doi.org/10.26434/chemrxiv.8943251.v1

Loading Dependent Structural Model of Polymeric Micelles bySolid-State NMRAnn-Christin Pöppler, Michael M Lübtow, Jonas Schlauersbach, Johannes Wiest, Lorenz Meinel, RobertLuxenhofer

Submitted date: 17/07/2019 • Posted date: 17/07/2019Licence: CC BY-NC-ND 4.0Citation information: Pöppler, Ann-Christin; Lübtow, Michael M; Schlauersbach, Jonas; Wiest, Johannes;Meinel, Lorenz; Luxenhofer, Robert (2019): Loading Dependent Structural Model of Polymeric Micelles bySolid-State NMR. ChemRxiv. Preprint.

Physically loaded polymeric micelles - from bulk properties to a molecular level understanding: In this work,solid-state NMR complemented by PXRD and quantum chemical calculations is used to learn about the shortrange order and structural arrangement in polymeric micelles formed by amphiphilic triblock copolymers.From changes in chemical shift and line widths, we could observe that at higher loadings, not just the micellarcore but also the hydrophilic corona is involved in coordination of poorly soluble guest molecules, which canexplain the poorer dissolution properties. This can serve as a platform for the future, targeted modification ofsuch polymers.

File list (1)

download fileview on ChemRxivLoading dependent Structural Model of Polymeric Micelles ... (2.47 MiB)

Loading dependent Structural Model of Polymeric Micelles by

Solid-State NMR

Ann-Christin Pöppler,* Michael M. Lübtow, Jonas Schlauersbach, Johannes Wiest, Lorenz Meinel, Robert

Luxenhofer

* to whom correspondence is to be addressed:

Jun.-Prof. Ann-Christin Pöppler

Institute of Organic Chemistry, University of Würzburg

Am Hubland, 97074 Würzburg, Germany

Email.: ann-christin.poeppler@uni-wuerzburg.de

Abstract: Detailed insight into the internal structure of drug loaded polymeric micelles is scarce, but

important for developing optimized delivery systems. We observed that an increase in curcumin loading

of triblock copolymers based on poly(2-oxazolines) and poly(2-oxazines) results in poorer dissolution

properties. Using solid-state NMR and complementary tools we can propose a loading dependent

structural model on the molecular level providing an explanation for these pronounced differences.

Changes in the chemical shifts and the respective line widths give evidence for the involvement of the

hydrophobic polymer block in the curcumin coordination for low loadings, while for higher loadings an

increase in the interaction with the hydrophilic polymer blocks is observed. The involvement of the

hydrophilic compartment may be critical for ultra-high loaded polymer micelles and can help to rationalize

specific polymer modifications to improve the performance of similar drug delivery systems.

Only if molecular level understanding of polymer drug formulations is available, it is possible to make

targeted changes to the components with the aim to optimize physico-chemical properties. Ideally, drug

delivery platforms should carry large amounts of cargo, while simultaneously maintaining suitable stability

and efficient release. In practice, drug formulations such as solid dispersions or soluble drug delivery

systems (DDS) comprising polymeric micelles have attracted a lot of attention[1] and several made their

way onto the market in the form of drug delivery platforms, health care products and biomaterials.[2]

However, the large body of published reports on DDS is not mirrored by therapeutic advances and benefit

to the patient.[3] Complexity and reproducibility are important points to be considered for polymers in

nanomedicine.[4] A molecular level understanding of these combined “macromolecule/small molecule”

materials that would help to systematically address these points is difficult to obtain due to their complex

nature and the lack of long-range order. Therefore, the prevailing picture across literature for self-

assembled polymeric micelles encapsulating drug molecules is that of a well-defined core-shell particle,

where the hydrophobic core contains the drug molecules and the hydrophilic shell of the particle has a

protection and solubilisation function.[5] This image is dominated by the look from the outside. In this

context, the work of Callari et al. was inspirational, because they were amongst the first to analyse how

the structure of a polymer micelle might be affected by the presence of drug molecules.[6] Their study

shows that a higher loading reduces the cellular uptake and cytotoxicity in vitro, attributed to an increased

packing density of the particle. In this case, a glycopolymer with discrete anchoring points for the Pt-drug

was used. However, the majority of DDS is relying on physical encapsulation featuring less well-defined

interactions. For such physically loaded micelles, this effect – a decrease in dissolution with increasing drug

loading – is also well known.[7] Here, we utilize solid-state NMR in combination with complementary tools

to generate detailed insights into ultra-high drug loaded polymeric micelles on the molecular level to

understand the interplay between drug loading and the bulk properties, e.g. dissolution behaviour, of such

formulations. We envisage this structure-property relationship to act as a basis for defined and systematic

modifications on the way to get the best from both worlds in future development – high loading and high

release.

To build a bridge from experimental

observations to a structural model, we used

an amphiphilic triblock copolymer poly(2-

methyl-2-oxazoline)-block-poly(2-n-propyl-2-

oxazine)-block-poly(2-methyl-2-oxazoline)

(pMeOx-b-pPrOzi-b-pMeOx = A-pPrOzi-A = P)

(Scheme 1). Due to the weak hydrophobic

character of pPrOzi, the polymer chains self-

assemble into micelles in aqueous solution

only in the presence of hydrophobic guest

molecules.[8] The natural product curcumin

(CUR) is encapsulated as a model compound

due to its very low aqueous solubility,

straightforward spectroscopic detection at

very low concentrations and the absence of

signal overlap with the polymer in the NMR spectra. Additionally, A-pPrOzi-A (P) can incorporate large

quantities of curcumin (>50 wt.%) enabling the preparation of formulations with different, well-defined

loadings.[9] A set of three different formulations, CUR-2-P, CUR-6-P and CUR-11-P, was prepared.[9-10] This

corresponds to 2, 6 or 11 g/L CUR per 10 g/L of the polymer before freeze-drying (Chapter S1 of the ESI).

For the highest loading, the number of curcumin molecules per polymer chain is exceeding the number of

repeating units of the inner, more hydrophobic polymer block.

Table 1. Experimentally determined dissolution rates in μmol/(min*cm²) with

lag time as well as results from hydrophobicity testing.

sample lag time

[min]

dissolution rate

[μmol/(min*cm²)] [a]

water uptake

(wt.%) at 80% RH[b]

CUR-2-P 0 6.2 ± 0.5[d] 30

CUR-6-P 24 2.6 ± 0.8[d] 16

CUR-11-P N/A 0.025 ± 0.008[e] 15

amorphous

CUR N/A 0.001 ± 0.0004[e] -

[a] mean ± SD (n = 3). [b] n = 1.

In standard pharmacopeia dissolution tests no detectable dissolution was observed for crystalline

curcumin within two hours, while amorphous CUR exhibited a low but noticeable dissolution (Table 1 and

Scheme 1. Structural formula of the components used in this

study: The amphiphilic block copolymer P encapsulates

curcumin by self-assembly into polymeric micelles

(schematic drawing on the right).

Figure S2). Formulating CUR with A-pPrOzi-A improved the dissolution rate up to 6000-fold (CUR-2-P) with

the dissolution behaviour strongly depending on the drug loading (Table 1 and Figure S3).

