“Clickase” Single-Chain Nanoparticles: Effect of Intra-Chain Distribution of Catalytic Sites on Catalytic Activity

“Clickase” single-chain nanoparticles (Ck-SCNPs) are folded, enzyme-mimetic unimolecular polymeric nano-objects containing copper (Cu) ions able to catalyze the azide-alkyne Huisgen cycloaddition reaction in water and/or selected organic solvents, often in the presence of a reductant. Herein, we investigate the effect of morphology on catalytic activity of Ck-SCNPs synthesized by means of two different routes. An amphiphilic random copolymer composed of oligo(ethylene glycol) methyl ether methyl methacrylate (OEGMA) and 2-acetoacetoxy ethyl methacrylate (AEMA) units was used as precursor of these Ck-SCNPs. Folding was promoted through metal complexation between Cu(II) ions and beta-ketoester-containing AEMA moieties. The first route resulted in Ck-SCNPs1 containing Cu ions homogeneously distributed within each nanoparticle, whereas the second one promoted intra-chain clustering of Cu ions inside Ck-SCNPs2. A model fluorogenic “click” reaction between 9-(azidomethyl)anthracene and phenylacetylene, which was catalyzed either by Ck-SCNPs1 or Ck-SCNPs2, was used to unravel the effect of morphology on catalytic activity. This work paves the way to improve the catalytic activity of metallo-folded SCNPs through control of the intra-chain distribution of catalytic sites.

In spite of these advances, investigations about how the intra-chain distribution of catalytic sites affects the catalytic activity of SCNPs are very scarce. We envisioned that controlling the spatial distribution of catalytic sites in metallo-folded SCNPs should be critical for the rational design of improved catalytic soft nano-objects. Based on the use of amphiphilic random copolymers and two different SCNP synthesis procedures involving selective or nonselective solvents, we reported previously a pathway for tuning the internal structure of metallo-folded SCNPs [64]. The first SCNP synthesis procedure involved the conventional synthesis in good solvent (method 1). The second one was based on transfer, after SCNP formation, from selective to good solvent conditions (method 2). By combining size exclusion chromatography (SEC) with triple detection, small-angle Xray scattering (SAXS) and molecular dynamics (MD) computer simulations we unraveled the SCNP size, sparse morphology in good solvent and spatial distribution of catalytic sizes for Cu-containing SCNPs ("clickase" SCNPs, Ck-SCNPs) synthesized by method 1 (Ck-SCNPs1) and method 2 (Ck-SCNPs2). Interestingly, we observed a homogeneous distribution of catalytic sites in the case of Ck-SCNPs1 but the presence of clusters of catalytic sites in the case of Ck-SCNPs2 in good solvent [64]. However, we did not investigate at that time the effect of these very different distributions of active sites on the catalytic activity of Ck-SCNPs1 and Ck-SCNPs2. Herein we report the results obtained from a model fluorogenic "click" reaction between 9-(azidomethyl)anthracene and phenylacetylene, which was catalyzed either by Ck-SCNPs1 or Ck-SCNPs2. Additionally, we investigate the effect of the nature of the solvent in which this model "click" reaction is carried out. The results obtained are of great interest for further advance the field of enzyme-mimetic SCNPs.

1 H Nuclear Magnetic Resonance ( 1 H NMR)
1 H NMR spectra were recorded at room temperature on a Bruker spectrometer operating at 400 MHz using CDCl3 as solvent.

Dynamic Light Scattering (DLS)
DLS measurements were carried out at room temperature on a Malvern Zetasizer Nano ZS apparatus.

Differential Scanning Calorimetry (DSC)
DSC measurements were carried out on 5-10 mg of sample using a Q2000 TA Instrument. A liquid nitrogen cooling system (LCNS) was used with a 25 mL/min helium flow rate. Measurements were performed using hermetic aluminum pans from -150 °C to 100 °C, at a scanning rate of 10 °C/min.

Thermal Gravimetric Analysis (TGA)
TGA measurements were performed in a Q500-TA Instruments apparatus at a heating rate of 10 °C/min under nitrogen atmosphere from room temperature to 800 °C.

Fluorescence Spectroscopy (FS)
Photoluminiscence spectra were recorded at room temperature on an Agilent Cary Eclipse spectrometer at an excitation wavelength of 370 nm.

