Subcomponent Exchange Transforms an FeII4L4 Cage from High- to Low-Spin, Switching Guest Release in a Two-Cage System

Subcomponent exchange transformed new high-spin FeII4L4 cage 1 into previously-reported low-spin FeII4L4 cage 2: 2-formyl-6-methylpyridine was ejected in favor of the less sterically hindered 2-formylpyridine, with concomitant high- to low-spin transition of the cage’s FeII centers. High-spin 1 also reacted more readily with electron-rich anilines than 2, enabling the design of a system consisting of two cages that could release their guests in response to combinations of different stimuli. The addition of p-anisidine to a mixture of high-spin 1 and previously-reported low-spin FeII4L6 cage 3 resulted in the destruction of 1 and the release of its guest. However, initial addition of 2-formylpyridine to an identical mixture of 1 and 3 resulted in the transformation of 1 into 2; added p-anisidine then reacted preferentially with 3 releasing its guest. The addition of 2-formylpyridine thus modulated the system’s behavior, fundamentally altering its response to the subsequent signal p-anisidine.


Materials and Methods
Reagents and solvents were purchased from commercial suppliers and used without further purification, unless otherwise specified. All manipulations involving cage 1 were carried out in a glovebox using CD 3 CN that had been dried over calcium hydride and distilled in vacuo.
Centrifugation of cage samples was carried out using a Grant-Bio LMC-3000 low speed benchtop centrifuge.
UV/visible spectra were recorded on a Perkin Elmer Lambda 750 UV-Vis-NIR spectrophotometer fitted with a PTP-1 Peltier temperature controller accessory. Spectra were obtained in double beam mode recording the spectra using the front beam with air in the rear beam. A background spectrum of CH 3 CN was recorded using the analyte beam prior to each experiment and a baseline correction was applied using the Perkin Elmer WinLab software suite.
IR spectra were recorded on a Perkin Elmer Spectrum One FT-IR Spectrometer fitted with a University ATR Sampling Accessory. A background spectrum was recorded prior to each experiment, and spectra were obtained over 16 scans. 1.1.1 Paramagnetic 1 H NMR spectra Paramagnetic 1 H NMR spectra were recorded using the zg30 pulse program with a 407 ppm sweep width centred at 130 ppm. The delay, D1, was set at a value five times the longest T1 value of the signals and a minimum number of 120 scans were recorded. The NMR spectra were processed applying a line broadening of 20 Hz.

T 1 Measurements
It was not possible to measure the T 1 values for all cage signals simultaneously in the 240 ppm range of the paramagnetic 1 H NMR spectrum. Therefore, the T 1 value for each signal was measured with a sweep width of 10 ppm centred on the signal or group of signals within the sweep width. Initially, the T 1 value for each signal was estimated using the t1ir1d pulse program in order to set an appropriate D1 value (five times the estimated T 1 value) and variable delay list for measuring an accurate T 1 value using the t1ir pulse program. Data was collected for a minimum of 25 delays using 8 scans and processed using Dynamics Center 2.3.3. As a small sweep width had been used, the data was fit using the inversion recovery with partial inversion fitting function (Eq. 1) with an error estimation by fit to obtain the T 1 value.
Eq. 1 1.1.3 COSY Spectra COSY NMR spectra for cage 1 were recorded using a cosyqf90 pulse program with a 120 ppm spectral width centred at 10 ppm using 28 scans, 2048 increments and acquisition times of 0.34 s and 0.021 s in the F2 and F1 dimensions, respectively. The COSY spectra were processed using a line broadening of 20 Hz in the F2 dimension, and a sine function with a sine bell shift of 2 in the F1 dimension.

Magnetic Susceptibility Measurements
Magnetic susceptibility measurements in solution were determined by the Evans' method 1 using variable temperature NMR data obtained using the same parameters as for the paramagnetic 1 H NMR spectra (Section 1.1.1). Cage 1 binds a variety of guests in CD 3 CN and commonly used internal references for mass susceptibility measurements, such as tert-butanol and cyclohexane, were found to bind in the cavity of the cage. p-Xylene was chosen as the internal reference since it only binds in traces amounts according to host-guest studies in Section 3 in order to minimise the interactions with the cage at the concentrations of internal reference required for the measurements. In a glovebox, a 1 mL solution of p-xylene (30.0 µL) and CD 3 CN was prepared in a volumetric flask. This solution was added to a 5 mm NMR tube. In a glovebox, a 1 mL solution of 1 (6.64 mg, 1.50 μmol), p-xylene (30.0 µL), and CD 3 CN was prepared in a volumetric flask. This solution was used to fill a standard melting point capillary, which was then inserted into the NMR tube containing the p-xylene/CD 3 CN solution. Data and results are included in Section 9.2.

