of Compound 2a. Acetylated monomer 1a (1.12 g, 2.9
mmol, 1.2 equiv) was dissolved in 6 mL of 1:1 TFA:CH2Cl2 and
stirred at room temperature for 2 h. The reaction mixture was then
diluted with CH2Cl2, and the solvent was removed under reduced
pressure. This dilution/evaporation cycle was repeated an additional
three times to remove residual TFA, yielding the deprotected-
acetylated monomer 1d. 4-Nitrophenyl-activated activated monomer
1c37 (1.25 g, 2.5 mmol, 1.0 equiv) was then added, and the resulting
mixture was dissolved in anhydrous toluene and cooled to 0 °C.
DIPEA (1.3 mL, 7.4 mmol, 3.0 equiv) and DMAP (35 mg, 0.29 mmol,
0.12 equiv) were sequentially added, and the solution was stirred at 0
°C for 8 h. The reaction was allowed to warm to room temperature
and stirred an additional 8 h. The reaction mixture was then diluted
with CH2Cl2 and washed with 1 M HCl followed by 1 M Na2CO3.
The organic layer was dried over MgSO4, filtered, and concentrated in
vacuo. The resulting oil was then purified by silica gel flash
chromatography (1:4 EtOAc:CH2Cl2, gradient to 3:2 EtOAc:CH2Cl2)
1
Boc end-cap. H NMR (CDCl3, 600 MHz): δ 7.46−7.26 (m, 51H),
7.18−6.95 (m, 44H), 5.18−4.97 (m, 50H), 3.70−3.37 (m, 103H),
3.25−2.80 (m, 158H), 1.50−1.41 (m, 9H, rotamers). SEC: Mn = 5250
g/mol, Mw = 7730 g/mol, PDI = 1.47 (PMMA standards).
Synthesis of Polymer 7. Activated monomer 1c37 (1.25 g, 2.49
mmol, 1 equiv) was dissolved in 6 mL of 1:1 TFA:CH2Cl2 and stirred
at room temperature for 2 h. The solvent was then removed under a
stream of nitrogen in the fume hood prior to subjecting the reaction
mixture three times to a repeat cycle of dilution with CH2Cl2 followed
by concentration under reduced pressure to remove residual TFA and
provide the deprotected monomer. End-cap 1c37 (0.013 g, 0.025
mmol, 0.01 equiv) was added, and the resulting mixture was dissolved
in 12 mL of anhydrous toluene and cooled to 0 °C. NEt3 (1.73 mL,
12.43 mmol, 5 equiv) and DMAP (0.066 g, 0.54 mmol, 0.22 equiv)
were sequentially added, and the solution was stirred at 0 °C for 24 h.
The reaction mixture was then diluted with CH2Cl2 and washed with 1
M HCl followed by 1 M Na2CO3. The organic layer was dried over
MgSO4, filtered, and concentrated in vacuo to provide the crude
polymer (0.60 g, 92%). The crude polymer was then dissolved in 5 mL
of DMF and dialyzed against DMF for 24 h (500 mL, 1 solvent
change) using a regenerated cellulose membrane (25 000 g/mol
MWCO). The contents of the dialysis membrane were then
concentrated in vacuo and lyophilized to afford polymer 7 (0.25 g,
37%). 1H NMR indicated a degree of polymerization of ∼101 by
integrating the benzylic peak against the Boc end-cap. 1H NMR
(CDCl3, 600 MHz): δ 7.39−7.29 (m, 181H), 7.12−6.99 (m, 178H),
5.17−5.02 (m, 203H), 3.70−3.36 (m, 409H), 3.21−2.79 (m, DMF,
700H), 1.49−1.42 (m, 9H, rotamers). SEC: Mn = 13 600 g/mol, Mw =
21 500 g/mol kDa, PDI = 1.58 (PMMA standards).
1
to afford 2a as a colorless oil (1.34 g, 84%). H NMR (CDCl3, 400
MHz): δ 7.38−7.27 (m, 4H), 7.11−7.00 (m, 4H), 5.13−5.01 (m, 4H),
3.64−3.36 (m, 8H), 3.12−2.82 (m, 12H), 2.06 (s, 3H), 1.52−1.37 (m,
9H, rotamers). FT-IR (νmax/cm−1): 2932, 1722, 1699, 1512. HRMS:
calcd [M + Na]+ (C32H44N4O10Na): 667.2955. Found (ESI):
667.2929. RP-HPLC: tR = 5.1 min (purity >99%). SEC: Mn = 875
g/mol, PDI = 1.03 (PMMA standards).
Synthesis of Compound 2b. Acetylated dimer 2a (0.64 g, 1.0
mmol, 1.0 equiv) and LiOH·H2O (59 mg, 1.4 mmol, 1.4 equiv) were
dissolved in 15 mL of 3:2 THF:H2O and stirred at room temperature
for 16 h. Upon completion of the reaction, the solvent mixture was
poured over 1 M HCl and the product was extracted with CH2Cl2.
The combined organic layers were dried over MgSO4, filtered, and
concentrated in vacuo to yield compound 2b (0.55 g, 92%) as a white
solid. 1H NMR (CDCl3, 400 MHz): δ 7.38−7.17 (m, 4H), 7.10−6.89
(m, 4H), 5.13−5.03 (m, 2H, rotamers), 4.58 (d, J = 5.4, 2H,
rotamers), 3.67−3.37 (m, 8H), 3.16−2.78 (m, 12H), 1.48−1.37 (m,
9H, rotamers). FT-IR (νmax/cm−1): 3468, 2928, 2872, 2856, 1718,
1701, 1510. HRMS: calcd [M + Na]+ (C30H42N4O9Na): 625.2850.