Synthesis of "Clickase" SCNPs by Method 1 (Ck-SCNPs1)
In a typical reaction, P1 (100 mg, 0.06 mmol) was dissolved in THF (90 ml) at room temperature. Then, a solution of Cu(OAc)2 (6 mg, 0.03 mmol Cu) in 10 ml of THF was added, and the mixture was maintained under stirring for 24 h. After reaction completion to give Ck-SCNPs1, the system was concentrated and precipitated in hexane (twice). Finally, Ck-SCNPs1 were dried in a vacuum oven at r.t. under dynamic vacuum [64].

Results and Discussion
The aim of this work is to determine how the intra-chain distribution of catalytic sites affects the catalytic activity of metallo-folded SCNPs. As precursor of the SCNPs, we synthesized an amphiphilic random copolymer denoted as P1 decorated with -ketoester units via RAFT polymerization of MMA and AEMA, following a procedure previously optimized by our group [64]. SEC measurements with triple detection of P1 revealed a weight average molecular weight (Mw) of 72.1 kDa and a very narrow dispersity of Ð = 1.02. P1 contained 17 mol % of -ketoester moieties as determined by 1 H NMR spectroscopy. It is well-known that -ketoester groups are efficient ligands for Cu(II) ions to give Cu(-ketoester)2 complexes. When complexation takes place within an individual polymer chain decorated with -ketoester units at high dilution in a good solvent, metallofolded SCNPs are obtained (see Figure 1, method 1) [31]. As reported by Zimmerman and coworkers [54], Cu(II)-containing SCNPs under reducing conditions can function as highly efficient catalyst of the azide-alkyne Huisgen cycloaddition reaction (i.e., "ckickase" SCNPs, Ck-SCNPs). Herein we denote as Ck-SCNPs1 the Cu-containing SCNPs obtained from P1 via method 1 (see Figure 1 and section 2.3). From our previous study combining SEC, SAXS and MD simulations [64], a uniform distribution of Cu catalytic sites is expected for metallo-folded SCNPs obtained using method 1. Conversely, the presence of clusters of Cu catalytic sites is expected in metallo-folded SCNPs synthesized via method 2 when dissolved in a non-selective, good solvent (e.g., THF) (see Figure 1, method 2). We denote the Cu-containing SCNPs obtained from P1 via method 2 as Ck-SCNPs2 (see Figure 1 and section 2.3).
Characterization of Ck-SCNPs1 and Ck-SCNPs2 by means of SEC measurements in THF revealed an increase in retention time (i.e., reduction of hydrodynamic size) when compared to precursor P1, as illustrated in Figure 2A  of Cu incorporated into Ck-SCNPs1 and Ck-SCNPs2 was very similar, although as stated previously the distribution of catalytic sizes would be rather different in Ck-SCNPs2 when compared to Ck-SCNPs1 due to the presence of clusters in the former. TGA results obtained following the method reported in ref. [31] showed almost complete formation of the theoretical amount of Cu(-ketoester)2 complexes in both Ck-SCNPs1 (>99%) and Ck-SCNPs2 (>99%). DSC measurements ( Figure 3C,D) revealed a similar increase in glass transition temperature (Tg) for SCNPs1 (Tg = -41.3 °C) and Ck-SCNPs2 (Tg = -41.5 °C) when compared to that of P1 (Tg = -47.6 °C). The increase in Tg is also a signature of the efficient formation of intra-chain Cu(-ketoester)2 complexes that restrict the mobility of the AEMA chain segments involved and, probably, also that of near-neighbor segments. Consequently, the hydrodynamic size, content of Cu(-ketoester)2 complexes and thermal behavior of Ck-SCNPs1 and Ck-SCNPs2 was very similar even having a different internal distribution of Cu catalytic sites. Subsequently, we investigated the catalytic activity of Ck-SCNPs1 and Ck-SCNPs2 using a model fluorogenic "click" reaction between non-fluorescent phenylacetylene (1) and 9-(azidomethyl)anthracene (3) to give the fluorescent compound 3-(anthracen-9ylmethyl)-5-phenyltriazole (4) (see Scheme 2). Two different reaction media were selected to guarantee the solubility of reagents 1 and 3 as well as the product 4: aqueous THF (a mixture of THF and H2O at a volume ratio 3:1) and neat DMSO. It is worth mentioning that Ck-SCNPs1 and Ck-SCNPs2 are completely soluble in both aqueous THF and DMSO without the presence of aggregates, as determined by DLS measurements (see Figure 4). Control reactions for comparison were performed by replacing Ck-SCNPs1 and Ck-SCNPs2 by CuSO4 as catalyst. Figure 5A illustrates the fluorescence observed from the  product of reaction, 4, after 1h of reaction time by using Ck-SCNPs1 (blue trace), Ck-SCNPs2 (green trace) or CuSO4 (red trace) as catalysts of the fluorogenic "click" reaction between 1 and 3 in aqueous THF ( Figure 5A). Figure 5B illustrates the results obtained in DMSO at longer reaction time. In this sense, the fluorogenic "click" reaction was found to be faster in aqueous THF than in DMSO and Ck-SCNPs2 being more efficient than Ck-SCNPs1 and CuSO4 in both solvents. The high reproducibility of the results and the high sensitivity of the fluorescence spectroscopy technique allowed us to perform such a reliable comparison between different catalytic systems. Remarkably, the difference in catalytic activity observed (Ck-SCNPs2 > Ck-SCNPs1 > CuSO4) was more notorious in DMSO than in aqueous THF. The slower reaction rate in DMSO can be attributed -to a large extent-to the oxidant power of this solvent [67] that oxidizes a certain amount of Cu(I) ions generated by NaAsc to Cu(II) ions, the latter being inactive for the "click" reaction. Interestingly enough, the presence of cluster of Cu ions in the case of Ck-SCNPs2 was found beneficial for improving the stability and effectiveness of the "clickase" SCNPs under the deactivating, oxidative nature of the DMSO solvent. The above experiments with a model fluorogenic "click" reaction and different reaction media provide solid support that the intra-chain distribution of catalytic sites has an important effect on the catalytic activity of "clickase" SCNPs, especially in solvents with significant oxidative nature like DMSO [67].

