Synthesis of rhodamine 6G-based compounds for the ATRP synthesis of fluorescently labeled biocompatible polymers

Article abstract of DOI:10.1021/bm200311s

Facile derivatization of rhodamine 6G in the 2′ position by direct reaction with secondary amines is reported. If the secondary amine contains a hydroxy group, the hydroxyl-functional intermediate can be readily esterified to give either fluorescent initiators for atom transfer radical polymerization (ATRP) or a fluorescent methacrylic comonomer. In contrast to rhodamine dyes functionalized using primary amines, which are only fluorescent at low pH, these compounds are highly fluorescent at physiological pH. These new compounds were subsequently used to prepare a range of fluorescently labeled biocompatible polymers based on the biomimetic monomer, 2-(methacryloyloxy)ethyl phosphorylcholine (MPC), for biomedical studies.

Full text of DOI:10.1021/bm200311s

ARTICLE
pubs.acs.org/Biomac
Synthesis of Rhodamine 6G-Based Compounds for the ATRP Synthesis
of Fluorescently Labeled Biocompatible Polymers
Jeppe Madsen,* Nicholas J. Warren, and Steven P. Armes*
Dainton Building, Department of Chemistry, University of Sheffield, Brook Hill, Sheffield, Yorkshire, S3 7HF, United Kingdom
Andrew L. Lewis
Biocompatibles UK Ltd., Chapman House, Farnham Business Park, Weydon Lane, Farnham, Surrey, GU9 8QL, United Kingdom
S Supporting Information
b
ABSTRACT: Facile derivatization of rhodamine 6G in the 20
position by direct reaction with secondary amines is reported. If
the secondary amine contains a hydroxy group, the hydroxyl-
functional intermediate can be readily esterified to give either
fluorescent initiators for atom transfer radical polymerization
(ATRP) or a fluorescent methacrylic comonomer. In contrast
to rhodamine dyes functionalized using primary amines, which
are only fluorescent at low pH, these compounds are highly
fluorescent at physiological pH. These new compounds were
subsequently used to prepare a range of fluorescently labeled
biocompatible polymers based on the biomimetic monomer,
2-(methacryloyloxy)ethyl phosphorylcholine (MPC), for bio-
medical studies.
’ INTRODUCTION
literature examples of polymerizable vinylic rhodamine deriva-
tives,11ꢀ14 including at least one commercially available rhoda-
mine-labeled monomer.15 Usually, such dye-functionalized
monomers are copolymerized with conventional vinyl mono-
mers to give statistical copolymers with relatively low dye
contents. In contrast, using monofunctional fluorescent initiators
allows the chromophore to be placed precisely at the polymer
chain end.
There appear to be no reports of rhodamine-based initiators for
atom transfer radical polymerization (ATRP), although a number of
other fluorescent dye initiators have been used to prepare labeled
copolymers.16ꢀ23 For example, a 2-bromoisobutyrate ester of fluo-
rescein allowed good control to be obtained for the polymerization
of N-isopropylacrylamide.16 However, the relatively poor photo-
stability of fluorescein5 combined with the hydrolytic instability of
aromatic esters24 suggests that such labeled polymers may not be
ideal for biomedical applications that require prolonged monitor-
ing over extended time periods (days to weeks) in aqueous solu-
tion. According to Zhang and co-workers, an ATRP initiator based
on phenyloxazole17 is reasonably efficient. However, this chro-
mophore has an emission maximum at 370 nm, and autofluores-
cence of cellular constituents is likely to be problematic at this
wavelength. Similarly, the anthracene-based initiator reported
The use of synthetic polymers for the intracellular delivery of
drugs requires a detailed knowledge of the final fate of the
macromolecular vector. One important tracking approach is to
fluorescently label the polymer chains. This allows their diffusion
within tissue and live cells to be monitored in situ using
established techniques such as fluorimetry and confocal fluores-
cence microscopy.1ꢀ4 Ideally, the dye label should emit in a
spectral region where there is minimal autofluorescence from
either cells or body fluids/organs. In addition, high quantum
yields are clearly advantageous, since this can either increase
sensitivity or minimize the degree of labeling required. In the
latter case, this allows significant cost savings for relatively
expensive dyes and reduces possible toxic side effects due to
the dye label. Finally, dyes with high photostabilities are pre-
ferred. Rhodamines are one class of dyes that fulfill all of the
above requirements, because they have high quantum yields,
emit in the red part of the visible spectrum, are relatively cost-
effective, and offer good photostability.5
In general, there are several possible methods for the covalent
attachment of a dye label onto a polymer chain. A range of
functionalized rhodamines are commercially available for cou-
pling via various chemistries.6 These dyes are commonly used for
labeling specific sites in biological macromolecules.7,8 Recently,
alternative dyes have also been used for labeling synthetic polymers.9
However, such reactive labels are significantly more expensive than
their nonreactive counterparts.10 In addition, there are several
Received: March 7, 2011
Revised:
April 8, 2011
Published: April 11, 2011
r
2011 American Chemical Society
2225
dx.doi.org/10.1021/bm200311s Biomacromolecules 2011, 12, 2225–2234
|

Biomacromolecules
Scheme 1. Base-Induced Conversion of 20-Substituted Rhodamine 6G from its Hydroquinone Form to its Spirolactone Form
ARTICLE
Scheme 2. General Reaction between Rhodamine 6G and Amines to Form Various Substituted Amidesa
a Numbers in parentheses are isolated yields after purification by recrystallization.
by Klumperman’s group18 presumably has an emission max-
imum at around 400 nm (i.e., similar to that of native an-
thracene), which may also lead to autofluorescence problems.