Found (ESI) 625.2858. RP-HPLC: tR = 3.4 min (purity >99%).
Synthesis of Compound 2c. Unactivated dimer 2b (0.50 g, 0.84
mmol, 1.0 equiv) and pyridine (20 μL, 2.5 mmol, 3.0 equiv) were
dissolved in 6 mL of anhydrous CH2Cl2. 4-Nitrophenyl chloroformate
(0.34 g, 1.7 mmol, 2.0 equiv) was added slowly to the reaction flask,
and the solution was stirred at room temperature until completion (∼3
Simulation Studies. Log-normal polymer distributions were
simulated using the Statistics Toolbox in MATLAB (R2012A).
Solutions to the set of differential equations (eqs 2, 3, and 5) were
solved numerically in MATLAB using a nonstiff ordinary differential
equation solver (ode45) using simulated log-normal weight-fraction
distributions for the initial concentrations of each polymeric species.
C
dx.doi.org/10.1021/ma4009753 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
Scheme 1. Iterative Convergent Synthesis of Monodisperse Oligomers
Degradation Studies. Degradation of Oligomers 1d−4d.
Acetylated oligomers 1a−4a were Boc-deproected using the same
procedures described above to yield compounds 1d−4d. Deprotected
oligomer (10 mg) was dissolved in CH2Cl2 and washed with a 1:1
mixture of brine:1 M citric acid. The product was then re-extracted
from the aqueous layer five times with CH2Cl2. The combined organic
layers were dried over MgSO4, filtered, and concentrated in vacuo. The
product was then taken up in 1 mL of 0.1 M phosphate buffer
(D2O):acetone-d6 (3:2) preheated to 37 °C and filtered through a
Promax PTFE membrane (0.22 μm). The filtered solution was then
incubated at 37 °C using an INOVA variable temperature controller
Following depolymerization, the pH of the solution was tested to
ensure that the sample did not fall outside of the buffer region during
the degradation process. The extent of depolymerization was
quantified by integrating the methyl peak of the N,N′-dimethylimida-
zolidinone degradation product relative to the (CHD2)(CD3)CO in
the sample. The plateau region corresponding to 100% degradation
was defined as the region in which the mean fluctuation between
successive integral values was less than or equal to 5%. Polymer
degradation data were treated by nonlinear regression and fit to the
self-immolative degradation model (eqs 2, 3, and 5) to assess the
degradation kinetics.
Statistics. Nonlinear least-squares regression was performed using
a Gauss−Newton algorithm contained within the R stats package.
Unless otherwise noted, the presented data represent the average plus
or minus the standard deviation determined through triplicate
measurement.
1
calibrated using ethylene glycol. H NMR spectra were recorded at 3
min intervals over a period of 4 h. Following depolymerization, the pH
of the solution was tested to ensure that the sample did not fall outside
of the buffer region during the degradation process. The extent of
depolymerization was quantified by integrating the methyl peak of the
N,N′-dimethylimidazolidinone degradation product relative to the
(CHD2)(CD3)CO in the sample. The plateau region corresponding to
100% degradation was defined as the region in which the mean
fluctuation between successive integral values was less than or equal to
2%. Oligomer degradation data were treated by nonlinear regression
and fit to the appropriate form of eq 8 to assess the degradation
kinetics.
Degradation of Polymers 6 and 7. Polymer (10 mg) was dissolved
in 1 mL of 1:1 TFA:CH2Cl2 and stirred at room temperature for 2 h.
The solvent was removed under a stream of nitrogen in the fume
hood, and the product was taken up in CH2Cl2 and washed with a 1:1
mixture of brine:1 M citric acid. The deprotected polymer was then re-
extracted from the aqueous layer five times with CH2Cl2. The
combined organic layers were dried over MgSO4, filtered, and
concentrated in vacuo. The product was then taken up in 1 mL of
0.1 M phosphate buffer (D2O):acetone-d6 (3:2) preheated to 37 °C
and filtered through a Promax PTFE membrane (0.22 μm). The
filtered solution was then incubated at 37 °C using an INOVA variable
temperature controller calibrated using ethylene glycol. 1H NMR
spectra were recorded at 5 min intervals over a period of 8 h.
RESULTS AND DISCUSSION
■
Choice of Model System. The chemical structure of our
previously reported linear self-immolative polycarbamate
derived from alternating N,N′-dimethylethylenediamine and
4-hydroxybenzyl alcohol-based spacers is shown in Figure 1b.37
Selective removal of the N-tert-butoxycarbonyl (Boc) end-cap
reveals a terminal amine that cyclizes to form N,N′-
dimethylimidazolidinone, exposing a phenol, which then
spontaneously undergoes 1,6-elimination followed by a
decarboxylation to regenerate an amine on the polymer
terminus. This cascade of alternating cyclization, 1,6-
elimination, and decarboxylation reactions is repeated along
the length of the polymer backbone until complete
depolymerization of the polymer into N,N′-dimethylimidazoli-
dinone, 4-hydroxybenzyl alcohol, and CO2 occurs. This
polymeric framework was selected as the model system for
the kinetic study as its mechanism of degradation has been
D
dx.doi.org/10.1021/ma4009753 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
previously investigated in a number of different solvent
systems,21,37 and the cyclization of N,N′-dimethylethylenedi-
amine spacers is well studied and occurs over a moderate time
frame under physiological conditions.48 Additionally, the
polymer is sufficiently stable as a trifluoroacetic acid (TFA)
salt following end-cap removal. As a result, end-cap removal can
be effectively decoupled from the degradation of the polymer
backbone, allowing the kinetics of self-immolation to be studied
independent from the end-cap removal.