Mass Spectrometry
Low resolution electrospray ionisation mass spectrometry (ESI-MS) was carried out on a Micromass Quattro LC (cone voltage 4-14 eV, desolvation temp. 313 K, ionisation temp. 313 K) infused from a Harvard syringe pump at a rate of 10 μL per minute.
High resolution electrospray ionisation mass spectrometry (ESI-MS) was carried out by the EPSRC UK National Mass Spectrometry Facility at Swansea University on a LTQ Orbitrap XL hybrid ion trap-orbitrap mass spectrometer.
Gas-chromatography-mass spectrometry (GC-MS) was carried out using a Shimadzu QP2010-SE fitted with a SHIM-5MS column (30 m, 0.25 mm, 0.25 μm film) for MS analysis.    The signals for H f and H g were too broad at low temperatures to accurately determine chemical shifts.

Proton Assignment through T 1 Measurements
In order to assign the paramagnetic signals of 1, T 1 relaxation values were measured: T 1 is inversely proportional to Σ(r ij ) -6 , where r ij is the distance from the paramagnetic center to the proton, according to the Solomon equation. 3 For cage 1, the T 1 values varied from 2.91 ms to 101 ms, although the broadness of the peak at -18.8 ppm precluded measurement of the T 1 value (Table S1, Figure S7). In the absence of a crystal structure for cage 1, the Fe II -proton distances from the crystal structure of the low spin analogue 2 2 were used to calculate relative T 1 values (normalised to the imine peak at 193.6 ppm) and assign the proton signals, following the methods employed by Raehm 4 and Ward 5 for paramagnetic Co II complexes. It was not possible to calculate the relative T 1 values normalised to the smallest measured T 1 value of 2.91 ms as cage 2 does not have an equivalent Fe II -proton a distance in the crystal structure. There is good agreement (within a factor of 1.3) between the measured and calculated T 1 values with the exception of proton h. This discrepancy could be due to the increased flexibility of the cage in solution compared with the solid state.
Additional confirmation of the proton assignments by 2D NMR analysis was rendered difficult by the short T 1 values and wide spectral width of the paramagnetic spectrum. The cross-peaks observed in the COSY spectrum are consistent with our assignment of the pyridine protons ( Figure S3), although correlations between protons f and g were not observed due to the broadness of the signal for proton f. Comparison to the calculated chemical shifts for a related high spin Fe II mononuclear complex provided additional support for our 1 H NMR assignments. 6 S12 Figure S7. T 1 measurements for each signal in the paramagnetic 1 H NMR spectrum of [1](OTf) 8 in CD 3 CN at 298 K. S13

Host-Guest Complexes with Cage 1
In a glovebox, 10-15 equivalents of the guest were added to a 2 mM solution of cage 1 in CD 3 CN (0.5 mL) and the solution was transferred to a J Young NMR tube, sealed and left to equilibrate at 298 K. Paramagnetic 1 H NMR spectra were recorded over time and equilibration occurred in less than 4 hours. In some cases, signals corresponding to cage protons f and g were broadened into the baseline or obscured by resonances of excess non-encapsulated guest.

Comparison of Host-Guest Complexes
The paramagnetic Fe II centers in cage 1 enabled the sensitive detection of guest encapsulation by 1 H NMR spectroscopy via two mechanisms. First, paramagnetism reduces the T 1 relaxation times of the guest and cage nuclei, allowing more scans in comparison with a diamagnetic analog to be recorded over the same time by reducing the acquisition time and relaxation delay. 7 Second, as noted by Ward, 8 paramagnetism also enhances the observation of host-guest complexes because the NMR signals are spread over a wider chemical shift range, thus reducing signal overlap and improving dispersion upon encapsulation.
Signals for the encapsulated guest were observed in all cases between -10 and -20 ppm and T 1 values were measured for signals having sufficient intensity (Table S2). Their values were of a similar magnitude to the T 1 value for proton h and reflect the isotropic shifts experienced by the guests within the paramagnetic host cavity. S14 [a]

Guest Signal Chemical Shifts and T 1 Measurements
[a] [a] [a] [a] [a] [a] [a]                             Only the empty cage is observed in the ESI mass spectrum. Figure S42. Paramagnetic 1 H NMR spectrum of [toluene  1](OTf) 8 in CD 3 CN at 298 K. Red labels refer to [toluene  1](OTf) 8 and blue labels refer to empty cage 1.        8 and blue labels refer to empty cage 1.

Competition Host-Guest Studies
In a glovebox, a competing guest (10-15 equivalents) was added to a solution of the host-guest complexes prepared in Section 3. The J Young NMR tube was sealed and left to equilibrate at 298 K. Paramagnetic 1 H NMR spectra were recorded over time and generally equilibration times were several hours.             In the Absence of a Guest In a glovebox, 24 equivalents of 2-formylpyridine (2.28 μL) was added to a 2 mM solution of cage 1 in CD 3 CN (0.5 mL) in a J Young NMR tube. The NMR tube was sealed and left to equilibrate for 16 h at 323 K and a colour change from orange to red-purple was observed. 1 H, paramagnetic 1 H and 19 F NMR spectra were recorded. In a glovebox, D 2 O (25 μL, 5% v/v) was added to the solution, the J Young tube was sealed and left to equilibrate at 323 K for 1 day. NMR spectra were recorded periodically to monitor the equilibration process.    Figure S69. Paramagnetic 1 H NMR spectra of the transformation from cage 1 (orange labels) to cage 2 upon sequential addition of 24 eq. of 2-formylpyridine and 5% D 2 O and heating at 323 K in CD 3 CN. Figure S70. 1 H NMR spectra for the transformation from cage 1 (orange labels) to cage 2 (purple labels) upon sequential addition of 24 eq. of 2-formylpyridine and 5% D 2 O and heating at 323 K in CD 3 CN. Figure S71. 19 F NMR spectra of the transformation from cage 1 to cage 2 upon sequential addition of 24 eq. of 2-formylpyridine and 5% D 2 O and heating at 323 K in CD 3 CN.