Initiators based on substituted naphthalimides exhibit maximum
emissions at around 500 nm,18,21 which is close to that of
fluorescein. Thus, the former could also be a useful alternative
label to the latter commonly used dye. On the other hand,
rhodamine dyes are generally more photostable than fluorescein5
and are also relatively water soluble compared to naphthalimide-
based dyes.25,26
Here we focus on modifying the commercially available dye,
rhodamine 6G, because this compound has an exceptionally high
quantum yield (> 0.9) compared to rhodamine B, which typically
has quantum yields of the order of 0.3ꢀ0.4.10 Moreover, given
the very high absorption coefficient of rhodamine 6G (in excess
of 100 000 Mꢀ1 cmꢀ1), only a minimal amount of label is
necessary, which should minimize any adverse effects that might
be caused by the dye label.27
as a comonomer. Indeed, we have recently reported using such
rhodamine-labeled copolymers for monitoring intracellular up-
take of vesicles28ꢀ30 and also for studying vesicle diffusion into
tissue-engineered human oral mucosa.31
Rhodamine dyes exist in their fluorescent hydroquinone form
at neutral/acidic pH and in their nonfluorescent spirolactone
form at basic pH (Scheme 1).32 Amidation of rhodamine esters
in the 20 position has been reported in several recent papers and
patents.10,32ꢀ36 Primary amines react directly with the cyclic
ester to form secondary amides under mild conditions. For these
compounds, conversion to the cyclic spirolactam occurs at a
lower solution pH than for the rhodamine ester precursor, and no
significant fluorescence was observed above pH 6.37 Thus, such
compounds are not likely to be useful as fluorescence probes
within living tissue at pH 7.4. However, they may offer some
potential as fluorescent pH indicators.38,39 A synthetically elegant
solution to this problem has been reported, where a rhodamine-
based secondary amide was coupled to a fluorescein dye.37 This
fluorophore emits light over a wide pH range at a wavelength that
depends on the solution pH. However, this approach requires a
multistep synthesis.
In principle, chemical modification of rhodamine 6G should
allow this biomedically relevant chromophore to be readily
incorporated into polymer chains, either as ATRP initiators or
2226
dx.doi.org/10.1021/bm200311s |Biomacromolecules 2011, 12, 2225–2234

Biomacromolecules
ARTICLE
Scheme 3. Esterification of Three Hydroxy-Functional Rhodamine Derivatives to Produce Various Fluorescently-Labeled ATRP
Initiators and a Fluorescently-Labeled Methacrylic Monomer
If a tertiary amide rhodamine dye derivative is used instead of a
secondary amide, then internal amide formation at high pH is
prevented. Thus conjugation is retained and no loss of fluores-
cence is observed in neutral or alkaline solution.10 Formation of
the tertiary amide does not occur under mild conditions, but it
has been reported using either benzotriazole coupling agents35,36
or highly reactive Lewis acids.10 In contrast, other commonly
used amidation reagents such as carbodiimides afforded only low
yields.10 In addition, the synthesis of rhodamine-based acid
halides has been described in the patent literature. These highly
reactive compounds have been used to prepare a range of tertiary
amide derivatives.33,34 However, this approach precludes the use
of functional amines such as γ-aminoalcohols, unless protecting
group chemistry is employed. Therefore, an additional synthetic
step is required to prepare hydroxy-functional rhodamine dyes
that exhibit pH-independent fluorescence.10 On the other hand,
direct reaction between cyclic lactones and secondary amines has
been reported to afford high yields under relatively mild condi-
tions, particularly if a large excess (up to 20 equiv) of the amine is
used.40 In practice, the amine can be used as a reactive solvent. As
far as we are aware, this attractive approach has not previously
been reported for the preparation of rhodamine-based tertiary
amides.
amide linkages (Scheme 2). In addition, protocols for the
esterification of both hydroxy-functional secondary amides and
tertiary amides to produce a series of fluorescent 2-bromoisobu-
tyryl esters are described (Scheme 3). These compounds can be
used as fluorescent ATRP initiators for the controlled polymer-
ization of methacrylic monomers (Scheme 4). In addition, a
rhodamine 6G methacrylic ester has also been synthesized.
The labeled initiator was evaluated by preparing a range of
fluorescently labeled poly(2-methacryloyloxy)ethyl phos-
phorylcholine) (PMPC) homopolymers by ATRP. In addition,
two pH-responsive, vesicle-forming block copolymers compris-
ing PMPC and poly(2-(diisopropylamino)ethyl methacrylate)
(PDPA) were prepared (Scheme 5).41 One was prepared using
the fluorescent ATRP initiator (with the MPC being polymer-
ized first) and therefore had a single rhodamine 6G label at the
PMPC terminus. The other was prepared by statistically incor-
porating the fluorescent methacrylic monomer into the pH-
responsive PDPA block. The pH-dependent emission of these
two copolymers depended on the spatial location of the fluo-
rophore. Thus, if the labeled initiator was used, self-assembly led
to a reduction in fluorescence to approximately half of the
original intensity of the initially dissolved copolymer chains.
On the other hand, if the labeled comonomer was incorporated
into the pH-responsive PDPA block, self-assembly led to com-
plete fluorescence quenching.
Herein we report a convenient one-step synthesis of a range
of hydroxy-functional rhodamine 6G-based dyes with tertiary
2227
dx.doi.org/10.1021/bm200311s |Biomacromolecules 2011, 12, 2225–2234

Biomacromolecules
ARTICLE
Scheme 4. Synthesis of Rhodamine-PMPCn Homopolymers by ATRP Using Initiator 7 or 8
’ RESULTS AND DISCUSSION
the reaction was also attempted using rhodamine B instead of
rhodamine 6G. However, little or no tertiary amide was formed
with the former dye. The main difference here is that the
rhodamine 6G precursor is an ethyl ester, whereas the rhodamine
B starting material was present in its free carboxylic acid form.
Secondary amines are both more basic and sterically congested
than 3-aminopropan-1-ol. Thus, in addition to the desired
tertiary amide, there is also some lactone formation due to
internal cyclization. Unfortunately, although this byproduct can
react with primary amines,10 it is unreactive toward secondary
amines, which therefore reduces the overall yield.
Esterification of Hydroxy-Functional Rhodamine 6G Deri-
vatives. Synthetic routes to various rhodamine 6G-based esters
are shown in Scheme 3. The secondary amide, 1, was isolated in
its nonprotonated spirolactam form. Addition of excess 32%
aqueous HCl to a suspension of this compound in acetonitrile
gave a deep red solution, indicating protonation of the amine
groups and formation of the conjugated hydroquinone form.