different oligomer species would invalidate the assumption of
monodispersity in the subsequent model development and
validation. As shown in Figure 2a, SEC traces for 2a, 3a, and 4a
Synthesis. In order to test the effect of chain length on the
kinetics of linear self-immolative degradation, a series of
monodisperse oligomers were synthesized using a convergent
iterative scheme (Scheme 1). The use of monodisperse
oligomers was driven by the fact that chain length and
polydispersity are inherently difficult to control in the
previously reported polycondensation of the model polymer.
The use of a monodisperse sample set also eliminates any
competing effects of polydispersity on the overall degradation
kinetics, thereby allowing for a more accurate assessment of the
influence of chain length on self-immolative degradation. A
monomer and three oligomers, dimer, tetramer, and octamer,
were prepared using a modification of the route used to
synthesize the model polymer.37 As in the previously reported
monomer synthesis,37 a Boc group was selected as a protecting
group and end-cap for the amine termini of the oligomers. An
acetyl protecting group was chosen for the hydroxyl termini of
the oligomers because it can be selectively removed in the
presence of carbamates and can tolerate TFA deprotections of
the Boc group.41 As shown in Scheme 1, the previously
reported unactivated monomer (1b)37 was reacted with acetyl
chloride to provide the acetyl-protected monomer (1a). The
Boc group of 1a was then removed by treatment with 1:1
TFA:CH2Cl2 to provide monomer 1d. The monomer 1d was
then reacted with the 4-nitrophenyl carbonate-activated
monomer (1c)37 to afford the acetyl-protected dimer (2a).
From here, a convergent iterative strategy was employed in
which the acetyl-protected oligomers (2a, 3a) were treated with
LiOH·H2O in 3:2 THF:H2O to hydrolyze the acetate and
provide the unactivated alcohol species (2b, 3b). These
compounds were then activated with 4-nitrophenyl chlor-
oformate to afford the activated carbonate species (2c, 3c),
which were subsequently reacted with the Boc-deprotected
acetyl compounds (2d, 3d) in the presence of DIPEA and
DMAP to produce the acetyl-protected oligomers (3a, 4a). For
example, the tetramer 3a was obtained from the reaction of
dimer 2c with dimer 2d, while the octamer 4a was obtained
from tetramer 3c and 3d.
Figure 2. (a) Size exclusion chromatograms of compounds 2a (− · −),
3a (- - -), and 4a () prior to the removal of the Boc-protecting group
(black) and after complete degradation (gray) in 0.1 M phosphate
buffer (D2O):acetone-d6 (3:2) at 37 °C. (b) Size exclusion
chromatograms of polymers 6 (Mn = 5250 g/mol, PDI = 1.47)
(− −) and 7 (Mn = 13 600 g/mol, PDI = 1.58) () before (black)
and after degradation (gray). All chromatograms were acquired at a
sample concentration of 5 mg/mL and calibrated using PMMA
standards.
exhibited monomodal distributions and very narrow poly-
dispersities ranging from 1.02 to 1.03, as expected for the
convergent multistep oligomer syntheses. The number-average
molecular weights for 2a, 3a, and 4a were 875, 1940, and 3660
g/mol relative to PMMA standards, in comparison with the
theoretical molecular weights of 644, 1173, and 2230 g/mol,
respectively.
The oligomers were characterized by NMR spectroscopy, IR
spectroscopy, SEC, and HPLC. Notably, the absence of N,N′-
dimethylimidazolidinone that would be formed from the
competing intramolecular cyclization reaction during each
coupling step was assessed using 1H NMR spectroscopy.
Because of the presence of many carbamate rotameric centers,
standard 13C NMR analysis could not be used to assess purity
in each of the oligomer samples above the monomer
generation. Therefore, the absence of different oligomeric
species from each sample was assessed using reverse phase
HPLC analysis, and it was demonstrated that the purity of each
oligomer was >95% (Supporting Information). Ensuring the
absence of both cyclic urea and unwanted oligomer
contaminants was crucial to allow proper experimental model
validation, as the presence of cyclic urea would result in an
overestimation of the degradation rate and the presence
Polymer Synthesis. In addition to the oligomers that
possessed exact molecular weights due to their convergent
syntheses, it was also a goal of this work to demonstrate the
influence of chain length on the degradation time in longer
polymeric systems. As shown in Scheme 2, the polycarbamate
can be synthesized from the activated carbonate 5, along with
1c as an end-cap in the presence of DMAP and a tertiary amine
base.37 As this polymerization follows a step-growth process,
the molecular weights of the resulting polymers are not easy to
control. However, in order to obtain a reference set of polymers
displaying varying chain lengths, a two-level, four-factor
designed experiment was conducted to investigate the effects
of end-cap ratio, temperature, reaction duration, and the
nucleophilicity of the amine base on the degree of polymer-
ization. The results of this study (Supporting Information)
were used to select adequate polymerization conditions to
E
dx.doi.org/10.1021/ma4009753 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
elimination, the degradation of the model polymer can be
treated as a series of first-order intramolecular cyclization
reactions between successive chain lengths as follows:
Scheme 2. Synthesis of Polymers 6 and 7
−M
−M
−M
−M
P ⎯⎯→⎯ P ⎯⎯⎯→ ··· ⎯⎯→⎯ P ⎯⎯→⎯ M
n
n−1
1
kn
kn−1
k2
k1
(1)
where Pi is the polymer chain of length i, M is the released
monomer unit, and ki is the rate of intramolecular cyclization
for a terminal cyclization spacer on a polymer chain of length i.