Conclusions
We investigate the effect of intra-chain distribution of catalytic sites on catalytic activity of Ck-SCNPs as folded, enzyme-mimetic unimolecular polymeric nano-objects containing Cu ions able to catalyze the azide-alkyne Huisgen cycloaddition reaction under appropriate reaction conditions. By using an amphiphilic random copolymer composed of 83 mol% of OEGMA units and 17 mol% of -ketoester-containing AEMA units, we synthesize Ck-SCNPs by two different methods. The first one results in Ck-SCNPs1 containing Cu ions homogeneously distributed within each nanoparticle in the form of individual Cu(-ketoester)2 complexes, whereas the second method promotes intra-chain clustering of Cu(-ketoester)2 complexes inside Ck-SCNPs2. To unravel the effect of morphology on catalytic activity, we have evaluated the efficiency of Ck-SCNPs1 and Ck-SCNPs2 in a model fluorogenic "click" reaction between non-fluorescent 9-(azidomethyl)anthracene and phenylacetylene to give the fluorescent compound 3-(anthracen-9-ylmethyl)-5-phenyltriazole. Fluorescence spectroscopy experiments allowed us to determine the catalytic efficiency of Ck-SCNPs1 and Ck-SCNPs2 when compared to a classical catalyst such as CuSO4. Using aqueous THF or neat DMSO in the presence of NaAsc as reducing agent, the catalytic activity was found to follow the order: Ck-SCNPs2 > Ck-SCNPs1 > CuSO4. The fluorogenic "click" reaction was faster in aqueous THF than in DMSO -a solvent that promotes the oxidation of Cu(I) to Cu(II)-with Ck-SCNPs2 being more efficient than Ck-SCNPs1 (or CuSO4) in both solvents. The presence of clusters of Cu ions in the case of Ck-SCNPs2 was beneficial for improving the stability and effectiveness of "clickase" SCNPs in DMSO solvent. In this sense, control of the intra-chain distribution of catalytic sites could be also a useful strategy to improve the catalytic activity of other metallo-folded SCNPs containing metal ions different from Cu ones.