CUR-2-P with the lowest CUR loading immediately began to dissolve after the start of the experiment with

a dissolution rate of 6.19 μmol/(min*cm²), which was the highest among the tested samples. Upon

increasing the CUR loading (CUR-6-P), the dissolution rate drops by a factor of two and a lag period of 24

min is observed. This change is even more pronounced for the highest CUR loading (CUR-11-P) with the

dissolution rate decreasing by a factor of 100. Although this dissolution rate is still a factor of 20 faster

than pure, amorphous CUR, it shows that the highest loading of a compound may not always be the most

desirable delivery form for (oral) administration. A two-sided t-test confirmed that the observed

differences between CUR-2-P and CUR-6-P as well as between CUR-11-P and amorphous CUR are

statistically significant (p ≤ 0.05). To explain the differences in the dissolution behaviour, crystallization

and pH effects could be discarded based on PXRD and NMR, respectively. The high stirring speed (4800

rpm) used here also suffices to exclude a resident water layer as diffusion barrier. The dissolution studies

were supplemented by water uptake experiments at 80% relative humidity. After 24 h, the formulations

with lowest loading showed a 30% weight gain, while weight gains of 16% and 15% were observed for

CUR-6-P and CUR-11-P, respectively, illustrating an increasing hydrophobicity for the latter.

To derive a structural model, which can explain

these experimentally observed bulk properties,

first the freshly prepared and freeze-dried

formulations were characterized by powder X-ray

diffraction (Figure 1a). For neat CUR, a distinct

pattern of clear diffraction peaks was observed in

agreement with the most stable, monoclinic crystal

form (CSD ref code: BINMEQ05[11]). In contrast, all

formulations lacked long-range order; they were X-

ray amorphous. The typical amorphous halo is

generally similar in all three formulations, but

trends are observed, as the diffraction peaks

become broader and shift to higher 2θ values from

low (yellow) to high loading (red). As the PRXD

patterns depend on sample composition and

arrangement of the individual components,[12] this

indicates changes in local order. NMR spectroscopy

is particularly sensitive to the local environment of

NMR active nuclei and can thus be a powerful

probe of such local or short-range order

phenomena. Therefore, the samples were

dissolved in the non-selective solvent CDCl3 (no

micelles, individual components) or selective D2O

(CUR loaded micelles) and subsequently analysed

by NMR in solution (Figure S4 and Figure 1b). While all components of the formulation were clearly

observed in the spectra obtained from the CDCl3 solution, the CUR signals could barely be observed in D2O.

Only the polymer resonances were clearly distinguishable, suggesting that CUR in the micellar core

behaves more solid-like, which agrees with recent fluorescence spectroscopic analysis at very low

Figure 1. a) PXRD patterns recorded for as-received, crystalline curcumin and the formulations, CUR-2-P, CUR-6-P and CUR-11-P. b) Variable temperature 1H NMR spectra of CUR-2-P in D2O at 14.1 T.

loadings.[13] Increasing the mobility of the overall system by increasing the temperature resulted in

narrower linewidths for the CUR signals, but they remained very broad features (Figure 1b, inset). This

hampers the detailed analysis of proximities and intermolecular interactions. Diffusion (DOSY) NMR

measurements of the three formulations in solution yielded diffusion coefficients and thus approximated

radii, which agree very well with previously determined values from dynamic light scattering (DLS, Table

S2).[9] This shows that information on the size and exterior of these micelles is readily available, while

information on the molecular arrangement of and within the micellar core is more difficult to obtain.

Therefore, we turned our attention to

solid-state NMR, which has been

shown to be a very powerful analytical

technique in the pharmaceutical

context studying amorphous dosage

forms in general[14] and which is

particularly sensitive to intermolecular

interactions and subtle change in the

local arrangement. Freeze-dried

formulations were subjected to 1H and 13C CP/MAS NMR experiments at

24 kHz MAS and 14.1 T. To ensure

sample stability upon MAS, the 1H NMR

was observed at different times and

PXRD was measured after completion

of the NMR experiments (Figures S8b).

For a first proof-of-principle, the NMR

data from the three formulations were

compared to spectra for the individual

components (pure CUR and pure

polymer), a 1:1 physical mixture and

quench-cooled amorphous curcumin

(Figure 2, scaled according to individual

number of scans). The 13C CP/MAS NMR spectra of pure, crystalline CUR and the polymer used for sample

preparation are shown in grey and blue. The resonances were assigned based on NMR in solution and

additional GIPAW (CASTEP)[15] calculations for curcumin (Figure S12), which agree well with literature.[16]

The spectrum of the 1:1 physical mixture (green) is the simple sum of the spectra of the individual

components. In contrast, the spectra of the three formulations (yellow-orange-red) clearly differ from the

other spectra in peak number, position, width and relative intensities. Increasing curcumin loading results

in a corresponding increase in the relative CUR signal intensities and increasing resemblance with the

spectrum of quench-cooled amorphous CUR. Together, this shows that solid-state NMR provides a solid

ground to systematically analyse these polymer-drug self-assemblies, allowing us a detailed look at

distinct, loading dependent changes. In the following, several regions in the spectra (Figure 2, grey

highlights) will be discussed in more detail to point out specific aspects of the formulations (Figures 3 and

4). For this discussion, the scaling of the spectra was adjusted for each signal area to show signals with

practically equal heights. This facilitates the identification of changes in chemical shift and the respective

line widths. For an overview over all extracted spectral parameters, the reader is referred to the SI (Table

S4). As a general trend, the signals in the formulations attributed to the polymer are all broader than those

Figure 2. 13C CP/MAS NMR spectra of amorphous CUR (black), the

three formulations (yellow, orange, red with increasing CUR

concentration), pure polymer (blue), crystalline CUR (grey) and a 1:1

mixture of the two components (green). All spectra were recorded at

14.1 T and 24 kHz. Scaling is done according to the number of scans

of the individual datasets.

observed for the pure compound, while those resulting from CUR are narrower than fully amorphous CUR.

This corroborates the concept of an X-ray amorphous material with different degrees of short-range order.

Firstly, the signals corresponding to the quaternary C-OH/C-OMe carbon atoms at ~150 ppm as well as the

signals of the CH3 groups of curcumin (~ 56 ppm) are compared as they appear to be affected most by

increased loading (Figure 3, grey ellipsoids). These are the polar groups, which are expected to be involved

in intermolecular hydrogen bonding, while the enol moiety is mostly involved in intramolecular

interactions. This was also underlined by a comparison of theoretical chemical shifts in a full crystal and

for an isolated molecule

(Figures S13/S14). From low

(CUR-2-P) to high loading (CUR-

11-P), the corresponding 13C

chemical shifts decrease

approaching the one of

amorphous CUR and the line

widths increase. Hence, with

higher loadings, the curcumin

molecules experience a more

diverse molecular

environment. Interestingly, the

changes in absolute values are

larger between CUR-2-P and

CUR-6-P than between the two

higher loadings (6 vs. 11).

Secondly, the signals of CUR

between 110 and 130 ppm give

additional insights. These

signals belong to the backbone of the CUR molecule (highlighted in light blue). Curcumin in its most stable

crystalline form (monoclinic space group P2/n) does not show any 13C signals around 120 ppm (Figure

S12a) as indicated by the calculated chemical shifts (Figure 3, black sticks). In contrast, analogous

calculations for the two other known polymorphs, in which CUR shows a significantly smaller torsion angle

(16.1° vs. 46.0°), predict 13C resonances in this area of the spectrum (blue sticks and Figure S12a).

Therefore, we can deduce that CUR inside the micelles adopts a range of largely planar conformations.

