Poly(ethylene glycol) (PEG) brush-arm star polymers (BASPs) with narrow mass distributions and tunable nanoscopic sizes are synthesized in via ring opening metathesis polymerization (ROMP) of a PEG-norbornene macromonomer followed by transfer of portions of the resulting living brush initiator to vials containing varied amounts of a rigid, photo-cleavable bis-norbornene crosslinker.
Convenient methods for the rapid, parallel synthesis of diversely functionalized nanoparticles will enable discovery of novel formulations for drug delivery, biological imaging, and supported catalysis. In this report, we demonstrate parallel synthesis of brush-arm star polymer (BASP) nanoparticles by the "brush-first" method. In this method, a norbornene-terminated poly(ethylene glycol) (PEG) macromonomer (PEG-MM) is first polymerized via ring-opening metathesis polymerization (ROMP) to generate a living brush macroinitiator. Aliquots of this initiator stock solution are added to vials that contain varied amounts of a photodegradable bis-norbornene crosslinker. Exposure to crosslinker initiates a series of kinetically-controlled brush+brush and star+star coupling reactions that ultimately yields BASPs with cores comprised of the crosslinker and coronas comprised of PEG. The final BASP size depends on the amount of crosslinker added. We carry out the synthesis of three BASPs on the benchtop with no special precautions to remove air and moisture. The samples are characterized by gel permeation chromatography (GPC); results agreed closely with our previous report that utilized inert (glovebox) conditions. Key practical features, advantages, and potential disadvantages of the brush-first method are discussed.
Polymeric nanoparticles have been widely studied for their potential use as platforms for drug delivery, supported catalysis, biological imaging, and self-assembly1-3. Modern applications require that nanoparticle syntheses be facile, reproducible, compatible with chemical functionalities, and amenable to diversification4,5. Ring-opening metathesis polymerization (ROMP) of strained olefins is a powerful methodology for the synthesis of functional polymeric nanostructures with controlled sizes and narrow mass distributions1,6-8. For example, norbornene-functionalized poly(ethylene glycol) (PEG) macromonomers (MMs) can be efficiently polymerized via ROMP to generate water soluble bottle-brush polymers. Using this approach, nanostructures that carry multiple releasable drug molecules, fluorophores, and spin-contrast agents can be prepared rapidly and in parallel6,9,10.
ROMP has also been used for the "arm-first" synthesis of star polymers. In the arm-first method, linear polymers are crosslinked with a multi-functional crosslinker to give spherical nanostructures with polymeric arms. Schrock and co-workers reported the first arm-first ROMP synthesis of star polymers via crosslinking of norbornene, dicarbomethoxynorbornadiene, and trimethylsilyl protected dicarboxynorbornene linear polymers with a bifunctional norbornene crosslinker.11,12 Buchmeiser has extended this methodology for the synthesis of materials with a range of applications that include supported catalysis, tissue-engineering, and chromatography13-17. Otani and coworkers have made star polymer nanoparticles with functional surfaces via a related "in-out" polymerization strategy18,19.
Most arm-first polymerizations involve a complex interplay of monomer, polymer, and star coupling reactions. The latter proceeds via a step-growth mechanism that typically leads to broad molecular weight (MW) distributions. To overcome this limitation in related arm-first atom transfer radical polymerization reactions, Matyjaszewski and coworkers performed arm-first crosslinking of preformed polymeric MMs to provide star polymers with very narrow MW distributions20. In this case, the steric bulk of the MMs, and the increased ratio of star arms to initiation sites, inhibited poorly controlled star+star coupling processes, and led to a living, chain growth mechanism.
When we attempted the same strategy in the context of ROMP with a norbornene-terminated PEG-MM and a bis-norbornene crosslinker, star polymers with very broad, multi-modal MW distributions were obtained. This result suggested that in this system the MM alone was not sufficiently bulky to inhibit star+star coupling. To increase the steric bulk of the star arms, and potentially limit this uncontrolled coupling, we attempted to first polymerize the MM to form bottle-brush polymers in the absence of crosslinker and then add the crosslinker. We were pleased to find that under certain conditions, this "brush-first" method provided straightforward access to "brush-arm star polymers" (BASPs) with narrow MW distributions and tunable core and corona functionalities.