Heating to reflux afforded better solubility and addition of
2-bromoisobutyryl bromide gave the target product in 94% yield
within 3 h (Scheme 3). The resulting ATRP initiator, 2, was
isolated in sufficient purity (g95% by 1H NMR and HPLC) to be
used directly for polymer syntheses. Further purification (>99%)
could be achieved either by recrystallization from methanol or by
preparative reverse phase HPLC.32,37
Reaction between Rhodamine 6G and 3-Aminopropan-1-ol.
The literature reaction32,37 between rhodamine 6G and primary
amines is shown in Scheme 2a. This reaction was reported to
proceed spontaneously at room temperature in DMF, with
spirolactam yields ranging from 54 to 92%, depending on the
primary amine used. In our hands, replacing DMF with aceto-
nitrile gave an isolated yield of 89% when using 3-aminopropan-
1-ol. Because acetonitrile is much more volatile (and hence
easier to remove) than DMF, the former solvent was preferred
for such reactions. As both reactants are water-soluble, purifica-
tion was readily achieved by washing the water-insoluble product
1
with excess water. H and 13C NMR spectra and ES-MS anal-
ysis were all consistent with the target structure (see Supporting
Information).
Reaction between Rhodamine 6G and Secondary Amines.
The direct reaction between rhodamine 6G and a secondary
amine is depicted in Scheme 2b. The secondary amine was used
as a reactive solvent in this synthesis, typically using 1.0 g
rhodamine 6G dye per gram of amine. Maintaining the reaction
mixture at 90 °C for 17ꢀ23 h gave the desired tertiary amide in
52ꢀ75% yield. The main byproduct was the cyclic lactone, as
determined by electrospray mass spectroscopy (ES-MS). Var-
ious secondary amines were evaluated, as indicated in Scheme 2.
Rhodamine 6G is highly soluble in protic solvents such as
alcohols.42 In this context, it is perhaps noteworthy that the
synthesis of a similar hydroxy-terminated rhodamine B derivative
has been reported in two steps, with an overall yield of 50%.10
Thus, our one-step protocol produces comparable or better
yields without requiring protecting group chemistry. All products
were highly water-soluble, whereas the cyclic lactone byproduct
is water-insoluble. Similar aqueous solubility has also been
reported for a related rhodamine B-based compound.10 Thus,
Using the same protocol with tertiary amide 3 as a substrate
gave the desired product but in a much lower yield. ES-MS
analysis indicated that amide hydrolysis was prevalent in this case
and that the acid was the main byproduct. This indicates that the
tertiary amide is significantly more prone to acidic hydrolysis
than the secondary amide. In addition, it was found by ES-MS
that if the amine hydrochloride salt form of the dye was used, a
significant amount of the 2-chloroisobutyryl ester was obtained,
presumably due to halogen exchange occurring during the
2228
dx.doi.org/10.1021/bm200311s |Biomacromolecules 2011, 12, 2225–2234

Biomacromolecules
ARTICLE
Scheme 5. Synthesis of pH-Responsive Diblock Copolymers 7-PMPC25-PDPA90 and PMPC25-P(DPA66-91) Using Sequential
Monomer Addition
reaction. In general, the overall yield of the targeted ester was
only around 10%. Several modifications of the reaction condi-
tions were examined. For example, addition of base leads to
deprotonation of the secondary aromatic amine. This entity
reacts with the acid bromide to give an amide in addition to
the desired ester. This route was not synthetically useful, because
the resulting products were difficult to separate. On the other
hand, use of 2-bromoisobutyric anhydride instead of the 2-
bromoisobutyryl bromide afforded 2-bromoisobutyric acid as a
byproduct, instead of HBr. Because the former acid is weaker,
it caused virtually no amide hydrolysis. Unfortunately, this
reaction was very slow in common organic solvents such as
acetonitrile and DMF, with only 10ꢀ20% conversion being
achieved over 4ꢀ5 days even at 80ꢀ90 °C; this is probably
related to the low solubility of the tertiary amide in these aprotic
solvents. However, using 2-bromoisobutyric acid as solvent
significantly improved the yield. This acid melts at 47 °C, hence
it is necessary to work above this temperature. This approach is
illustrated in Scheme 3. At 70 °C, a conversion of around 70%
was obtained after 24 h for the reaction with 3 (see Supporting
Information, Figure S1). Because the resulting ester is highly
soluble in dichloromethane after neutralization, purification is
relatively straightforward. The deprotonated form of the piper-
azine-based initiator 8 proved to be less soluble in water than the
deprotonated initiator 7, as judged by the reduced coloration of
the aqueous phase in the former case; this is believed to be the
main reason for the higher isolated yields of 8.
A similar approach was used for the preparation of rhodamine-
based methacrylic monomer 9 (Scheme 3). This reaction was
conducted at 20 °C, which is above the m.p. of methacrylic acid
(16 °C) but sufficiently low to avoid thermal polymerization.
Compound 5 had relatively low solubility in pure methacrylic
acid, thus, it was necessary to add chloroform as a cosolvent in
this particular case. This approach gave a yield of 76%.
Absorption Maxima and Absorption Coefficients Ob-
tained for Various Rhodamine Derivatives. Table S1 shows
the wavelengths of maximum absorption and absorption coeffi-
cients determined for the various rhodamine 6G derivatives. For
the two dyes containing a secondary amide group (1 and 2), the
maximum absorption wavelength is essentially the same as for
rhodamine 6G, both in water and in acidified methanol. The
absorption coefficient of 1 in methanol containing 0.1% trifluor-
oacetic acid (TFA) is similar to that of rhodamine 6G in
methanol, whereas it is significantly lower for 2. The absorption
coefficients observed in 0.1 M HCl are significantly lower than
those reported for compounds prepared using diamines, which
are typically of the order of 60000.37 On the other hand, a more
complex adduct of normetanephrine and rhodamine 6G has an
absorption coefficient of 41400 cmꢀ1 Mꢀ1 32
which is much
,
closer to the value of 34000 ( 2000 cmꢀ1 Mꢀ1 obtained for 2.