This assumption is valid, as the rate of 1,6-eliminations in
hydroxybenzyl alcohol-based system are typically quite rapid
with complete conversion to the quinone methide occurring in
several seconds,46 while the cyclization of N,N′-dimethylethy-
lenediamine in aqueous systems is much slower with a
previously reported half-life of 36.3 min at 37 °C and pH
7.4.48 In the simplest case, the intramolecular cyclization rate
constant is independent of chain length, and eq 1 can be
reduced to the following set of linear ordinary differential
equations:
synthesize both a low and high molecular weight polymer. It
was found that using our previously reported conditions37
involving 0.05 equiv of end-cap, DIPEA as a base, and room
temperature conditions for 6 h provided a relatively low
molecular weight polymer 6 that was purified by dialysis in
DMF against a 6000−8000 g/mol molecular weight cutoff
(MWCO) membrane to yield a polymer with a Mn of 5250 g/
d[P ]
dt
n
= −k[P ]
n
(2)
1
mol and a PDI of 1.47 (Figure 2b). Analysis by H NMR
spectroscopy revealed a monomer to end-cap ratio of 25:1,
which was similar to the monomer feed ratio of 20:1. It was
found that a lower end-cap ratio (0.01 equiv), NEt3 as a base,
and polymerization at 0 °C for 24 h provided a higher
molecular weight polymer 7, which was purified by dialysis in
DMF against a 25 000 g/mol MWCO membrane to yield a
polymer with a Mn of 13 600 g/mol and a PDI of 1.58 (Figure
2b). Analysis by 1H NMR spectroscopy revealed a monomer to
end-cap ratio of 101:1, which was similar to the monomer feed
ratio of 100:1.
d[P]
dt
i
= k([Pi+1] − [P]) ∀ i ≤ n − 1
i
(3)
where [Pi] is the concentration of polymer chains of length i
and t is the time elapsed in the degradation process. This
assumption has been previously shown to hold true for
dendritic systems21 and will later be validated through our
experimental systems. Using an integrating factor, eqs 2 and 3
can be solved for all chain lengths and generalized into the
following form:
Linear Self-Immolative Degradation Model. As part of
this work, we aimed to probe the kinetics of linear self-
immolative degradation using computational simulations. It has
long been inferred that the degradation kinetics of self-
immolative polymers are dictated by the kinetics of the
monomer repeat units, yet no conclusive study has been
conducted to relate the two on a theoretical and quantitative
basis. The kinetics of linear self-immolative degradation have
been previously explored on a number of oligomers based on
1,4- and 1,6-elimination spacers and on elongated cyclization
spacers; however, analysis of the degradation kinetics has either
been qualitative or limited to the fitting of higher order
models.44,46,47 Moreover, these studies often fail to isolate end-
cap removal from polymer degradation and rely on composite
degradation plots to infer kinetic relationships. The Shabat
group has shown that the kinetics between successive
dendrimer generations in systems derived from cyclization
and elimination spacers approximately follows first-order
cyclization kinetics and used this observation to derive
theoretical release models for second-generation dendrons.21,30
In this work we sought to build off the promising results
observed for these simple dendrimer systems and derive a
theoretical model for the degradation of linear self-immolative
polymers that relates monomer kinetics to the overall kinetics
of polymer degradation and use this model to predict and
explain experimental degradation trends.
i
(kt)i−j
(i − j)!
[Pn−i] = e−kt
[P ]
n−j 0
∑
j=0
(4)
Since N,N′-dimethylimidazolidone and 4-hydroxybenzyl
alcohol are ultimately released upon every cyclization, the
concentration of either released monomer unit can be used as a
measure of polymer degradation as follows:
n
d[M]
dt
= k
[P]
∑
i
(5)
i=1
M(t)
M∞
M(t)
D(t) =
=
n
i=1
∑
i[P]
(6)
i 0
where D(t) is the relative degradation of the self-immolative
polymers at time t. Although not a true measure of molecular
weight, the trend in the appearance of degradation byproducts
is inversely related to the number-average molecular weight.
Tracking the degradation in this manner also allows for a more
robust method of practical measurement in that small changes
in the concentration of degradation products can be measured
with a higher degree of accuracy compared to small changes in
Mn or Mw.
In the case of initial monodispersity, eqs 2, 3, and 5 can be
solved algebraically to yield the following expression for the
evolution of degradation products during the linear self-
immolative process:
The linear self-immolative degradation of the model polymer
may be described as a sequence of elementary reactions
proceeding from the chain terminus. Assuming that the rate of
cyclization is significantly slower than the rate of 1,6-
F
dx.doi.org/10.1021/ma4009753 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
Figure 3. Mixed-mode degradation profile for the depolymerization of linear self-immolative polymers involving an initial zero-order domain
followed by a gradual transition toward first-order behavior.
⎛
⎞
the degradation time, the computational simulations suggest
that the time for degradation is proportional to rn at a fixed PDI
(Table S1). Analysis of the effect of PDI, however, is slightly
more complicated. The mixed-mode degradation kinetics are
most pronounced for monodisperse samples and become
skewed as PDI is increased (Figure S1d). This can be explained
by considering that the fraction of both low and high molecular
weight species increases as PDI is raised at a fixed rn.