Focussing on the polymer, the resonance of the amidic C=O group (~ 172 ppm) which can serve as

hydrogen bond acceptor site is clearly sensitive to the presence of CUR (Figure 4a). For the neat polymer,

only one resonance is observed in the 1D 13C CP/MAS NMR spectrum, so that pMeOx and pPrOzi cannot

be distinguished. However, in the 1H-13C HETCOR spectrum with long contact times (1.5 ms) three distinct

resonances at 172.8, 171.7 and 171.0 ppm are observed, which can be assigned to the pPrOzi carbonyl

moiety and the cis/trans isomers of pMeOx by comparison with 13C NMR data in solution (Figure 4b). The

existence of cis/trans isomers is well known for tertiary amides.[17] Upon CUR loading, the lines broaden

and overlapping resonances can be observed. This is due to the distribution of environment and the

presence of uncoordinated as well as coordinated amide moieties. With increasing loading, this trend

becomes more pronounced and two additional signals can be observed at lower ppm values for CUR-11-P

(arrows in Figure 4a, Figure 4c). This could be an indication that not just one type of amide, namely the

pPrOzi core, but also the amide groups of hydrophilic pMeOx polymer blocks are involved in the

Figure 3. Enlarged extracts from the overlay of the 13C CP/MAS NMR spectra of CUR-2-P (yellow), CUR-6-P (orange) and CUR-11-P (red) compared to amorphous CUR (black) as depicted in Figure 2. Here, the signals intensities were scaled to approximately equal height for each individual signal area to facilitate direct comparison. Calculated chemical shifts are shown as sticks.

coordination of CUR. The latter should not be strongly affected if the CUR molecules were only to be

located in and interact with the micellar core. To investigate this more closely, changes in chemical shift

and line widths for the CH3 group of the pMeOx polymer blocks are analysed (Figure 4d). The line width

increases significantly from the pure, unassembled polymer to self-assembled CUR-2-P and further to the

CUR-6-P and CUR-11-P formulations, which lie in the same range. The chemical shift shows only a minor

change from the pure polymer to the formulation with low loading (0.1 ppm) as expected from the

simplified image that CUR is located in the micellar core in the proximity of the more hydrophobic PrOzi

residues. With increasing CUR loading, however, a more pronounced change in chemical shift is observed.

This also indicates that for higher loadings, the hydrophilic polymer blocks become involved in the

coordination of CUR. In agreement with this picture, the chemical shift as well as the linewidth of the

propyl-CH3 moiety do not differ between the CUR-6- and CUR-11-P formulations, for which we hypothesize

that the interactions with the hydrophobic micellar core are largely saturated (Figure S7).

Figure 4. a) Extract from the 13C CP/MAS NMR spectra in Figure 2 with a different scaling showing the signals with approximately equal height. The signal area corresponds to the CO resonance of the amide moiety. b) Comparison of the chemical shift and line width of the CH3 group of the hydrophilic polymer block for the pure polymer and the three formulations. c) Amide region of the 1H-13C HETCOR spectrum of the pure polymer alongside 13C NMR spectra in solution as well as d) analogous extract from the HETCOR spectrum of CUR-11-P. Long contact times were used to obtain the HETCOR data.

Based on these observations, we can now assemble the puzzle pieces from the various experiments to an

overall loading dependent, molecular level structural model (Figure 5): The unimers self-assemble into

micelles only in the presence of curcumin. For the lowest concentration of CUR (CUR-2-P), the poorly

water-soluble, largely planar CUR molecules are mainly located in the micellar core with a distribution of

rather defined environments due to hydrogen bonding between the phenolic OH-moieties and the amide

moiety of the hydrophobic block (depicted in red, Figure 5). As these preferred interaction sites become

saturated with increasing CUR loading (CUR-6-P), the local order of CUR molecules is reduced.

Furthermore, CUR molecules are now envisaged to be located at the hydrophilic-hydrophobic interface (in

agreement with the raspberry model fitted for SANS data of a comparable system[18]) and therefore

interact increasingly with the repeating units of the hydrophilic block as evidenced by distinct changes in

chemical shift for the pMeOx carbonyl carbon and CH3 group. These amide functionalities of the

hydrophilic blocks (Figure 5, in blue) otherwise hydrate and stabilize the micelle. Consequently, in the

hydration/dissolution process, CUR molecules now block the hydrogen bond acceptor sites within the shell

of the micelle. This correlates with the observed retardation of the dissolution for CUR-6- and CUR-11-P.

Finally, for the highest loading, not only the core is essentially filled with CUR molecules (smallest degree

of short-range order in the NMR), also the hydrophilic amide units become saturated, essentially physically

crosslinking the hydrophilic corona, significantly affecting hydration. This is in accordance with the strongly

decreased dissolution rate as observed for CUR-11-P1. This overall picture would suggest an increased

hydrophobicity of the micelles at higher loadings in agreement with the results in Table 1 and with a

different study on glycopolymers using SAXS and SANS.[19]

Figure 5. Schematic model of the structural changes of the polymeric micelles upon loading with curcumin based on the solid-state NMR data and complementary insights. For each loading stage, the additionally occurring interaction site is depicted.

In general, structural information and understanding on the molecular level is difficult to obtain for drug

loaded polymeric micelles. In this work, we could show that solid-state NMR is a versatile toolbox, which

can – complemented by other approaches – facilitate the detailed analysis of polymeric micelles improving

our structural understanding. NMR data probed changes in the interaction profile with increasing loading.

At higher loadings, the hydrophilic polymer blocks were found to participate in the coordination of CUR,

which is presumably of critical importance to obtain such unusually high drug loadings of approx. 50 wt.%

observed with this platform. This can now serve as a starting point for the rational modification of

polymers to maintain ultra-high loadings, while not having to compromise the release behaviour.

Experimental studies will now need to verify if and how optimized nanoparticles for drug delivery can be

obtained through this approach. Additionally, the NMR experiments used here can be expanded by high-

resolution proton-detected correlations at fast MAS, which should be even more sensitive to the

underlying key intermolecular interactions. This would allow the drawing of an even more accurate picture

for this and other drug-polymer arrangements. A systematic investigation by varying both polymer and

guest molecules potentially also employing isotopic labelling schemes is necessary to, for example, explain

extremely different loading efficiencies observed for structurally similar polymers.[9, 20] Moreover,

exploration of the complementary insights from pair-distribution functions based on PXRD or probing the

internal micellar structure by SANS should also be very valuable to learn more about these interesting

systems.

Acknowledgements

We thank Prof. Roland Mitric for access to his workgroup cluster for the GIAPW (CASTEP) calculations and

in particular to Matthias Wohlgemuth for helping us with the setup of the program package. We thank Dr.

Matthias Grüne for helpful and stimulating discussions. We further thank Dominik Heuler for his support

with the PXRD measurements and Malik Salman Haider for sharing his experience with respect to quench-

cooling of curcumin. Marvin Grüne and Tessa Lühmann are acknowledged for discussion and proof

reading. Financial support from the Fonds der Chemischen Industrie (FCI) in form of a material cost

allowance is acknowledged (A.-C. P.). M. M. L thanks the Evonik Foundation for providing a doctoral

fellowship. Both experimental NMR and calculated data for this study will be available through archivexxx.

Keywords: Solid-state NMR • Micelles • Polymers • Dissolution rates • Short-range order

[1] a) R. Duncan, Nat. Rev. Drug Discov. 2003, 2, 347-360; b) Y. Huang, W.-G. Dai, Acta Pharm. Sin. B.

2014, 4, 18-25; c) G. S. Kwon, K. Kataoka, Adv. Drug Delivery Rev. 2012, 64, 237-245; d) H. Cabral,

K. Kataoka, J. Control. Release 2014, 190, 465-476.

[2] a) V. Wagner, A. Dullaart, A.-K. Bock, A. Zweck, Nat Biotech 2006, 24, 1211-1217; b) N. Kamaly, Z.