Both of these dyes were isolated in their spirolactam form
(Scheme 1), which is not directly soluble in water at neutral
pH. These dyes only dissolved very slowly in 0.1 M HCl, even
with heating and ultrasonic treatment. Dissolution was rapid in
32% HCl, which could then be diluted with water without
precipitation. However, the absorption coefficients determined
using this protocol were of the order of 5000 Mꢀ1 cmꢀ1, which is
significantly lower than that obtained for rhodamine 6G and also
for the rhodamine derivatives prepared with tertiary amides.
Both esters could be purified by recrystallization from THF.
On the other hand, attempts to purify these products by silica
column chromatography were unsuccessful, because both
column adsorption and a significant degree of hydrolysis were
observed.
2229
dx.doi.org/10.1021/bm200311s |Biomacromolecules 2011, 12, 2225–2234

Biomacromolecules
ARTICLE
Instead, stock solutions for spectroscopic studies were prepared
by dissolving these chromophores in methanol containing 0.1%
v/v TFA followed by serial dilution using 0.1 M aqueous HCl.
These observations suggest that the conversion of each dye to its
fluorescent hydroquinone form is relatively slow and may not go
to completion in aqueous acid. Moreover, poorer solvent quality
is known to reduce the absorption coefficient.43
Solvency effects were observed for almost all the rhodamine
derivatives in Table S1. Compounds 3 and 4 have absorption
coefficients close to those of unmodified rhodamine 6G in
methanol, whereas the corresponding values observed in 0.1 M
aqueous HCl are generally lower, indicating that water is a poorer
solvent than methanol.
The absorption coefficient for compounds 5 and 6 in metha-
nol is around 10% lower than that for rhodamine 6G. This may be
due to increased steric congestion or, in the case of compound 6,
it may be due to reduced solvation of the n-butyl group in
methanol. Similar observations were made in the same solvent
for ATRP initiators 7 and 8 and also for the monomer, 9, which
each exhibit absorption coefficients around 80% of that of
rhodamine 6G in methanol.
Figure 1. Absorption spectra obtained for 7 in methanol and 0.1 M
HCl.
The absorption coefficients of compounds 5ꢀ7 in 0.1 M HCl
is significantly lower than those observed in methanol, which is
probably related to its reduced solvation (as for compounds 3
and 4). Figure 1 shows absorption spectra recorded for 7
dissolved in both 0.1 M HCl and methanol. Despite the
differing concentrations, the maximum absorbance is almost
identical for these two solutions. However, the relative absor-
bance at 508 nm is significantly less in methanol than
in 0.1 M HCl. This feature is directly related to the aggregation
of the dye molecules.26,44 According to Figure 1, there is a
significantly higher degree of dye aggregation in 0.1 M HCl,
which confirms that solvent quality influences fluorescent
intensity, as expected.
In contrast, the piperazine-based ATRP initiator and mono-
mer (8 and 9) exhibit very similar absorption coefficients in 0.1 M
HCl and methanol. These compounds have additional amine
functionality due to the piperazine moiety. This extra amine
becomes protonated at low pH, which enhances the aqueous
solubility of such compounds relative to the other derivatives.
pH-Dependence of Absorption and Emission Behavior of
Rhodamine 6G Derivatives. Figure 2 shows typical normalized
absorption and emission fluorescence spectra obtained for the
derivatized rhodamine dyes in acidic aqueous solution. These are
similar to those reported for rhodamine 6G.26,44
The effect of increasing the solution pH on the emission and
absorption spectra of a 10ꢀ5 M solution of 1 in dilute HCl is
shown in Figure 3a. The maximum emission and absorption at
530 nm both increase monotonically from pH 1 to pH 4. This is
because pH adjustment involves the addition of a dilute aqueous
base (see Supporting Information), which shifts the dye’s
unimerꢀdimer equilibrium in favor of the unimers.26,44 Because
dimers act as fluorescence quenchers,44 increasing the relative
unimer concentration leads to an increase in both the absorption
at 530 nm and therefore also the emission intensity, provided
that the increase in unimer concentration is larger than the
dilution factor. Raising the pH leads to precipitation, which
increases the background scattering in these absorption spectra.
This is due to formation of the water-insoluble nonfluore-
scent spirolactam form of the dye. Above pH 4, this becomes
the dominant factor in the attenuation of the absorption and
emission spectra. This effect is also evident in digital
Figure 2. Normalized absorption and emission spectra of 3 in aqueous
HCl at pH 2.0. The emission spectrum was recorded using an excitation
wavelength of 530 nm.
photographs of the aqueous solutions/suspensions (see Sup-
porting Information, Figure S3). Figure 3b shows the variation
of the relative emission and the relative A530/A508 ratio as a
function of pH for a 10ꢀ5 M solution of 3. The shoulder at
approximately 508 nm is attributed to dimer formation and is
somewhat higher than the reported value of 496 nm for the
rhodamine 6G dimer.44 This spectral shift may be due to
substituent effects for these dyes, but overlapping peaks make
the precise location of such features rather problematic and the
precise determination of the wavelength for dimer absorption is
outside the scope of this paper. Nevertheless, the change in the
A530/A508 ratio corresponds to a change in the unimer/dimer
ratio. Increasing the pH from 1.5 to 10 more than doubles the
emission intensity, despite concomitant dilution of the solu-
tion. This is believed to be due to a shift in the unimerꢀdimer
equilibrium.