Consequently, the degradation kinetics become more heavily
weighted toward both small and large chain lengths, which
results in a minor increase to the relative rate of degradation
early in the process and a substantial reduction to the rate
during the latter stages (Figure S1 and Table S1).
n
i
(kt)j−1
−kt
⎜
⎟
⎟
⎠
M(t) = [P ] n − e
∑ ∑ (j − 1)!
n 0
⎜
⎝
i=1 j=1
(7)
where n is the chain length of the initial monodisperse polymer
species. Substituting into eq 6 to provide a measure of self-
immolative degradation:
⎛
⎞
n
i
M(t)
(kt)j−1
= [P ] 1 − e−ktn−1
⎜
⎟
⎟
⎠
D(t) =
∑ ∑ (j − 1)!
n 0
⎜
n[P ]
n 0
⎝
i=1 j=1
(8)
Mixed-Mode Degradation Kinetics. Analysis of the
system of differentials describing the linear self-immolative
degradation process reveals a mixed-mode kinetic phenomenon
unique to linear self-immolative polymers. Unlike traditional
biodegradable polyesters that display pseudo-first-order degra-
dation kinetics, linear self-immolative polymers exhibit mixed-
mode kinetics in which the degradation displays an initial zero-
order phase followed by a gradual transition toward first-order
behavior over the course of degradation. This theoretical
behavior is consistent with our previous experimental findings
for the degradation of linear self-immolative polymer systems
that required fitting of a modified first-order model to account
for the sharp rise in the initial zero-order domain.39 Such
behavior can be explained by considering that the concen-
tration of polymer chains is not affected until complete
degradation of a given chain into monomer units occurs
(Figure 3). This causes the concentration of polymer chains to
remain relatively constant during the initial stages of
degradation, thereby imparting pseudo-zero-order character-
istics to eq 5. The length and duration of the apparent zero-
order phase are dependent on PDI, but also on the length of
polymer chains, as the resistance to first-order behavior
(polynomial term in eq 4) is enhanced by the presence of
high molecular weight species.
Oligomer Degradation Kinetics. Based on the promising
results of the simulation studies, the degradation kinetics for the
oligomer samples were investigated and compared to the
expected theoretical behavior. The solution degradation
1
kinetics for compounds 1d−4d were studied by H NMR
spectroscopy in 3:2 0.1 M pH 7.4 phosphate buffer
(D2O):acetone-d6 at 37 °C. The Boc protecting groups of
compounds 1a−4a were cleaved by treatment with TFA/
CH2Cl2, and the solvent was removed to afford compounds
1d−4d (Scheme 3). The oligomer−TFA salts were then
Scheme 3. Degradation of Compounds 1d−4d To Form
N,N′-Dimethylimidazolidinone and 4-Hydroxybenzyl
Alcohol
Simulation Studies. Prior to the experimental investigation
of the degradation kinetics for the oligomer and polymer
samples, computational simulations were conducted to
investigate the effects of degree of polymerization (rn) and
PDI on the kinetics of self-immolative degradation. The set of
linear differentials describing the self-immolation process (eqs
2, 3, and 5) were solved numerically on a series of artificial log-
normal polymer distributions using the experimentally
determined rate constant for intramolecular cyclization of
1.61 × 10−1 min−1 (Supporting Information). As shown in
Figure S1, the degradation time of linear self-immolative
polymers is influenced by rn and PDI. Using the times required
to reach 50% (t50) and 95% (t95) completion as indicators of
washed with 1 M citric acid/brine and concentrated under
reduced pressure to remove any residual TFA immediately
prior to dissolving the resulting oligomers in 1 mL of
buffer:acetone. Incorporation of the wash step was necessary
to ensure the pH was maintained at 7.4 during the degradation
process. Citric acid was selected as its pKa (3.15) was
G
dx.doi.org/10.1021/ma4009753 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
sufficiently high to remove the residual TFA but also acidic
enough to ensure no premature degradation occurred during
the wash step (Figures S21 and S22). The relative degradation
of the oligomeric species was quantified and assessed by
integrating the methyl peak of N,N′-dimethylimidazolidinone
relative to residual solvent peaks in the samples (Figure 4a).
Table 1. Kinetic Parameters for the Degradation of
Monodisperse Oligomers 1d−4d and Polymers 6 and 7
a
b
c
compd
n
k × 101 (min−1
)
t50 (min)
1d
2d
3d
4d
6
1
2
4
8
1.61 0.37
1.48 0.10
1.60 0.38
1.73 0.34
5.26 0.55
4.79 0.24
4.5 0.9
8.1 0.9
13.2 1.7
23.9 4.3
17.4
20
51
7
41.5
a
Number of repeat units (based on estimated Mn measured by SEC
b
for the polymers, sample was polydisperse). Rate constant for
monomer cyclization. Time for 50% polymer degradation. Depicted
c
errors correspond to standard deviations determined through triplicate
measurement for oligomers and 95% confidence intervals for the
polymers determined through nonlinear regression (t50 was not a
regressed parameter, so confidence intervals are not reported for
polymers).