Xiao, P. M. Valencia, A. F. Radovic-Moreno, O. C. Farokhzad, Chem. Soc. Rev. 2012, 41, 2971-3010.

[3] J.-C. Leroux, Angew. Chem. Int. Ed. 2017, 56, 15170-15171.

[4] R. Luxenhofer, Nanomedicine 2015, 10, 3109-3119.

[5] a) H. Chen, S. Kim, L. Li, S. Wang, K. Park, J.-X. Cheng, Proc. Natl. Acad. Sci. 2008, 105, 6596-6601;

b) H. Cho, T. C. Lai, K. Tomoda, G. S. Kwon, AAPS PharmSciTech 2014, 16, 10-20; c) Y. T. Tam, J.

Gao, G. S. Kwon, J. Am. Chem. Soc. 2016, 138, 8674-8677; d) H. Cabral, K. Miyata, K. Osada, K.

Kataoka, Chem. Rev. 2018, 118, 6844-6892.

[6] M. Callari, P. L. De Souza, A. Rawal, M. H. Stenzel, Angew. Chem. Int. Ed. Engl. 2017, 56, 8441-8445.

[7] a) R. Gref, Y. Minamitake, M. T. Peracchia, V. Trubetskoy, V. Torchilin, R. Langer, Science 1994, 263,

1600; b) M. Polakovič, T. Görner, R. Gref, E. Dellacherie, J. Control. Release 1999, 60, 169-177.

[8] M. M. Lübtow, L. C. Nelke, J. Seifert, J. Kühnemundt, G. Sahay, G. Dandekar, S. L. Nietzer, R.

Luxenhofer, J. Control. Release 2019, 303, 162-180.

[9] M. M. Lübtow, L. Hahn, M. S. Haider, R. Luxenhofer, J. Am. Chem. Soc. 2017, 139, 10980-10983.

[10] R. Luxenhofer, A. Schulz, C. Roques, S. Li, T. K. Bronich, E. V. Batrakova, R. Jordan, A. V. Kabanov,

Biomaterials 2010, 31, 4972-4979.

[11] P. Sanphui, N. R. Goud, U. B. R. Khandavilli, S. Bhanoth, A. Nangia, Chem. Commun. 2011, 47, 5013-

5015.

[12] S. Bates, G. Zografi, D. Engers, K. Morris, K. Crowley, A. Newman, Pharm. Res. 2006, 23, 2333-2349.

[13] M. M. Lübtow, H. Marciniak, A. Schmiedel, M. Roos, C. Lambert, R. Luxenhofer, Chemistry – A

European Journal, just accepted.

[14] a) R. Lefort, A. De Gusseme, J. F. Willart, F. Danède, M. Descamps, Int. J. Pharm. 2004, 280, 209-

219; b) T. N. Pham, S. A. Watson, A. J. Edwards, M. Chavda, J. S. Clawson, M. Strohmeier, F. G.

Vogt, Mol. Pharm. 2010, 7, 1667-1691; c) J. Brus, M. Urbanova, I. Sedenkova, H. Brusova, Int. J.

Pharm. 2011, 409, 62-74.

[15] a) C. J. Pickard, F. Mauri, Phys. Rev. B 2001, 63, 245101; b) S. J. Clark, M. D. Segall, C. J. Pickard, P.

J. Hasnip, M. I. J. Probert, K. Refson, M. C. Payne, Z. Kristallogr. 2005, 220, 567; c) J. R. Yates, C. J.

Pickard, F. Mauri, Phys. Rev. B 2007, 76, 024401.

[16] X. Kong, A. Brinkmann, V. Terskikh, R. E. Wasylishen, G. M. Bernard, Z. Duan, Q. Wu, G. Wu, J. Phys.

Chem. B 2016, 120, 11692-11704.

[17] a) M. Sisido, Y. Imanishi, T. Higashimura, Biopolymers 1972, 11, 399-408; b) Q. Sui, D. Borchardt,

D. L. Rabenstein, J. Am. Chem. Soc. 2007, 129, 12042-12048; c) B. C. Gorske, J. R. Stringer, B. L.

Bastian, S. A. Fowler, H. E. Blackwell, J. Am. Chem. Soc. 2009, 131, 16555-16567.

[18] A. Schulz, S. Jaksch, R. Schubel, E. Wegener, Z. Di, Y. Han, A. Meister, J. Kressler, A. V. Kabanov, R.

Luxenhofer, C. M. Papadakis, R. Jordan, ACS Nano 2014, 8, 2686-2696.

[19] C. Cao, J. Zhao, F. Chen, M. Lu, Y. Y. Khine, A. Macmillan, C. J. Garvey, M. H. Stenzel, Chem. Mater.

2018, 30, 5227-5236.

[20] M. M. Lübtow, L. Keßler, A. Appelt-Menzel, T. Lorson, N. Gangloff, M. Kirsch, S. Dahms, R.

Luxenhofer, Macromol. Biosci. 2018, 18, 1800155.

S1

Supporting Information

Loading dependent Structural Model of Polymeric Micelles

by Solid-State NMR

Ann-Christin Pöppler,* Michael M. Lübtow, Jonas Schlauersbach, Johannes Wiest, Lorenz Meinel, Robert Luxenhofer

* to whom correspondence is to be addressed:

Jun.-Prof. Ann-Christin Pöppler

Institute of Organic Chemistry, University of Würzburg

Am Hubland, 97074 Würzburg, Germany

Email.: ann-christin.poeppler@uni-wuerzburg.de

Content:

S1. Preparation of the CUR-Polymer Formulations S2

S2. Dissolution Tests S2

S3. Hygroscopicity S5

S4. Characterization of the Formulations by NMR in Solution S6

S5. Characterization by Multinuclear solid-state NMR S8

S6. Characterization of the Formulations by Powder X-ray diffraction S11

S7. CASTEP Calculations S13

S8. References S17

S2

S1. Preparation of the CUR-Polymer Formulations

Curcumin (CUR) powder from Curcuma longa (Tumeric) was purchased from Sigma-Aldrich and used

as received. The polymer, poly(2-methyl-2-oxazoline)-block-poly(2-n-propyl-2-oxazine)-block-poly(2-

methyl-2-oxazoline) (PMeOx-b-PPrOzi-b-PMeOx ≡ A-pPrOzi-A ≡ P), was synthesized via living cationic

ring-opening polymerization and characterized as described in detail in the literature.[1] This

amphiphilic block-co-polymer forms micelles in the presence of guests in aqueous solution. The

formulations with CUR were prepared using the thin film method as described in the same and

previous work of this group.[1-2] Three different formulations with defined CUR loadings were prepared

(Table S1).

Table S1: Summary of the formulations used in this work and how they are referred to in the manuscript. The polymer A-pPrOzi-A was used to encapsulate Curcumin in its more hydrophobic core resulting in polymer micelles with a defined loading.

Amount of loaded CUR per 10 g/L polymer Name of Formulation Loading in wt%

2 g/L CUR-2-P 16.7

6 g/L CUR-6-P 37.5

11 g/L CUR-11-P 52.5

For the highest loading presented in this table, almost 25 molecules of CUR were incorporated, which

is more than there are repeating units in the more hydrophobic inner polymer block.