Relative quantum yields of the rhodamine 6G derivatives that
were soluble at physiological pH were determined using the
method described by Fery-Forgues and Lavabre.10,45 These
values are listed in Table S1. In general, the quantum yields
observed for the dyes with hydrophilic substituents (i.e., com-
pounds 3ꢀ6 and 10) are comparable to those of rhodamine 6G,
whereas the quantum yields obtained for dyes with hydrophobic
2230
dx.doi.org/10.1021/bm200311s |Biomacromolecules 2011, 12, 2225–2234

Biomacromolecules
ARTICLE
rhodamine is attached to the polymer. These homopolymers
were purified by dialysis against methanol using membranes with
a molecular weight cutoff of 1000. This protocol inevitably
removes oligomers, which leads to a higher Mn for the purified
homopolymer. However, homopolymers prepared using a target
DP of 50 (Table S2, entries 2 and 6) also have lower initiator
contents than anticipated, as judged by 1H NMR. Thus around
75% of the aromatic groups are intact for 7-PMPC50, whereas for
8-PMPC50 this value is around 62%. Although the experimental
error in the integrated aromatic signals in these longer chains is
higher, these results indicate that loss of oligomers during dialysis
is not the main reason that the apparent molecular weight is
higher than targeted. There are two alternative explanations for
this observation. Either the initiator efficiency is less than 100%,
that is, not all initiators are incorporated into the chain, or the
initiator end groups are partially lost during polymerization and
purification. If the initiator efficiency is less than 100%, the
molecular weight should increase proportionally, that is, the
polymer made from the initiator with the lowest efficiency should
be longer. Thus, in principle, this hypothesis can be evaluated by
comparing 1H NMR and GPC homopolymers made from either
7 or 8 with the same target degrees of polymerization. For
example, entries 1 and 5 (which each had a target DP of 20) or
entries 2 and 6 (both target DP of 50) in Table S2 can be
compared. The 1H NMR data indicate that the polymers based
on initiator 8 (see entries 2 and 6 in Table S2) should have higher
molecular weights than those based on initiator 7 (entries 1 and 5
in Table S2). However, the GPC data indicates that, in fact, 8-
PMPC20 is shorter than 7-PMPC20 and, similarly, that 8-
PMPC50 is shorter than 7-PMPC50. Of course, an imperfect
initiator efficiency cannot be excluded on the basis of these
results because the GPC data are relative to the poly(ethylene
oxide) calibration standards. Indeed, GPC analysis of a PMPC20
homopolymer synthesized using a previously reported morpho-
line-based ATRP initiator48 gave a lower Mn value (∼12000 g
molꢀ1, data not shown), indicating that the initiator efficiency is
Figure 3. (a) Effect of increasing the solution pH on the maximum
emission normalized with respect to pH 1.0 and absorbance at 530 nm
for an aqueous solution initially containing 5 ꢁ 10ꢀ5 M 1. (b) Effect of
increasing the pH on the maximum emission and the absorbance ratio at
530 and 508 nm respectively for an aqueous solution initially containing
1 ꢁ 10ꢀ5 M of 3.
substituents (i.e. 7ꢀ9) are significantly lower. This indicates that
relatively poor solvation adversely affects the relative quantum
yield, as expected.
Use of Rhodamine-Labeled ATRP Initiators and a
Methacrylic Monomer To Prepare PMPC Homopolymers.
Initial experiments with the pH-dependent fluorescent ATRP
initiator, 2, confirmed that this compound efficiently initiated
polymerization of MPC. However, because it is only fluorescent
below pH 4, it is not particularly relevant for most biological
studies, hence it was not explored further. The pH-independent
ATRP initiators 7 and 8 were used to prepare PMPC homo-
polymers via a previously reported protocol, as illustrated in
Scheme 4 for initiator 7.46,47 Table S2 summarizes the character-
ization data obtained for the various PMPC homopolymers
prepared using the rhodamine-labeled initiators 7 and 8. The
maximum absorption wavelength was red-shifted by 5ꢀ10 nm
for all molecularly dissolved copolymers relative to that of their
corresponding initiators. This indicates a change in the local
environment,43 which is presumably due to the presence of the
highly hydrophilic polymer chains.
1
around 80%. This corresponds fairly well to the H NMR DP,
hence, the loss of end groups during polymerization appears to
be a relatively minor problem.
In view of the above observations, loss of aromatic initiator
signals during polymerization may explain the observed discre-
pancies for polymers prepared using initiators 7 and 8. It has
previously been shown that tertiary amine methacrylates can
undergo partial transesterification in methanol to form methyl
methacrylate and the corresponding alcohol.49 Similar transes-
terification of the initiator would lead to formation of methyl
2-bromoisobutyrate and the corresponding rhodamine alcohol.
However, no evidence for such transesterification side-reactions
were observed by either 1H NMR or HPLC in solutions of 7 or 8
in perdeuterated methanol over 24 h (data not shown). The
addition of two equivalents of 2,20-bipyridine had no effect and
neither did the addition of both 2,20-bipyridine and copper(II)
bromide (data not shown). However, in the presence of the
ATRP catalyst (i.e., 1 equiv of copper(I) bromide and 2 equiv of
2,20-bipyridine in nitrogen-purged perdeuterated methanol),
rapid transesterification was observed (as shown for initiator 8,
Figure S6). Thus, some transesterification is likely to occur
during the PMPC polymerization, which would account for the
higher DP values that are observed.
Both initiators give well-defined homopolymer chains with
relatively low polydispersities for target degrees of polymeriza-
tion up to 100 (Table S2 and Figure S4). However, reduced
control is achieved if a degree of polymerization of 200 is
targeted, as evidenced by an increase in polydispersity and the
appearance of a multimodal GPC trace.
For PMPC homopolymers prepared with a target DP of 20
using initiators 7 and 8, the actual DP determined by 1H NMR
corresponds fairly well to that expected (see Table S2, entries 1
and 5, and Figure S5). In both cases, this experimental DP
In general, the mean DP determined by absorption spectros-
copy using the absorption coefficient of the initiator was much
higher than the targeted DP values and also somewhat higher
1
determined by H NMR is slightly higher than that targeted,
which suggests that only around 80% of the theoretical amount of
2231
dx.doi.org/10.1021/bm200311s |Biomacromolecules 2011, 12, 2225–2234

Biomacromolecules
ARTICLE
than the values determined by 1H NMR, where applicable (Table S1).
This was the case for both initiators, although there was a closer
correlation when using 8 than 7 (Table S2, compare entries 1ꢀ4
with entries 5ꢀ8). Thermogravimetric analyses (see Supporting
Information, Figure S7) indicated that these homopolymers
contain around 15% water, even after extensive drying under
vacuum for 24 h at 90 °C. This is not unexpected, because it is
well-known that water binds tenaciously to PMPC.50 Unfortu-
nately, even allowing for such relatively high water contents
cannot account for the high DP values observed. A possible
solvent effect was examined by determining the mean DP in
methanol for selected polymers. However, this gave an essen-
tially identical DP to that obtained in 0.1 M HCl (data not
shown). An alternative explanation may be that the absorption
coefficient of the initiator is reduced when conjugated to a
polymer chain, due to steric congestion of the fluorophore.51 If
that is the case, initiator 7 should be more affected than initiator
8, because the latter has a longer spacer group between the
conjugated ring system and the initiator moiety. This hypothesis
is consistent with the absorption data presented in Table S2.