We recently reported the brush-first ROMP synthesis of PEG BASPs using Grubbs 3rd generation catalyst A (Figure 1)21. In this work, exposure of PEG-MM B to catalyst A generated a living brush macroinitiator with defined backbone length (B1, Figure 1). Transfer of aliquots of the B1 to vials that contained different amounts of crosslinker C initiated BASP formation. The MW, and therefore the size, of the BASPs increased geometrically with the amount of C added. We provided a mechanistic hypothesis for this geometric growth process and demonstrated that functional, nitroxide core- and corona-labeled BASPs could be readily prepared without the need for post-polymerization modification steps or sequential monomer additions. However, in all of the reported examples, we were concerned about catalyst deactivation; we carried out all reactions under N2 atmosphere inside a glovebox.
Since our initial report, we have found that the brush-first method is very effective for the formation of BASPs from a wide range of norbornene-terminated MMs and functional crosslinkers. We have also discovered that the method can be performed on the benchtop with no special precautions to remove air or moisture.
Herein, a series of three BASPs of differing MWs will be synthesized by the brush-first method under ambient conditions. In brief, 10 equivalents of B will be exposed to 1.0 equivalents of catalyst A (Figure 1a) for 15 min to yield a BI with an average degree of polymerization (DP) of 10. Three aliquots of this batch of BI will be transferred to separate vials that contain 10, 15, and 20 equivalents (N, Figure 1b) of C. After 4 hr, the polymerizations will be quenched via addition of ethyl vinyl ether. The star polymer MWs and MW distributions will be characterized using a gel permeation chromatography instrument equipped with a multi-angle laser light scattering detector (GPC-MALLS).
We first describe the synthesis and purification of PEG-MM B from 3 kDa O-(2-aminoethyl)polyethylene glycol (PEG-NH2) and norbornene-N-hydroxysuccinimidyl (NHS) ester. The former compound can be purchased from Sigma Aldrich Inc., or prepared via anionic polymerization according to literature procedures22,23. The latter compound can be prepared in two steps according to a published procedure21. Next we describe a synthesis of catalyst A from commercially available Grubbs 2nd generation catalyst. We then demonstrate the use of this complex for brush-first BASP synthesis. This experiment details the procedure for making BASPs with N = 10, 15, and 20 from a BI with DP = 10. All reactions were performed in a fume hood using standard scintillation vials.
CAUTION: Always wear gloves, a lab-coat, and lab glasses, and follow common laboratory safety practices when working with hazardous chemicals. Any organic solvent must be handled in a fume hood. Solids can be weighed out on a balance outside the fume hood. Chemicals should not come into contact with skin, eyes, or mouth. It is strongly recommended to read the MSDS for every solvent and solid used in this procedure before beginning.
1. Preparation of PEG-MM B
2. Purification of PEG-MM
In our previous report, the PEG-MM B was prepared from commercially available PEG-NH2 and was used for BASP synthesis without further purification after drying (i.e., after step 1.7). In this study, we vary the PEG-NH2 source (commercial versus homemade), and we compare BASP formation results before and after more rigorous preparative high performance liquid chromatography (prep-HPLC) MM purification. In the remainder of this study, the dried MM obtained after step 1.7 is referred to as B1. Prep-HPLC was used to purify B1 to give B2. An analogous prep-HPLC purified MM synthesized in our laboratory via anionic polymerization is referred to as B3. Prep-HPLC was performed using a Beckmann Coulter HPLC (127p solvent module and 166p detector module) with a 1-ml sample loop and an Agilent Zorbax 300SB-C18 PrepHT reverse-phase column at room temperature.