S2. Dissolution tests

Dissolution rates were measured with a Sirius T3 instrument (Sirius Analytical, Forest Row, UK) as

described earlier.[3] Tablets discs (diameter 0.07 cm2 and provided by the manufacturer of the

machine) with defined surfaces were prepared by compression of 3–10 mg of each sample under a

weight of 0.18 tons for 5 min with a manual hydraulic tablet press (Paul Weber, Stuttgart, Germany).

The release of drug substance from the tablet discs allows data collection with a standardized surface

area (0.07 cm2) required to calculate dissolution rates.[3] Dissolution rates were determined

photometrically at room temperature in simulated intestinal fluid (SIF) pH 6.8 (USP 26), for which the

ionic strength was adjusted to 0.15 M with potassium chloride, at a stirring speed of 4800 rpm

following manufacturer’s instructions. The amount of dissolved CUR was calculated by the Beer-

Lambert law using the spectroscopic data obtained by a fibre optic dip probe connected to a diode

array detector. The linear part of the release profile was used for calculation of the dissolution rate

(dissolved substance per time and surface area). In total, five samples were measured: crystalline and

amorphous curcumin as well as three formulations. Each experiment was repeated three times. A

two-sided t-test with a significance level of p ≤ 0.05 was successfully performed for CUR-2-P and CUR-

6-P as well as for CUR-11-P and amorphous CUR.

S2.1 pKa determination of Curcumin

A photometrical pKa assay as described[4] is necessary before performing the dissolution tests. The

curcumin concentration was calculated with the mean molar extinction coefficients of the differently

S3

charged curcumin species derived from this assay. The pKa values of curcumin (Figure S1) were in

accordance to literature 8.03 (± 0.14), 8.16 (± 0.03) and 10.16 (± 0.04) respectively. As the used buffer

in the dissolution assay has a pH of 6.8, the predominant curcumin species in solution is H3A

(uncharged).

Figure S1: pH dependent distribution of Curcumin species. Blue H3A (uncharged), red H2A-, green HA2- and black A3-

S6.2 Pure crystalline and amorphous Curcumin

a)

b)

Figure S2: Exemplary dissolution profile for a) crystalline and b) amorphous curcumin according to the protocol described

above. For crystalline curcumin no dissolution could be detected within the duration of the experiment, because the

concentration was below the limit of detection (only signal noise).

For crystalline curcumin, the concentration was below the limit of detection (Figure S2a), which

prevented the determination of the dissolution rate.

In the case of amorphous curcumin (obtained via quench-cooling), a higher dissolution rate could be

expected. Exemplarily for the three repetition experiments, this is observed in Figure S2b. From the

three experiments, an average extrapolated dissolution rate of 1.01 nmol/min*cm² can be

determined. The individual values 0.78, 1.51 and 0.75 nmol/min*cm² show that the deviation is quite

large. Residual small particles on the tablet disc can already interfere with the measurements for

samples with such low amounts of detectable substance.

S4

S2.3 Formulations CUR-2/6/11-P

For the three formulations, dissolution profiles with varying onset could be recorded for all samples

(Figure S3).

a)

b)

c)

Figure S3: Dissolution profiles for the three samples a) CUR-2-P, b) CUR-6-P and c) CUR-11-P recorded for a pressed tablet at pH 6.8 (n =3). While the y-axis is displayed in μmol for a) and b), the axis for CUR-11-P in c) is given in nmol.

S5

S3. Hygroscopicity

Storage of part of the sample under ambient conditions already indicated that the three formulations

show different degrees of hygroscopicity (Figure S4a). Therefore, the hygroscopicity of the three

formulations was measured gravimetrically as explained in Ph. Eur: The samples inside a glas vial are

placed in a closed desiccator containing a saturated solution of ammonium chloride. Thus, a fixed

relative humidity of 80% at 25°C is achieved.[5] After 24 h, the samples were removed from the

desiccator and carefully weighed. From this, the weight gain [%] was determine for each sample.

Figure S4b shows pictures of the three samples before and after being stored in the desiccator. The

data is summarized in Table S2. Comparably small sample volumes were used. However, a general

trend is distinguishable, and the difference observed between CUR-2-P and CUR-6-P is very

pronounced, which agrees with dissolution data and our structural model.

a)

b)

Figure S4: a) Set of samples after storage at ambient conditions. b) The three samples before (top) and after (bottom) storage for 24 h at about 80% relative humidity.

Table S2: Experimental results from the hygroscopicity test according to Ph. Eur. Weighing of the vials was carried out with a high-precision balance. The temperature was 24 °C and thus slightly lower than the temperature given in the Ph. Eur. A saturated solution of ammonium chloride was used to generate the 80% RH atmosphere in the desiccator.

empty glass [mg]

filled glass [mg]

after 24h [mg]

Sample [mg]

Hygroscopicity [%]

CUR-2-P 7675.3 7682.9 7685.2 7.6 30

CUR-6-P 7685.7 7694.5 7695.9 8.8 16

CUR-11-P 7710.7 7719.6 7720.9 8.9 15

S6

S4. Characterization of the Formulations by NMR in solution

All NMR experiments were performed on a 14.1 T standard bore Bruker Avance III instrument at room

temperature. A 5 mm BBFO probe equipped with z-gradient and a temperature unit was used.

NMR in solution using CDCl3 (readily dissolves polymer and CUR) and D2O (micelles present) was used

for a first assignment of the 1H and 13C chemical shifts for the different components.

Figure S5: 1H and 13C NMR spectra of the polymer in CDCl3 (blue) and D2O (red) recorded on a 600 MHz spectrometer. The chemical formula is shown on top of the proton data and assignment of the resonances is indicated by either colour or text fragments.

In the case of as received curcumin dissolved in CDCl3, we saw the presence of the enol form only as

previously shown.[6] The 1H as well as 13C NMR spectra of the pure polymer in solution (no micelles)

including assignment of the chemical shifts to the respective fragments in the polymer is shown in

Figure S5. For the assignment a standard set of 1D and 2D NMR data (1H, 13C, 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, 1H-1H NOESY) was used.

DOSY

DOSY NMR data were measured at 298 K with a Bruker Avance III HD 600 MHz spectrometer and a 5

mm BBFO probe containing a z-axis gradient coil with a maximum gradient strength of 50 G cm−1.To

avoid the influence of convection effects on the results (due to temperature gradients in the coil of

the probe), two different pulse sequences, a stimulated echo BPP-LED pulse sequence[7] with sample

rotation and a corresponding double stimulated echo pulse sequence[8] without sample rotation were

used. No differences in diffusion data obtained from the two pulse sequences were observed. For all

experiments, the diffusion gradients were linearly incremented in 16 steps from 2 to 98%. Diffusion

times Δ of 100 ms and durations of a bipolar gradient pulse δ of 4 ms were optimized and then used

for all further experiments. All signals had decayed to below 5% of the initial signal area. The diffusion

coefficients were obtained by fitting of the signal area versus the gradient strength with two

exponentials using the Bruker software Topspin 3.5. As starting values for the fitting routine of the

diffusion coefficient 8·10-12 and 9·10-11 m²/s were chosen. The reported diffusion coefficients were

obtained from the measurements with the BPP-LED pulse sequence and converted to approximate

radii by using the viscosity of D2O at 298 K.[9] To extract the values shown in Table S3, the CH3(CO)

S7

resonance of the hydrophilic polymer block at 1.98 ppm was used as this functional group is located

in the shell of the micelles and the signal is clearly separated.