Use of Rhodamine 6G-Labeled ATRP Initiators and a
Methacrylic Monomer To Prepare pH-Responsive Diblock
Polymers. Initiator 7 was also used to prepare a pH-responsive
diblock copolymer comprising MPC and 2-(diisopropylaminoethyl)
methacrylate, DPA, with a targeted composition of 7-PMPC25-
PDPA90 (see Table S2, entry 9). Such diblock copolymers have
previously been shown to be molecularly dissolved under acidic
conditions but to self-assemble to form polymer vesicles at around
physiological pH.41
Figure 4. (a) Variation of hydrodynamic diameter and relative fluores-
cence intensity on increasing the pH of 0.20% w/w aqueous solutions
of 7-PMPC25-PDPA90 and PMPC25-(PDPA66-9) with 1 M NaOH.
(b) Relative change in the 530 and 508 nm absorption bands compared
to the maximum normalized fluorescence intensity and relative fluores-
cence intensity as a function of pH. The initial volume was 25 mL and the
total amount of added base was no more than 2 mL.
1
For 7-PMPC25-PDPA90, the H NMR data agrees with the
target block composition. It was not possible to determine the
amount of rhodamine directly in the final copolymer by this
method due to its relatively high DP. Based on the results
obtained for the homopolymers discussed above, the amount
of rhodamine in the polymer is probably less than one label per
chain. This copolymer had a unimodal GPC trace and a narrow
polydispersity of 1.22, indicating a well-controlled polymeriza-
tion. However, it was necessary to apply a different GPC-
protocol to that used for the PMPC homopolymers, as the
PDPA block does not dissolve in water at neutral pH. Therefore,
the GPC results are not directly comparable. The mean DP of 7-
PMPC25-PDPA90 determined by absorption spectroscopy is
around 25% higher than that targeted. This is somewhat closer
to the target value than that achieved for PMPC homopolymers
prepared using this initiator (compare entry 9 with entries 1ꢀ4 in
Table S2).
PMPC-PDPA diblock copolymers was investigated by dynamic
light scattering, fluorescence spectroscopy, and absorption spec-
troscopy. Figure 4 shows that the hydrodynamic diameter for
both diblock copolymers increases by an order of magnitude
between pH 6 and 7, indicating vesicle formation as previously
reported.41 The relative fluorescence intensity shows a marked
difference between the two copolymers. Up to pH 6, a small
reduction is observed for both copolymers, which is probably
largely due to dilution effects caused by the addition of base.
Between pH 6 and 7, a significant decrease in the relative
fluorescence intensity is observed in both cases, which correlates
with the increased hydrodynamic diameter. Above pH 7, there is
a marked difference between the two copolymers: The 7-
PMPC25-PDPA90 vesicles are still fluorescent, but the relative
fluorescence intensity is reduced to approximately half of that
observed below pH 6. In contrast, the fluorescence observed for
the PMPC25-P(DPA66-stat-91), is completely quenched above
pH 7. To understand this observation, absorption spectra were
recorded as a function of pH. More specifically, the maximum
absorbance of the rhodamine unimer (at approximately 540 nm)
was compared to that due to the rhodamine dimer at 508 nm
from pH 2 to pH 9 (Figure S8).44 These results are shown in
Figure 4b. From pH 2 to pH 6, the Amax/A508 ratio does not
change, indicating that the unimerꢀdimer equilibrium is not
affected. This was expected, because there is on average one
rhodamine per copolymer chain and, below pH 6, these chains
are molecularly dissolved, which is why the pH-dependence
should be similar to that of the native rhodamine dye (Figure 3b).
For 7-PMPC25-PDPA90, there is a maximum of one fluorophore
per chain (due to the initiator fragment). However, for PMPC25-
P(DPA66-stat-91), a certain fraction of chains will contain more
A similar diblock copolymer was prepared using monomer 9.
In this case, a nonfluorescent ATRP initiator was used to target a
PMPC25-P(DPA70-stat-91) copolymer and a total of 1 equiv of
1
monomer 9 was used per copolymer chain. H NMR studies
revealed that the actual copolymer composition corresponded to
PMPC25-P(DPA66-stat-91) (Table S2, entry 10). This alternative
approach to the synthesis of fluorescent diblock copolymers also
led to a relatively narrow polydispersity. As with the longer
PMPC-based homopolymers discussed above, the exact amount
of 9 incorporated into the copolymer could not be accurately
determined by 1H NMR due to its relatively low concentration.
However, the mean DP estimated from absorption spectroscopy
studies is very close to that targeted in this case.
Hydrodynamic Size and Fluorescence Emission of Rhoda-
mine 6G-Labeled PMPC-PDPA Diblock Copolymers as a
Function of pH. The pH-dependent behavior of the two
2232
dx.doi.org/10.1021/bm200311s |Biomacromolecules 2011, 12, 2225–2234

Biomacromolecules
ARTICLE
than one label, especially if the reactivity ratios of 9 and DPA
deviate significantly from unity. In this case, the dye labels may be
so close to each other that dimer formation is significant even for
the molecularly dissolved chains (Figure S8). This explains why
the Amax/A508 ratio is lower for PMPC25-P(DPA66-stat-91) than
for 7-PMPC25-PDPA90.
In related work with collaborators, we have already demon-
strated that these rhodamine 6G-labeled copolymers have the
appropriate molecular characteristics for a range of biological
studies, such as various intracellular delivery experiments.28ꢀ31
’ ASSOCIATED CONTENT
Above pH 7, the Amax/A508 ratio shifts significantly, indicating
dimer formation. This is more pronounced for PMPC25-P-
(DPA66-stat-91), whose absorption spectrum is very similar to
the literature spectra reported for rhodamine dimers.44 There-
fore, polymer self-assembly above neutral pH, combined with a
hydrophobic environment strongly favors formation of dimers
and possibly even higher order aggregates.52 In contrast, when
the pH is increased for a dilute aqueous solution of 7-PMPC25-
PDPA90, the relative distance between chromophores is greater
and these labels experience a highly polar environment. Thus,
these results indicate that dimer formation due to high local
fluorophore concentration and/or decreased solvent quality is
the main cause of fluorescence quenching. In addition, there may
be some contribution from the tertiary amines in the PDPA block
as these species are known to act as fluorescence quenchers.53
S
Supporting Information. Full experimental protocols
b
and characterization data for all reported compounds, transester-
ification kinetic data, additional absorption spectra, digital photo-
graphs, TGA curves. This material is available free of charge via
the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: p.madsen@sheffield.ac.uk; s.p.armes@sheffield.ac.uk.