This trend was quantified using the time to reach 50%
degradation (t50), which was found to be proportional to the
size of the oligomer species within experimental error (Table
1). Fitting the degradation data for each oligomer species to the
respective solutions to eq 8 using nonlinear regression provided
reasonable fits to all of the experimental data with significant
parameter estimations (p < 0.001) and low residual standard
errors (<5%) (Supporting Information). Additionally, the
fitting of all oligomeric species past the monomer stage to eq
8 offered better fits to the experimental data than the standard
first-order degradation model (eq 8, n = 1), evaluated based on
the Akaike information criterion (AIC) and Bayesian
information criterion (BIC) (Supporting Information). More
importantly, however, the monomer cyclization rate constants
determined through nonlinear regression (Table 1) displayed
no significant difference based on a one-way ANOVA test (p =
0.82), thereby validating the linear self-immolative degradation
model relating monomer kinetics to polymer degradation.
While a standard first-order model did provide a reasonable fit
for the oligomeric data below the tetramer stage, this model is
mechanistically inappropriate as the kinetics of linear self-
immolative degradation cannot be first-order above the
monomer stage. Such an observation also suggests that the
mixed-mode kinetic phenomenon is not clearly apparent in
short chain oligomers with a practical limit for the apparent
zero-order effect occurring at the octamer stage.
The half-life for monomer cyclization of 4.5 min determined
in this study was significantly shorter than the previously
reported half-life of 35 min.37 The large discrepancy in the
cyclization kinetics can likely be attributed to insufficient
buffering of residual TFA from the Boc deprotection step in the
previously reported case, as we found that the half-life was
significantly longer and pH values of the solution measured
using a microelectrode were lower when the citric acid wash
was not performed. The buffering capacity of the solvent used
for the NMR kinetics studies can be estimated to correspond to
less than 3 μL of TFA, suggesting the high sensitivity of the
system to residual TFA. Use of the citric acid wash to remove
TFA led to reproducible data for all oligomers, as demonstrated
in triplicate experiments (Supporting Information).
Figure 4. (a) Degradation kinetics of compounds 1d−4d, as measured
1
by H NMR spectroscopy in 0.1 M phosphate buffer (D2O):acetone-
d6 (3:2) at 37 °C. Representative samples from triplicate runs of 1d
■
◆
●
▲
( ), 2d ( ), 3d ( ), and 4d ( ). Solid lines correspond to the
regressed fits of eq 8 to each generation of oligomer (n = 1, 2, 4, 8).
■
▲
(b) Degradation kinetics of polymers 6 ( ) and 7 ( ), as measured
by H NMR spectroscopy in 0.1 M phosphate buffer (D2O):acetone-
1
d6 (3:2) at 37 °C. Overlayed lines correspond to the self-immolative
model fits for both polymers.
NMR spectra before and after degradation are provided in the
Supporting Information, showing conversion to the expected
products 4-hydroxybenzyl and N,N′-dimethylimidazolidinone
(Figures S23−S26). The complete degradation of all oligomer
species except the monomer was confirmed by SEC (Figure
2a). Model validation and kinetic analyses were carried out by
fitting the degradation model to the oligomeric degradation
profiles. As stated beforehand, application of the linear self-
immolative degradation model to a monodisperse sample set
allows for an algebraic solution to the system of differentials (eq
8) and requires no approximations in determining the initial
polymer distribution relative to chain length. This is particularly
important in comparing the fitted monomer cyclization rate
constant between oligomer samples of different lengths, as any
approximations in the calculation of chain length would alter
the form of the nonlinear degradation model and introduce an
artificial dampening or inflation to the rate constant.
Polymer Degradation Kinetics. The solution degradation
kinetics of polymers 6 and 7 were investigated under the same
1
As shown in Figure 4a and Table 1, the degradation time was
proportional to the length of the oligomer species, with longer
oligomers requiring more time to degrade than shorter ones.
conditions used for compounds 1d−4d (Figure 4b). H NMR
spectroscopy supported degradation of the polymers into the
expected degradation products (Figures S27 and S28);
H
dx.doi.org/10.1021/ma4009753 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
however, SEC analysis revealed the presence of a small fraction
of polymers 6 (Mn = 2020 g/mol, PDI = 1.26) and 7 (Mn =
8000 g/mol, PDI = 1.07) remaining after degradation (Figure
2b). Similar to other studies involving linear self-immolative
polymers, this nondegraded fraction is likely related to the
presence of non-end-capped cyclic oligomers that are formed
during polymerization.38−40 These peaks were neglected in the
treatment of the degradation data as fractional analysis revealed
that they were present at less than 3 wt % in the initial
distributions.
As shown in Figure 4b, the polymer degradation times were
proportional to Mn, with polymer 6 reaching 50% degradation
by 17.4 min and complete degradation by 150 min, while
polymer 7 reached 50% degradation by 41.5 min and complete
degradation by 200 min. These trends are in agreement with
the differences in Mn and follow the trends depicted in Figure
S1b, as polymer 7 displayed a 2.6-fold difference in Mn and
required 2.4 times longer to reach 50% degradation. Based on
the simulation results shown in Figure S1, the slight
discrepancy between the relative differences in Mn and t50 can
be attributed to the low molecular weight tailing observed for
polymer 6, which leads to an increased relative rate of
degradation in the early stages of the depolymerization process.
Although this slight tailing would have a minor influence on the
degradation time, the determining factor in these studies was
the large difference in Mn. Consequently, the results of this
study indicate that altering the chain length is an effective
means to tune the degradation time of linear self-immolative
polymers involving cyclization spacers.