Table S3: Diffusion coefficients and approximated radii of the three formulations with increasing drug loading obtained from 1H DOSY NMR data using two exponentials to fit the decay curves. The values shown were extracted for the CH3(CO) signal at 1.98 ppm. The radii are compared to diameters determined by dynamic light scattering (DLS) as found in literature.

Formulation Diffusion coefficients D

[m²/s]

Radii approximated from

D [nm]

Diameter obtained by DLS

[nm] and (drug loading)[1]

CUR-2-P

(10/2)

1.72·10-11

7.56·10-11

11.5

2.6 22.9 (10/2)

CUR-6-P

(10/6)

1.62·10-11

8.95·10-11

12.2

2.2 25.3 (10/5)

CUR-11-P

(10/11)

8.01·10-12

9.12·10-11

24.8

2.2 46.5 (10/11)

For all samples, diffusion coefficients in a similar range were determined. One of the two components

diffuses faster than the second species and thus has a smaller size. The size of this smaller particle was

determined to be 2.2-2.6 nm and possibly results from unimers (single polymer strands) present in

solution,[10] while the larger particles have varying radii depending on the CUR loading. With increasing

drug loading, the radius of the particle increases. Upon increasing the CUR loading, the radius of the

micelles formed in solution increases from 11.5 to 12.2 and 24.8 nm. This agrees very well with the

literature-known values observed by dynamic light scattering (DLS).[1] Alternative fitting procedures

based on the log normal or gamma model can also be used to account for the polydispersity of the

polymers.[11]

S8

S5. Characterization by Multinuclear NMR

The solid-state NMR measurements were performed using a 3.2 mm double-channel Bruker probe at

14.1 T and 24 kHz MAS. For the CP, a 2 ms ramp (50 to 100%) on the 1H channel was used during the

CP contact time for all samples. For heteronuclear decoupling during acquisition, SPINAL64 was

employed with a 100 kHz nutation frequency (1H). The chemical shifts were referenced using

adamantane (left signal at 38.48 ppm) by subsequent adjustment of the magnetic field.

For all following characterization by both NMR and calculations, the following numbering scheme is

used in accordance with previously published data on CUR (Figure S6a).[12]

a)

b)

Figure S6: a) Numbering scheme for the curcumin molecule as is used throughout this manuscript. H-atoms are labelled according to the carbon to which they are attached. b) 1H solid-state NMR spectra of the formulation CUR-2-P recorded at the beginning of the measurement block and after spinning the sample at 24 kHz for ~2.5 days.

In Figure S6b, the 1H solid-state NMR spectra of the formulation CUR-2-P are shown. Both spectra

were measured using the same conditions, yet at different times of the measurement block to ensure

that the sample is unchanged by the forces caused by rotating the sample at 24 kHz. At this MAS rate,

the signal overlap is substantial, which makes the interpretation of features and trends more difficult.

Higher MAS rates would be necessary to make use of powerful proton-detected solid-state NMR

experiments. Therefore, the focus in this manuscript is on the 13C NMR data.

S9

To look at the changes in the

polymer 13C NMR signals in more

detail, the respective area in the 13C

CP/MAS spectra of the pure polymer

P (blue) as well as the three

formulations is shown in Figure S7.

The OMe signal of CUR indicates the

increasing CUR loading of the

micelles. The corresponding amount

of polymer should be decreasing.

Interestingly, an increase in the

signal intensity is observed, which

can result from the less mobile

arrangement and/or a larger

number of proton-carbon contacts.

If we first focus on the CH3 group of

the propyl moiety (inner polymer

block), we see a trend similar to the

one observed for the amide group of

the outer polymer blocks. The chemical shift decreases with increasing loading with CUR-6-P and CUR-

11-P having very similar values, which is mirrored by the line widths (pure polymer: 96 Hz; CUR-2-P:

168 Hz; CUR-6-P: 269 Hz; CUR-11-P: 291 Hz). Again, the difference from pure polymer to the low and

medium loadings are larger than the difference between medium and high loading, for which we

propose that the core is already full and additional CUR needs to be located at the hydrophobic-

hydrophilic interface with the amide groups of the hydrophilic blocks participating in the coordination.

An indicator for the location of CUR at the interface could be the behaviour of the signal at 29 ppm,

which belongs to the inner CH2 group of the backbone. The signal broadens and for the two higher

loadings, the signal consists of two separate parts (~28 and 29 ppm).

A table summarizing all chemical shifts and (if possible to determine) linewidths for amorphous CUR,

the pure polymer and the three formulations CUR-2/6/11-P is shown in the following (Table S4)

Figure S7: Extract from the 13C CP/MAS NMR spectra of the formulations CUR-2/6/11-P as well as pure polymer for comparison. Only the signal areas of the aliphatic polymer signals and that of the OMe group of CUR are shown. Scaling was applied according to the individual number of scans.

S10

Tab

le S4: Su

mm

ary o

f the ch

emica

l shifts in

pp

m a

nd

line w

idth

s (if po

ssible a

s full w

idth

at h

alf m

axim

um

in H

z) determ

ined

from th

e 13C

C

P/M

AS N

MR

spectra

of th

e three fo

rmu

latio

ns C

UR

-2-P

, CU

R-6

-P a

nd

CU

R-1

1-P

, wh

ich a

re com

pa

red to

am

orp

ho

us cu

rcum

in a

s well a

s the

pu

re po

lymer. A

n in

crease co

mp

ared

to th

e pu

re com

po

un

d is in

dica

ted in

red, a

decrea

se in b

lue.

a) In

the 1

D d

ata

sets, wh

ich w

ere used

as b

asis fo

r this ta

ble, th

e pea

k po

sition

s an

d lin

e wid

ths o

f the ca

rbo

nyl reso

na

nce o

f the a

mid

e gro

up

a

re difficu

lt to d

etermin

e du

e to severe sig

na

l overla

p. P

lease see Fig

ure 4

in th

e ma

in text fo

r mo

re info

rma

tion

.

S11

S6. Characterization of the Formulations by Powder X-ray diffraction

Powder diffractometric studies were done with a Bruker Discover D8 powder diffractometer

(Karlsruhe, Germany) using Cu Kα radiation (unsplit Kα1+Kα2 doublet, mean wavelength λ = 154.19

pm) at a power of 40 kV and 40 mA, a focusing Goebel mirror and a 2.5° axial Soller slit. The scattered

X-ray beam went through a receiving slit (3,3°). Detection was done with a LynxEye-1D-Detector

(Bruker AXS) using the full detector range of 192 channels. Measurements were done in reflection

geometry in coupled two theta/theta mode with a step size of 0.025° in 2θ and 0.55 s (pure

compounds) or 0.71 s (formulations) measurement time per step in the range of 5–50°(2θ). Data

collection was done with the software package DIFFRAC.Suite (V2 2.2.690, BrukerAXS 2009–2011,

Karlsruhe, Germany). The diffraction data was subsequently converted into ASCII format and further

handled with Origin (OriginLab, Massachusetts, USA).

Directly after sample preparation as well as after the measurements undertaken (NMR, dissolution

tests), PXRD data was recorded to ensure that the samples did not change under the measurement

conditions (e.g. Magic Angle Spinning (MAS)).