’ ACKNOWLEDGMENT
Biocompatibles UK Ltd. is thanked for CASE support of J.M.
and N.J.W., for supplying the MPC monomer, and for permis-
sion to publish this work. The University of Sheffield is thanked
for a Ph.D. studentship to support J.M. BBSRC is thanked for a
Ph.D. studentship to support N.J.W. Dr. J. A. Balmer and Mr.
L. A. Fielding are acknowledged for assistance with the TGA
measurements. The three reviewers of this manuscript are
acknowledged for their useful comments.
’ CONCLUSIONS
A facile one-step protocol has been developed to prepare
hydroxy-functional rhodamine derivatives. Esterification of these
protonated precursors using 2-bromoisobutyryl bromide or
2-bromoisobutyric anhydride afforded three new rhodamine
6G-based ATRP initiators in good yields. One initiator exhibited
fluorescence below pH 4 but was nonfluorescent at higher pH,
while two initiators proved to be highly fluorescent over a wide
pH range (from pH 1 to pH 10). A new permanently fluorescent
rhodamine 6G-based methacrylic monomer was also synthesized
using a similar approach. All compounds exhibited similar
absorption characteristics to rhodamine 6G. In addition, the
quantum yields estimated for the hydrophilic labels were similar
to those obtained for rhodamine 6G in PBS at pH 7, whereas the
hydrophobic labels had somewhat lower quantum yields. The
two initiators were used to prepare reasonably well-defined,
rhodamine-labeled poly(2-(methacryloyloxy)ethyl phos-
phorylcholine) via ATRP in methanol at 20 °C. For both
initiators, mean degrees of polymerization calculated for these
homopolymers using absorption spectroscopy were significantly
higher than those estimated by 1H NMR spectroscopy. In
general, end-group analyses calculated from absorption spectros-
copy correlated more closely with 1H NMR data for initiator 8
than for initiator 7. The former has a longer spacer connecting
the chromophore to the polymer chain, which may lead to less
perturbation of the intrinsic chromophore signature.
In addition, two rhodamine 6G-labeled vesicle-forming di-
block copolymers were synthesized. One copolymer was pre-
pared using a rhodamine initiator and therefore contained a
maximum of one terminal fluorophore per chain. The other
copolymer was synthesized using a methacrylic rhodamine
monomer, which was statistically incorporated into the tunably
hydrophobic PDPA block. In both cases, increasing the solution
pH above 6 led to vesicle formation, which led to a significant
reduction in fluorescence. However, if the fluorophore was
attached to the terminus of the hydrophilic chains, only partial
reduction in fluorescence was observed, whereas statistical
incorporation of the fluorophore into the pH-responsive mem-
brane-forming block led to complete quenching.
’ REFERENCES
(1) Jensen, K. D.; Kopeꢀckovꢁa, P.; Bridge, J. H. B.; Kopeꢀcek, J. AAPS
PharmSci. 2001, 3, 62–75.
(2) Jensen, K. D.; Nori, A.; Tijerina, M.; Kopeꢀckovꢁa, P.; Kopeꢀcek, J.
J. Controlled Release 2003, 87, 89–105.
(3) Richardson, S. C. W.; Wallom, K.-L.; Ferguson, E. L.; Deacon,
S. P. E.; Davies, M. W.; Powell, A. J.; Piper, R. C.; Duncan, R. J. Controlled
Release 2008, 127, 1–11.
(4) Mangold, S. L.; Carpenter, R. T.; Kiessling, L. L. Org. Lett. 2008,
10, 2997–3000.
(5) Konstantinova, T.; Cheshmedjieva-Kirkova, G.; Konstantinov,
Hr. Polym. Degrad. Stab. 1999, 65, 249–252.
(6) Haugland, R. P. The Handbook—A Guide to Fluorescent Probes
and Labeling Technologies, 10th edition; Invitrogen: Carlsbad, CA,
http://probes.invitrogen.com/handbook/.
(7) Zhang, J.; Campbell, R. E.; Ting, A. Y.; Tsien, R. Y. Nat. Rev. Mol.
Cell Biol. 2002, 3, 906–918.
(8) Gonc-alves, M. S. T. Chem. Rev. 2009, 109, 190–212.
(9) York, A. W.; Scales, C. W.; Huang, F.; McCormick, C. L.
Biomacromolecules 2007, 8, 2337–2341.
(10) Nguyen, T.; Francis, M. B. Org. Lett. 2003, 5, 3245–3248.
(11) Costela, A.; Garcia-Moreno, I.; Figuera, J. M.; Amat-Guerri, F.;
Mallavia, R.; Santa-Maria, M. D.; Sastre, R. J. Appl. Phys. 1996,
80, 3167–3173.
(12) Nicolas, J.; San Miguel, V.; Mantovania, G.; Haddleton, D. M.
Chem. Commun. 2006, 4697–4699.
(13) Obata, M.; Morita, M.; Nakase, K.; Mitsuo, K.; Asai, K.;
Hirohara, S.; Yano, S. J. Polym. Sci., Part A: Polym. Chem. 2007,
45, 2876–2885.
(14) Yang, H.; Vasudevan, S.; Oriakhi, C. O.; Shields, J.; Carter, R. G.
Synthesis 2008, 6, 957–961.
(15) 2-(Methacryloxy)ethyl thiocarbamoyl rhodamine B, Poly-
sciences, Inc., Cat. No. 23591.
(16) Lu, X.; Zhang, L.; Meng, L.; Liu, Y. Polym. Bull. 2007,
59, 195–206.