The degradation profiles for polymers 6 and 7 were also fit to
the self-immolative model using a modified nonlinear
regression algorithm in which the numerical solution to eqs
2, 3, and 5 was applied to the SEC chromatograms for both
polymers. As shown in Figure 4b, the linear self-immolative
model offered reasonable fits to the degradation data with
standard errors of 6.6% and 4.7% and coefficients of
determination (R2) of 0.94 and 0.98 for polymers 6 and 7,
respectively (Figures S30 and S31). The regressed monomer
cyclization rate constants of 5.26 × 10−1 and 4.79 × 10−1 min−1
for polymers 6 and 7 were higher than expected based on the
oligomer data but were within experimental error based on the
spread of rate constants determined for the oligomer samples
(Table 1). Hence, the self-immolative degradation model was
an adequate predictor of polymer degradation and offered
quantitative evidence supporting the influence of chain length
on the rate of degradation. The one major limitation of this
model is that it requires an a priori knowledge of the chain
fraction distribution for a given polymer sample. Absolute
molecular weight data could not be obtained for these
polymers, thereby necessitating the use of relative molecular
weight data in the fitting of the self-immolative degradation
model. Although molecular weight data relative to PMMA
offered a more reasonable approximation compared to
polystyrene in DMF, the molecular weights reported in this
study were still overestimated by a factor of at least 1.6 based
on the limited SEC data obtained for the monodisperse
oligomers. Therefore, the rapid monomer cyclization rate
constants determined through nonlinear regression on the
polymeric data were subject to an artificial inflation due to the
use of relative molecular weight data. Work in our group is
currently underway to extend the results and improve the fitting
of the polymer degradation model to different systems to
further demonstrate the applicability of the self-immolative
degradation model as a means of predicting and explaining
experimental degradation profiles.
CONCLUSIONS
■
In conclusion, chain length was shown as an alternate means to
tune the time required for complete degradation in linear self-
immolative polymers. In this work we presented a new
convergent iterative synthesis of a set of monodisperse
oligomers based on the previously reported linear self-
immolative polymer backbone derived from alternating 4-
hydroxybenzyl alcohol and N,N′-dimethylethylenediamine
spacers. The degradation time of these oligomers in 0.1 M
pH 7.4 phosphate buffered (D2O):acetone-d6 (3:2) was shown
to be proportional to chain length. This finding was then
extended to the polydisperse case using a set of polymers
synthesized to display varying molecular weights. A theoretical
model describing linear self-immolative degradation was also
proposed and fit to the experimental degradation profiles to
provide quantitative evidence supporting the influence of chain
length on degradation time. This model, revealing the mixed
first-order and zero-order degradation kinetics of self-
immolative linear polymers, offers new insight into the
degradation process and should be applicable to a variety of
linear self-immolative polymers that undergo depolymerization
via a series of intramolecular reactions whose rates are
independent of chain length. It should also be possible in
future work to incorporate end-cap removal into the model. In
conjunction with rational monomer design, the controlled use
of chain length should allow for two-dimensional tuning of the
degradation time in linear self-immolative polymers and thus
significantly expand the usefulness of these materials for diverse
applications such as sensing and programmable release.
ASSOCIATED CONTENT
* Supporting Information
■
S
NMR spectra of compounds 1a, 2a, 2b, 2c, 3a, 3b, 3c, and 4a
as well as polymers 6 and 7; RP-HPLC chromatograms of
compounds 1a, 2a, 2b, 2c, 3a, 3b, 3c, and 4a; 1H NMR spectra
of compound 1d and deprotected polymer 7 prior to and
1
following citric acid wash; H NMR spectra of compounds 1d,
2d, 3d, and 4d as well as polymers 6 and 7 prior to and
following degradation; simulation studies; summary of DOE
factorial design conducted to vary polymer molecular weight;
and regression fits for all oligomeric compounds. This material
is available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*E-mail: egillie@uwo.ca (E.R.G.).
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank the Natural Science Research Council of
Canada (Discovery Grant Program and CGSM), the Ontario
Graduate Scholarship Program, and the Canada Research
Chairs Program for funding.
REFERENCES
■
(1) Kulkarni, R. K.; Moore, E. G.; Hegyeli, A. F.; Leonard, F. J.
Biomed. Mater. Res. 1971, 5, 169.
(2) Vert, M. Biomacromolecules 2005, 6, 538.
I
dx.doi.org/10.1021/ma4009753 | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
(3) Antheunis, H.; van der Meer, J.-C.; de Geus, M.; Heise, A.;
Koning, C. E. Biomacromolecules 2010, 11, 1118.
(4) Esser-Kahn, A. P.; Odom, S. A.; Sottos, N. R.; White, S. R.;
Moore, J. S. Macromolecules 2011, 44, 5539.
(5) Gao, W.; Chan, J. M.; Farokhzad, O. C. Mol. Pharmaceutics 2010,
7, 1913.
(6) Zhao, C.; Nie, S.; Tang, M.; Sun, S. Prog. Polym. Sci. 2011, 36,
1499.
(7) Xin, Y.; Yuan, J. Polym. Chem. 2012, 3, 3045.
(42) Schmid, K. M.; Jensen, L.; Phillips, S. T. J. Org. Chem. 2012, 77,
4363.
(43) Robbins, J. S.; Schmid, K. M.; Phillips, S. T. J. Org. Chem. 2013,
78, 3159.
(44) de Groot, F. M. H.; Loos, W. J.; Koekkoek, R.; van Berkom, L.
W. A.; Busscher, G. F.; Seelen, A. E.; Albrecht, C.; de Bruijn, P.;
Scheeren, H. W. J. Org. Chem. 2001, 66, 8815.
(45) Warnecke, A.; Kratz, F. J. Org. Chem. 2008, 73, 1546.