S6.1 Pure Curcumin and the Effect of MAS

Curcumin was first measured by PXRD as purchased and thus as it was used for the preparation of the

formulations (Figure S8a). As expected for a microcrystalline compound, discrete peaks were detected

(black), which agree well with simulated data from the literature known stable monoclinic form of

CUR (BINMEQ05).[13] Additionally, the diffraction pattern for the quench-cooled, amorphous sample

is shown in red. The sample is clearly X-ray amorphous and only very broad features centred around

16°, 24° and 43° (2Θ) can be observed. In a second step, diffractograms of the CUR-6-P formulation

directly after preparation and after one day at 24 kHz MAS are exemplarily shown (Figure S8b). There

are no pronounced differences between the two samples, which confirms that the forces acting on

the sample under MAS do not alter the general arrangement of the CUR-polymer assembly.

a)

b)

Figure S8: a) Powder X-ray pattern of as received curcumin with the corresponding simulated data from the monoclinic form BINMEQ05 alongside the data for the quench-cooled amorphous form. b) Comparison of the PXRD patterns before and after MAS at 24 kHz.

S12

S6.2 PXRD to ensure sample stability for dissolution testing

Similar to the examination before and after MAS, the pressing of the samples into the tablets prior to

the dissolution tests exposes them to considerable pressure. Therefore, PXRD was recorded for the

fresh samples and compared to the same samples after tablet pressing. The corresponding PXRD

patterns for pure curcumin as well as for the CUR-6-P formulation are shown below (Figure S9).

In the case of pure, crystalline CUR (a), only changes in the peak intensities, but not in the peak

positions are observed. This confirms that the cell and atom positions remain unchanged by the tablet

pressing, but that preferred orientations of crystallites are influenced. For CUR-6-P representing the

set of formulations (b), the general feature in the diffractogram is maintained.

Finally, PXRD data was recorded for the swollen yet

not dissolved tablet, which was observed for the

highest CUR loading of the micelles. The data is

compared to that of the pressed CUR-11-P tablet

(Figure S10). The general appearance of the PXRD

pattern does not change due to the dissolution

experiment. Furthermore, no indication of

crystalline, well-defined peaks is found within the

detection limit of this experiment despite addition of

three subsequent measurements for the swollen

sample. This shows that the delay in the dissolution

event is not likely to be due to crystallization

processes at the interface.

a)

b)

Figure S9: PXRD patterns for pure, crystalline curcumin (a) and the formulation CUR-6-P (b) before and after tablet pressing.

Figure S10: PXRD patterns of the pressed and subsequently swollen tablet of CUR-11-P.

S13

S7. GIPAW(CASTEP) calculations

GIPAW (CASTEP) calculations[14] were used for three purposes: a) identification of the initial crystal

form (together with PXRD) and subsequent assignment of the CUR NMR resonances, b) determination

dominant interactions in the crystalline materials, c) information on the influence of the geometry of

the CUR molecules on the chemical shifts. In the following, a general information on the calculations

will be given (Chapters S7.1 and S7.2) before specific data will be discussed (Chapter S7.3).

S7.1 General Information on Version, Parameters and Conventions used

All calculations were run on the local cluster of the workgroup of Prof. Roland Mitric. They were

performed using the CASTEP code,[14a] academic release version 17.2. All structures were first

geometry optimized with the unit cell parameters fixed before chemical shieldings were calculated

using the GIPAW method.[14b, 14c] In all calculations, the PBE exchange correlation was used.[15] For the

plane wave basis set with ultra-soft pseudopotentials,[16] a maximum cut-off energy of 800 eV was

used alongside a Monkhorst–Pack grid for sampling over the Brillouin zone with minimum sample

spacing 0.1 × 2π Å−1. The forces, energies and displacements were converged to better than 0.05 eV

Å−1, 0.00002 eV, and 0.001 Å, respectively. To handle and visualize NMR output the Magresview

environment was employed.[17] For a crystal vs. molecule comparison, an additional NMR calculation

was performed for a single molecule from the optimized structure in an enlarged unit cell, which is

thus no longer interacting with neighbouring molecules.

In this work, the chemical shielding tensors obtained by GIPAW (CASTEP) calculations are presented

according to the Haeberlen-Mehring-Spiess convention:

The principal components of the chemical shielding tensor are σxx, σyy and σzz, satisfying |σzz ‒ σiso| ≥

|σxx ‒ σiso| ≥ |σyy ‒ σiso|. The isotropic chemical shielding is an average over the principal components:

σiso = (σxx+ σyy+ σzz)

3 and the anisotropy is defined as σaniso = σzz ‒ σiso. The asymmetry is obtained

according to ηasym =σyy−σxx

σansio, with ηasym = 0 corresponding to axial symmetry.

To compare the calculated chemicals shieldings to experimental chemical shift data, referencing

according to the following scheme is required: δiso = σref ‒ σiso.[18] For the discussed nuclei 1H and 13C,

the references values and their origins are given below:

Table S5: Summary of the nuclei, for which NMR parameters were calculated using the GIPAW (CASTEP) approach and the corresponding references values used to convert chemical shieldings into chemical shifts.

nucleus σref ref.

1H 30.5 (BINMEQ05)

13C 171.34 (BINMEQ05), 170.81

(BINMEQ06 and BINMEQ07)

exact value by plotting exp. chem.

shifts and calc. shieldings (fixed slope

at −1)[18-19]

S7.2 Crystal vs. Molecule

To gain additional theoretical insight into the set of intermolecular interactions that are present for a

particular form or packing arrangement, a second set of calculations was performed on a single

molecule. To do so, one molecule from the fully geometry optimized structure was kept in the unit

cell, which was also enlarged by ~5 Å in each direction. Thereby it was assured that this molecule is no

S14

longer in proximity to any neighbouring molecules. Subsequently, another set of NMR parameters was

calculated. Thus, the obtained chemical shifts were compared to the corresponding ones for the full

crystal structure. This difference = crystal ‒ molecule then represents the change induced by the sum

of all interactions the molecule is involved in.

S7.3 GIPAW (CASTEP) calculations of Curcumin

Three different polymorphs, one monoclinic and two orthorhombic structures, of curcumin are

known. The molecule exists in its keto-enol form. For the calculations in this project, we used the

structures reported by Sanphui et al. in 2011.[13] They are deposited in the Cambridge Structural

Database[20] under the following reference codes: CSD-BINMEQ05, CSD-BINMEQ06 and CSD-

BINMEQ07 (Figure S11). Apart from the crystal system/space group and thus also the number of

individual molecules in the asymmetric unit, the three structures also vary in the degree to which the

two halves of the curcumin molecule are twisted (46° in BINMEQ05 vs. around 16° in BINMEQ06/07).

The full table of crystal parameters can be found below (Table S6).

a)

c)

b)

d)

Figure S11: 1a,b) Representation of the full unit cell and the asymmetric unit for the monoclinic CUR crystal structure (BINMEQ05). On the right, the two monoclinic structures with the reference codes BINMEQ06 (1c) and BINMEQ (1d) are shown. For those two structures, the CUR molecules are almost flat, while the torsion angle in 1b) is 46°.

Table S6: Summary of the characteristics of the three crystal forms of curcumin (BINMEQ05-07)[13] used for the calculations in this project. Interestingly, they differ in the degree to which the two halves of the CUR molecule are twisted.