2233
dx.doi.org/10.1021/bm200311s |Biomacromolecules 2011, 12, 2225–2234

Biomacromolecules
ARTICLE
(17) Zhang, L.; Xu, Q.; Lu, J.; Xia, X.; Wang, L. Eur. Polym. J. 2007,
43, 2718–2724.
(52) Martínez, V. M.; Arbeloa, F. L.; Prieto, J. B.; Arbeloa, I. L. J. Phys.
Chem. B 2005, 109, 7443–7450.
(18) Zhang, H.; Klumperman, B.; van der Linde, R. Macromolecules
2002, 35, 2261–2267.
(53) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 2nd ed.;
Kluwer Academic/Plenum Publishers: New York, 1999, p 238.
(19) Jin, Z. N.; Lu, J. M.; Xu, Q. F.; Wang, L. H.; Xia, X. W. e-Polym.
2007, Article 006.
(20) Jin, Z.; Xu, Q.; Li, N.; Lu, J.; Xia, X.; Yan, F.; Wang, L. Eur.
Polym. J. 2008, 44, 1752–1757.
(21) Bardajee, G. R.; Li, A. Y.; Haley, J. C.; Winnik, M. A. Dyes Pigm.
2008, 79, 24–32.
(22) Beijad, M.; Charreyre, M.-T.; Martinho, J. M. G. Prog. Polym.
Sci. 2011, 36, 568–602.
(23) Beija, M.; Charreyre, M.-T.; Martinho, J. M. G. Prog. Polym. Sci.
2011, 36, 568–602.
(24) Blay, G.; Cardona, M. L.; Garcia, M. B.; Pedro, J. R. Synthesis
1989, 438–439.
(25) Arbeloa, F. L.; Ojeda, P. R.; Arbeloa, I. L. J. Chem. Soc., Faraday
Trans. 2 1988, 84, 1903–1912.
(26) Arbeloa, F. L.; Gonzalez, I. L.; Ojeda, P. R.; Arbeloa, I. L.
J. Chem. Soc., Faraday Trans. 2 1982, 78, 989–994.
(27) Johnson, L. V.; Walsh, M. L.; Chen, L. B. Proc. Natl. Acad. Sci.
U.S.A. 1980, 77, 990–994.
(28) Lomas, H.; Massignani, M.; Abdullah, K. A.; Canton, I.; Lo Presti,
C.; MacNeil, S.; Du, J.; Blanazs, A.; Madsen, J.; Armes, S. P.; Lewis, A. L.;
Battaglia, G. Faraday Discuss. 2008, 139, 143–159.
(29) Massignani, M.; LoPresti, C.; Blanazs, A.; Madsen, J.; Armes,
S. P.; Lewis, A. L.; Battaglia, G. Small 2009, 5, 2424–2432.
(30) Murdoch, C.; Reeves, K. J.; Hearnden, V.; Colley, H.; Massignani,
M.; Canton, I.; Madsen, J.; Blanazs, A.; Armes, S. P.; Lewis, A. L. Macromol.
Biosci. 2010, 5, 1025–1036.
(31) Hearnden, V.; Lomas, H.; MacNeil, S.; Thornhill, M.; Murdoch,
C.; Lewis, A. L.; Madsen, J.; Blanazs, A.; Armes, S. P.; Battaglia, G. Pharm.
Res. 2009, 26, 1718–1728.
(32) Adamczyk, M.; Grote, J. Bioorg. Med. Chem. Lett. 2000,
10, 1539–1541.
(33) Haugland, R. P.; Singer, V. L.; Yue, S. T. U.S. Patent 6,399,392.
(34) Mayer, U.; Oberlinner, A. U.S. Patent 4,647,675
(35) Tomeiro, M.; Still, W. C. J. Am. Chem. Soc. 1995, 117, 5887–
5888.
(36) Tomeiro, M.; Still, W. C. Tetrahedron 1997, 53, 8739–8750.
(37) Adamczyk, M.; Grote, J. Bioorg. Med. Chem. Lett. 2003,
13, 2327–2330.
(38) Han, J.; Burgess, K. Chem. Rev. 2010, 110, 2709–2728.
(39) Albertazzi, L.; Storti, B.; Marchetti, L.; Beltram, F. J. Am. Chem.
Soc. 2010, 132, 18158–18167.
(40) Blay, G.; Cardona, L.; Garcia, B.; Garcia, C. L.; Pedro, J. R.
Tetrahedron 1996, 52, 10507–10518.
(41) Du, J. Z.; Tang, Y. P.; Lewis, A. L.; Armes, S. P. J. Am. Chem. Soc.
2005, 127, 17982–17983.
(42) Schmidt, T. Acta Histochem. 1970, 38, 250–263.
ꢁ
ꢀ
(43) Miljaniꢁc, S.; Cimerman, Z.; Frkanec, L.; Ziniꢁc, M. Anal. Chim.
Acta 2002, 468, 13–25.
(44) Ghasemi, J.; Niazi, A.; Kubista, M. Spectrochim. Acta, Part A
2005, 62, 649–656.
(45) Fery-Forgues, S.; Lavabre, D. J. Chem. Educ. 1999, 76, 1260–
1264.
(46) Lobb, E. J.; Ma, I.; Billingham, N. C.; Armes, S. P.; Lewis, A. L.
J. Am. Chem. Soc. 2001, 123, 7913–7914.
(47) Ma, I.; Lobb, E. J.; Billingham, N. C.; Armes, S. P.; Lewis, A. L.;
Lloyd, A. W.; Salvage, J. P. Macromolecules 2002, 35, 9306–9314.
(48) Bories-Azeau, X.; Armes, S. P.; van den Haak, H. J. W. Macro-
molecules 2004, 37, 2348–2352.
(49) Bories-Azeau, X.; Armes, S. P. Macromolecules 2002,
35, 10241–10243.
(50) Ishihara, K.; Nomura, H.; Mihara, T.; Kurita, K.; Iwasaki, Y.;
Nakabayasji, N. J. Biomed. Mater. Res. 1998, 39, 323–330.
(51) Nijegorodov, N. I.; Downey, W. S. J. Phys. Chem. 1994,
98, 5639–5643.
2234
dx.doi.org/10.1021/bm200311s |Biomacromolecules 2011, 12, 2225–2234