3. Preparation of Catalyst A
4. Preparation of Stock Solution of Living Brush Polymer (BI) with DP = 10
5. Formation of BASPs
6. GPC Sample Preparation
The GPC-MALLS results were obtained on an Agilent 1260 LC system equipped with a Shodex GPC KD-806M column, a Wyatt Dawn Heleos-II MALLS detector, and a Wyatt Optilab t-rEX refractive index detector at room temperature. DMF with 0.025 M LiBr at a flow rate of 1.0 ml/min was used as the eluent. Results were analyzed using Astra 6 software provided by Wyatt.
List of Abbreviations:
A: Grubbs 3rd generation bis-pyridine catalyst
B: poly(ethylene glycol) (PEG) macromonomer (MM)
B1: PEG MM prepared using commercially available (Aldrich) PEG-NH2 and used without HPLC purification.
B2: PEG MM prepared using commercially available (Aldrich) PEG-NH2 and used after HPLC purification.
B3: PEG MM prepared using newly synthesized PEG-NH2 and used after HPLC purification.
BASP: brush-arm star polymer
BI: living brush initiator
C: photodegradable crosslinker
Ð: molar mass dispersity index
DMF: N,N-dimethylformamide
DP: number average degree polymerization
GPC: gel permeation chromatography
Prep-HPLC: preparative high performance liquid chromatography
MALLS: multi-angle laser light scattering
MM: macromonomer
MW: molecular weight
Mw: weight average molar mass
N: number of crosslinker equivalents (ratio of C à A)
NHS: N-hydroxysuccinimidyl
PEG: polyethylene glycol
PEG-MM: norbornene-PEG macromonomer (also referred to as compound B)
ROMP: ring-opening metathesis polymerization
THF: tetrahydrofuran
Figure 2 shows GPC traces for a variety of BASPs prepared from B1, B2, and B3. In all cases, the data illustrate that increasing the equivalents of crosslinker (N) leads to an increase in the size of the BASP. As was observed in our previous report, 10 equivalents of crosslinker is not sufficient to achieve uniform BASPs; the N = 10 sample shows a clearly multi-modal GPC trace with a large amount of residual brush polymer especially in the case of unpurified MM B1 (Figure 2a). Greater amounts of crosslinker result in uniform MW distributions with very little residual brush and MM. The weight-average molar mass (Mw) approximately doubles in going from N = 15- 20. In the case of B3, no residual MM and less than 1% residual BI remains for the N = 15 and N = 20 cases.
Figure 1. Schematic for Brush-Arm Star Polymer (BASP) Synthesis. Panel (a) illustrates the synthesis of Grubbs' 3rd generation bispyridine catalyst (A) from commercially available Grubbs' 2nd-generation catalyst. Also shown are the structures of the PEG-MM (B) and crosslinker (C) used in this work. Panel (b) shows a schematic diagram of the brush-first process. Polymerization of PEG-MM (B) with catalyst (A) generates a 10-unit living brush initiator (BI), which is then added to crosslinker (C) resulting in the formation of a BASP. Click here to view larger image.
Figure 2. Representative GPC results of the N = 10, 15, and 20 BASPs prepared from various PEG MMs. Panels (a), (b), and (c) depict data for MMs B1, B2, and B3, respectively. Impurities from commercial PEG-NH2, unreacted MM, and residual BI are labeled with asterisks. Mw and dispersity index (Ð) values are provided in the inset tables. Note that Ð values obtained by GPC for highly branched nanostructures must be considered carefully24,25. The observation of monomodal, uniform peaks suggests a narrow distribution of particle radii. Click here to view larger image.
The key advantage of brush-first BASP synthesis is the unique ability to rapidly synthesize nanostructures of diverse size and composition in parallel without need for specialized equipment. In this study, we demonstrate the brush-first synthetic method using a norbornene functionalized PEG macromonomer (B, Figure 1) and a bis-norbornene nitrobenzyl ester crosslinker (C, Figure 1). The PEG chains from B impart water solubility to the final BASP structure. The nitrobenzyl-based crosslinker is photodegradable.