(46) Lee, H. Y.; Jiang, X.; Lee, D. Org. Lett. 2009, 11, 2065.
(47) Kevwitch, R. M.; Shanahan, C. S.; McGrath, D. V. New J. Chem.
2012, 36, 492.
(8) Bauhuber, S.; Hozsa, C.; Breunig, M.; Gopferich, A. Adv. Mater.
̈
2009, 21, 3286.
(48) Saari, W. S.; Schwering, J. E.; Lyle, P. A.; Smith, S. J.;
Engelhardt, E. L. J. Med. Chem. 1990, 33, 97.
(9) Meng, F.; Hennink, W. E.; Zhong, Z. Biomaterials 2009, 30, 2180.
(10) Wei, H.; Zhuo, R.-X.; Zhang, X.-Z. Prog. Polym. Sci. 2012, 38,
503.
(11) Zelzer, M.; Todd, S. J.; Hirst, A. R.; McDonald, T. O.; Ulijn, R.
V. Biomater. Sci. 2012, 1, 11.
(12) Wu, Q.; Wang, L.; Yu, H.; Wang, J.; Chen, Z. Chem. Rev. 2011,
111, 7855.
(13) de Gracia Lux, C.; Joshi-Barr, S.; Nguyen, T.; Mahmoud, E.;
Schopf, E.; Fomina, N.; Almutairi, A. J. Am. Chem. Soc. 2012, 134,
15758.
(14) Pasparakis, G.; Manouras, T.; Argitis, P.; Vamvakaki, M.
Macromol. Rapid Commun. 2011, 33, 183.
(15) Fomina, N.; Sankaranarayanan, J.; Almutairi, A. Adv. Drug
Delivery Rev. 2012, 64, 1005.
(16) Liu, G.; Liu, W.; Dong, C.-M. Polym. Chem. 2013.
(17) Mejia, J. S.; Gillies, E. R. Polym. Chem. 2013, 4, 1969.
(18) Wong, A. D.; DeWit, M. A.; Gillies, E. R. Adv. Drug Delivery Rev.
2012, 64, 1031.
(19) Blencowe, C. A.; Russell, A. T.; Greco, F.; Hayes, W.;
Thornthwaite, D. W. Polym. Chem. 2011, 2, 773.
(20) Peterson, G. I.; Larsen, M. B.; Boydston, A. J. Macromolecules
2012, 45, 7317.
(21) Amir, R. J.; Pessah, N.; Shamis, M.; Shabat, D. Angew. Chem., Int.
Ed. 2003, 42, 4494.
(22) de Groot, F. M. H.; Albrecht, C.; Koekkoek, R.; Beusker, P. H.;
Scheeren, H. W. Angew. Chem., Int. Ed. 2003, 42, 4490.
(23) Szalai, M. L.; Kevwitch, R. M.; McGrath, D. V. J. Am. Chem. Soc.
2003, 125, 15688.
(24) Amir, R. J.; Danieli, E.; Shabat, D. Chem.Eur. J. 2007, 13, 812.
(25) Sella, E.; Shabat, D. Chem. Commun. 2008, 5701.
(26) Weinstain, R.; Sagi, A.; Karton, N.; Shabat, D. Chem.Eur. J.
2008, 14, 6857.
(27) Sella, E.; Shabat, D. J. Am. Chem. Soc. 2009, 131, 9934.
(28) Amir, R. J.; Shabat, D. Chem. Commun. 2004, 1614.
(29) Shamis, M.; Lode, H. N.; Shabat, D. J. Am. Chem. Soc. 2004,
126, 1726.
(30) Perry, R.; Amir, R. J.; Shabat, D. New J. Chem. 2007, 31, 1307.
(31) Sagi, A.; Segal, E.; Satchi-Fainaro, R.; Shabat, D. Bioorg. Med.
Chem. 2007, 15, 3720.
(32) Zhang, H.; Yeung, K.; Robbins, J. S.; Pavlick, R. A.; Wu, M.; Liu,
R.; Sen, A.; Phillips, S. T. Angew. Chem., Int. Ed. 2012, 51, 2400.
(33) Weinstain, R.; Baran, P. S.; Shabat, D. Bioconjugate Chem. 2009,
20, 1783.
(34) Esser-Kahn, A. P.; Sottos, N. R.; White, S. R.; Moore, J. S. J. Am.
Chem. Soc. 2010, 132, 10266.
(35) Seo, W.; Phillips, S. T. J. Am. Chem. Soc. 2010, 132, 9234.
(36) Sagi, A.; Weinstain, R.; Karton, N.; Shabat, D. J. Am. Chem. Soc.
2008, 130, 5434.
(37) DeWit, M. A.; Gillies, E. R. J. Am. Chem. Soc. 2009, 131, 18327.
(38) DeWit, M. A.; Beaton, A.; Gillies, E. R. J. Polym. Sci., Part A:
Polym. Chem. 2010, 48, 3977.
(39) Chen, E. K. Y.; McBride, R. A.; Gillies, E. R. Macromolecules
2012, 45, 7364.
(40) de Gracia Lux, C.; McFearin, C. L.; Joshi-Barr, S.;
Sankaranarayanan, J.; Fomina, N.; Almutairi, A. ACS Macro Lett.
2012, 1, 922.
(41) DeWit, M. A.; Gillies, E. R. Org. Biomol. Chem. 2011, 9, 1846.
J
dx.doi.org/10.1021/ma4009753 | Macromolecules XXXX, XXX, XXX−XXX