CCDC Code BINMEQ05 BINMEQ06 BINMEQ07

Crystal System monoclinic orthorhombic orthorhombic

Space Group P2/n Pca21 Pbca

Tmeasure [K] 100 100 100

a [Å] 12.5676 35.417 12.536

b [Å] 7.0425 7.7792 7.9916

c [Å] 19.9582 12.6482 34.462

90 90 90

94.987 90 90

90 90 90

Z 4 8 8

Z' 1 2 1

Molecular Twist [°] 46.0 18.6, 14.2 16.1

S15

The three crystal structures discussed above were used as starting point for the GIPAW (CASTEP)

calculations. In agreement with PXRD data, the as received sample was identified as the monoclinic

form of curcumin. The calculated datapoints for the other two polymorphs do not fit the experimental

spectrum. To assign the solid-state NMR data, the calculated chemical shifts were used and the results

compared to the work published by Kong et al, which agrees with the data here (Figure S12a).[12]

Additionally, the agreement between experimental and calculated data was shown for a 2D 1H-13C

HETCOR spectrum, which was recorded using a short contact time of 50 μs to extract direct CH

contacts (Figure S12b). The calculated datapoints are shown as orange crosses and were generated

with a cutoff for the C-H distance of 1.2 Å.

a)

b)

Figure S12: a) 13C CP/MAS NMR spectra of crystalline, as received and quench-cooled amorphous CUR alongside GIPAW (CASTEP) calculated chemical shifts, which are shown as coloured sticks. The calculated data served as a basis for spectral assignment and the numbering scheme is as shown above and in agreement with the one chosen by Kong et al.[12] b) 1H-13C HETCOR spectrum recorded using a short contact time of 50 μs. Orange crosses represent GIPAW (CASTEP) calculated chemical shifts. All experimental data were recorded at 14.1 T and 24 kHz MAS.

For the quench-cooled amorphous sample, the lines are considerably broadened. Interestingly, there

is a broad feature at ~ 120 ppm, which is not observed in the initial untreated sample. The calculated

data for two orthorhombic crystal structures predicts resonances in this spectral region. Both

structures do not have the pronounced twist for the two halves of the molecule. Therefore, the

amorphous sample should mostly contain curcumin molecules with a relatively flat, untwisted

backbone. To learn more about key intermolecular interactions in the different polymorphs, which

will help to understand the environment of CUR within the polymeric micelles, calculations for the

crystal are compared to a single molecule calculated in an enlarged unit cell.

S6.4 GIPAW (CASTEP) calculations of Curcumin – crystal vs. molecule

After the initial full crystal geometry optimization, one molecule was selected and placed in an

enlarged unit cell to calculate the chemical shifts in the absence of intermolecular interactions. The

difference between this and the full crystal calculations are shown below (Figure S13 and Figure S14).

Generally, a positive difference for (1H) is indicative of hydrogen bonding interactions, while a

negative difference is observed if ring current effects are involved, e.g. for CH···π interactions.[21]

S16

BINMEQ05

BINMEQ07

Figure S13: Crystal vs. molecule GIPAW NMR shielding calculation for the two polymorphs BINMEQ05 and BINMEQ07 each containing one molecule of CUR in the asymmetric unit. The differences plotted are obtained from one calculation for the full crystalline environment and one for a single molecule only.

Figure S14: Crystal vs. molecule GIPAW NMR shielding calculation for the two individual molecules of CUR in the asymmetric unit of BINMEQ06. The differences plotted are obtained from one calculation for the full crystalline environment and one for each of the single molecule only.

Throughout all the different graphs, the OH protons at each side of the curcumin molecule (highlighted

in red) are most involved in hydrogen bonding as becomes apparent form the large positive difference

between the crystal and the single molecule calculation. In contrast, all other interactions are less

pronounced and also the other OH proton, which is part of the keto-enol group, is involved in strong

intramolecular interactions but not in strong intermolecular interactions. Therefore, in the

formulations, the CUR molecule is unlikely to interact with the polymer via this functional group.

S17

S8. References

[1] M. M. Lübtow, L. Hahn, M. S. Haider, R. Luxenhofer, J. Am. Chem. Soc. 2017, 139, 10980-10983.

[2] R. Luxenhofer, A. Schulz, C. Roques, S. Li, T. K. Bronich, E. V. Batrakova, R. Jordan, A. V. Kabanov, Biomaterials 2010, 31, 4972-4979.

[3] T. Gravestock, K. Box, J. Comer, E. Frake, S. Judge, R. Ruiz, Analytical Methods 2011, 3, 560-567.

[4] H. H. Tønnesen, J. Karlsen, Zeitschrift für Lebensmittel-Untersuchung und Forschung 1985, 180, 402-404.

[5] a) J. F. Young, J. Appl. Chem. 1967, 17, 241-245; b) L. Greenspan, Journal of research of the national bureau of standards 1977, 81, 89-96.

[6] F. Payton, P. Sandusky, W. L. Alworth, J. Nat. Prod. 2007, 70, 143-146. [7] D. H. Wu, A. D. Chen, C. S. Johnson, Journal of Magnetic Resonance, Series A 1995, 115, 260-

264. [8] A. Jerschow, N. Müller, J. Magn. Reson. 1997, 125, 372-375. [9] R. Evans, Z. Deng, A. K. Rogerson, A. S. McLachlan, J. J. Richards, M. Nilsson, G. A. Morris,

Angew. Chem. Int. Ed. 2013, 52, 3199-3202. [10] Q. Yu, D. Pichugin, M. Cruz, G. Guerin, I. Manners, M. A. Winnik, Macromolecules 2018, 51,

3279-3289. [11] X. Guo, E. Laryea, M. Wilhelm, B. Luy, H. Nirschl, G. Guthausen, Macromol. Chem. Phys.

2017, 218, 1600440. [12] X. Kong, A. Brinkmann, V. Terskikh, R. E. Wasylishen, G. M. Bernard, Z. Duan, Q. Wu, G. Wu,

J. Phys. Chem. B 2016, 120, 11692-11704. [13] P. Sanphui, N. R. Goud, U. B. R. Khandavilli, S. Bhanoth, A. Nangia, Chem. Commun. 2011, 47,

5013-5015. [14] a) S. J. Clark, M. D. Segall, C. J. Pickard, P. J. Hasnip, M. I. J. Probert, K. Refson, M. C. Payne, Z.

Kristallogr. 2005, 220, 567; b) C. J. Pickard, F. Mauri, Phys. Rev. B 2001, 63, 245101; c) J. R. Yates, C. J. Pickard, F. Mauri, Phys. Rev. B 2007, 76, 024401.

[15] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865-3868. [16] D. Vanderbilt, Phys. Rev. B 1990, 41, 7892-7895. [17] S. Sturniolo, T. F. Green, R. M. Hanson, M. Zilka, K. Refson, P. Hodgkinson, S. P. Brown, J. R.

Yates, Solid State Nucl. Magn. Reson. 2016. [18] R. K. Harris, P. Hodgkinson, C. J. Pickard, J. R. Yates, V. Zorin, Magn. Res. Chem. 2007, 45,

S174-S186. [19] G. N. M. Reddy, D. S. Cook, D. Iuga, R. I. Walton, A. Marsh, S. P. Brown, Solid State Nucl.

Magn. Reson. 2015, 65, 41-48. [20] C. R. Groom, I. J. Bruno, M. P. Lightfoot, S. C. Ward, Acta Crystallogr. B 2016, 72, 171-179. [21] a) J. R. Yates, T. N. Pham, C. J. Pickard, F. Mauri, A. M. Amado, A. M. Gil, S. P. Brown, J. Am.

Chem. Soc. 2005, 127, 10216-10220; b) A.-C. Uldry, J. M. Griffin, J. R. Yates, M. Pérez-Torralba, M. D. Santa María, A. L. Webber, M. L. L. Beaumont, A. Samoson, R. M. Claramunt, C. J. Pickard, S. P. Brown, J. Am. Chem. Soc. 2008, 130, 945-954.

download fileview on ChemRxivLoading dependent Structural Model of Polymeric Micelles ... (2.47 MiB)