This general procedure can be modified for other exo-norbornene based MMs and crosslinkers. We have prepared BASPs from several combinations of both. For example, we have used norbornene-PEG-based MMs that carry different anti-cancer drugs, nitroxides, and magnetic resonance imaging contrast agents 27. We have also used MMs comprised of polymers other than PEG. In our experience, the brush-first method can be applied to nearly any functional exo-norbornene imide terminated MM. In cases where high conversions of MM to BI (>95%) are not achieved, an MM impurity is the most likely culprit (as opposed to catalytic activity). More rigorous purification as outlined in this report (prep-HPLC) typically leads to successful ROMP. Note that we have not attempted ROMP polymerizations with MMs that bear unprotected functional groups that are known to interfere with catalyst A (e.g. free amines, olefins, azides, etc.). These groups can be introduced after the brush-first synthesis via post-polymerization modification27. For example, we have prepared azide-BASPs from alkyl halide MMs that were converted to azides after BASP formation. These azides were used for Cu-catalyzed azide-alkyne cycloaddition "click" reactions.
We sought to study the impact of MM purity in more detail. Small amounts of residual MM and BI were always observed in GPC traces when brush-first reactions were carried out using MM prepared from commercially available PEG-NH2 (B1, Figure 2a). We had learned from experience that completely pure MMs generally give quantitative MM conversion. Furthermore, we had noticed that the amount of residual MM varied depending on the batch number of the commercial PEG-NH2. We suspected that a non-functional PEG-NH2 impurity, perhaps simply PEG diol, was responsible for the apparent residual MM impurity. Therefore, we utilized prep-HPLC to purify B1 to give pure MM B2. Figure 2b shows that this purification process did indeed decrease the amount of residual MM (orange star) approximately two-fold; it did not remove it completely. Interestingly, B2 gave higher conversion of BI to BASPs as well; perhaps an impurity that led to catalyst deactivation was removed via prep-HPLC. Still dissatisfied with the amount of residual MM, we followed literature methods for the synthesis of PEG-NH2 via anionic polymerization of ethylene oxide from ethanolamine (CAUTION: Ethylene oxide should be handled by trained, experienced chemists; it is a highly flammable, explosive, and toxic gas!).22,23 MM prepared from this homemade PEG-NH2 (B3) yielded improved results compared to the commercial MMs. GPC analysis of the corresponding BASPs showed no detectable residual MM and very little (<1%) residual BI (Figure 2c). Thus, if high purity BASPs are required we recommend using the purest possible MM. Note that residual MM and BI can readily be removed from the larger BASPs through dialysis after brush-first synthesis.
We have also used crosslinkers other than C. For example, we have prepared BASPs from bisnorbornene metal complexes, polymerization initiators, acid-cleavable linkers, and supramolecular hosts. We find that crosslinkers with rigid spacers between the norbornenes tend to provide the most uniform BASPs; such crosslinkers are less likely to undergo intramolecular cyclization reactions that consume norbornenes but do not contribute to BASP growth.
Regardless of the MM and crosslinker combination, we find the following general practices will lead to the highest chance of brush-first success. First, before attempting brush-first synthesis with newly synthesized monomers, we recommend making the DP = 10 brush polymer alone and possibly longer brush polymers with DP = 25 and 50. If these tests are successful, there is an excellent chance that the brush-first method will also be successful. Second, the ideal concentration for the brush-first polymerization is dependent on monomer chemical composition and structure of the components. We recommend testing a few concentrations on small scale before making a large batch of BASP. Third, polymerizations carried out in dichloromethane or tetrahydrofuran appear to give the best results; monomers that are soluble in these solvents are ideal. As discussed above, if the crosslinker is poorly soluble in these solvents we recommend adding it as a solid rather than adding extra solvent. As long as the MM is soluble, we find that the crosslinking brings the crosslinker completely into solution within minutes. Fourth, though the polymerization does not require inert conditions, we recommend storage of the catalyst under inert atmosphere to increase its lifetime. Importantly, the catalyst will decompose over time in solution; the catalyst solution should be prepared fresh from the Grubbs third generation catalyst each time a series of ROMP reactions is performed. Finally, the amount of crosslinker required for uniform BASPs will vary widely with crosslinker and MM structure. As shown in Figure 2, 10 equiv of crosslinker C is not enough to provide complete BI conversions. In other cases, we find that addition of 1 equiv of crosslinker, and even up to 40 equiv, provides good results. Whenever a new crosslinker is to be used, we recommend running a series of small-scale reactions with various N values to identify optimal crosslinker amounts.
As a final note, it is important recognize that many alternative methods exist for making star shaped polymers (core-first, arm-first, etc.)25, 26. Each method has disadvantages and advantages, such as limits to size, purification requirements, and functional group compatibility. We argue that the broad functional group tolerance of ROMP, the ease of synthesis of norbornene-based functional monomers, and the ability to perform ROMP reactions on the benchtop rapidly, in parallel, and at room temperature, make the brush-first ROMP approach worth consideration for a variety of applications. In the future, we will continue to develop this method and BASP nanoarchitectures for various applications including drug and gene delivery, cellular imaging, and self-assembly. The full potential of these novel particles, and their capacity for combinatorial synthesis, has yet to be explored.
The authors have nothing to disclose.
We thank the MIT Department of Chemistry and the MIT Lincoln Labs Advanced Concepts Committee for support of this work.
Grubbs Second Generation Catalyst | Materia (or Sigma Aldrich) | C848 (Sigma Aldrich: 569747) |
Used as purchased from manufacturer. *Provided as a generous gift. |
Pyridine | Sigma Aldrich | 270970 | Used as purchased from manufacturer |
O-(2-aminoethyl)polyethylene glycol 3000 | Sigma Aldrich | 07969 | Used as purchased from manufacturer |
PEG-MM | N/A | N/A | Synthesized following reported procedures (Ref. 21, protocol 1) |
norbornene-N-hydroxysuccinimidyl (NHS) ester | N/A | N/A | Synthesized following reported procedures (Ref. 21) |
Bis-norb-NBOC Crosslinker | N/A | N/A | Synthesized following reported procedures (Ref. 21) |
Pentane | Sigma Aldrich | 158941 | Used as purchased from manufacturer |
Tetrahydrofuran (HPLC grade) | Sigma Aldrich | 34865 | Dried and purified over a solvent purification columns |
Dichloromethane | VWR | BDH1113-4LG | Used as purchased from manufacturer |
Acetonitrile (HPLC grade) | Sigma Aldrich | 34998 | Used as purchased from manufacturer |
Acetic Acid | Sigma Aldrich | A6283 | Used as purchased from manufacturer |
Sodium sulfate | Sigma Aldrich | 239313 | Used as purchased from manufacturer |
Diethyl ether | Sigma Aldrich | 673811 | Used as purchased from manufacturer |
Dimethylformamide (HPLC grade) | Sigma Aldrich | 270547 | Used as purchased from manufacturer |
Lithium Bromide | Sigma Aldrich | 213225 | Used as purchased from manufacturer |
MillQ Biocel A10 | Millipore | ||
Beckmann Coulter HPLC (127p solvent module, 166p detector) | Beckmann Coulter | ||
Zorbax 300SB-C18 PrepHT reverse phase column | Agilent | ||
1260 Infinity Liquid Chromatography | Agilent | ||
GPC KD-806M column | Shodex | ||
Dawn Heleos II Light Scatterer | Wyatt | ||
Optilab T-rEX Refractive Index Detector | Wyatt | ||
Glass Scintillation Vials – 40 ml | Chemglass | CG-4909-05 | |
Glass Scintillation Vials – 4 ml | Chemglass | CG-4904-06 | |
Glass Scintillation Vials (PTFE-lined cap) – 2 ml | Agilent | 5183-4518 | |
Stir-bars | VWR | 5894x | various sizes |
13 mm 0.45 µm Nylon Syringe filter | PerkinElmer | 02542903 | |
13 mm 0.45 µm polytetrafluoroethylene syringe filter | PerkinElmer | 02542909 | |
1 ml disposable syringes | VWR | 53548-001 | |
Swing bucket centrifuge or similar | Should be able to reach approximately 4,000 rpm | ||
Round bottom flask | |||
Fritted glass filter assembly | |||
Rotary Evaporator | |||
Balance |