In the Presence of a Guest, 1-Fluoroadamantane
In a glovebox, 1-fluoroadamantane (1.55 mg, 0.01 mmol) was added to a 2 mM solution of cage 1 (4.42 mg, 0.001 mmol) in CD 3 CN (0.5 mL) and the mixture was left to equilibrate at room temperature for at least 4 hours. 2-Formylpyridine (24 eq., 2.28 μL) was added to the host-guest complex and the mixture was left to equilibrate for 16 h at 323 K. D 2 O (25 μL, 5% v/v) was then added to the solution and the solution was left to equilibrate at 323 K for several days. NMR spectra were recorded periodically to monitor the equilibration process.   S62 Figure S74. 19 F NMR spectra of the transformation from [1-fluoroadamantane  1] 8+ (orange labels) to [1-fluoroadamantane  2] 8+ (purple labels) upon sequential addition of 24 eq. of 2-formylpyridine and 5% D 2 O and heating at 323 K in CD 3 CN.

S63
The equilibration process was slower for the full than for the empty cage, with kinetics depending on the amount of water added: equilibration approached completion on a time scale of more than 10 days for 2% D 2 O, 3 days for 5% D 2 O and 1 day for 10% D 2 O ( Figure S75). Figure S75. 1 H NMR spectra for the transformation from [1-fluoroadamantane  1] 8+ to [1-fluoroadamantane  2] 8+ upon heating at 323 K following addition of 24 eq. of 2-formylpyridine then addition of 2, 5 or 10% D 2 O in CD 3 CN.

6.1
Spectroscopic data were consistent with those reported in the literature.

Cage 2 Synthesis
Modified literature procedure: 2 In a glovebox, Fe(OTf) 2 (28.3 mg, 0.08 mmol), ligand A (35.3 mg, 0.08 mmol) and 2-formylpyridine (22.8 μL, 0.24 mmol) were combined with degassed MeCN (10 mL). The mixture was stirred at room temperature for 21 h. The reaction mixture was added dropwise to diethyl ether (60 mL) to precipitate 2 and the mixture was centrifuged, the supernatant decanted and the solid was washed with diethyl ether two times. The purple solid was dried in vacuo to give 2 (60.3 mg, 82%).  Assignments of quaternary cage carbons are based on assignments for other host-guest complexes for cage 2 2 since cross-peaks in the HMBC were not observed, most likely due to the broadness of the signals in the 1 H NMR spectrum. For this reason, it was also not possible to assign the quaternary carbon signals for the encapsulated 1-fluoroadamantane guest.    [BF 4  3](BF 4 ) 7 was prepared according to a modified literature procedure 9 where the cage was isolated by precipitation or trituration with diethyl ether.
Spectroscopic data (Figures S93 and S94) were consistent with those reported in the literature. 9 The BF 4 anion templates the formation of the Tsymmetric diastereomer. Small peaks in the 19 F NMR spectra are attributed to encapsulation of BF 4 -    - 3] 8+ (green labels, * is attributed to encapsulation of BF 4 within trace amounts of other diastereomers) in CD 3 CN following the addition of 24 eq. p-anisidine and equilibration at room temperature for 3 days.

Titrations of Cages 2 and 3
To a solution of [1-fluoroadamantane  2] 8+ (prepared as in Section 8.2 using 2.11 mg cage 2 and an equilibration time of at least 3 days) and/or [BF 4  3] 8+ (1.55 mg, 0.0005 mmol) was added aliquots of 6 eq. p-anisidine (10 μL, 300 mM). After each addition, the mixture was equilibrated at 323 K for 1 day and NMR spectra were recorded.

Magnetic Susceptibility
According to the Evans' method, 1 the mass susceptibility  g of the dissolved substance is given by equation 2 where f is the frequency shift in Hz of the reference compound, f is the fixed probe frequency of the spectrometer,   is the mass susceptibility of CD 3 CN (-0.534 x 10 -6 cm 3 g -1 ), m is the mass in g of the complex in 1 cm 3 of solution. The 3/2π factor in the original Evans equation has been replaced with -3/4π for a sample axis parallel to the magnetic field. The molar susceptibility  M was calculated according to equation 3 by multiplying the mass susceptibility by the molecular weight (M) .

Eq. 3
The molar susceptibility  M contains the diamagnetic contribution ( M dia ) and according to Piguet, 10 this contribution cannot be neglected for large supramolecular complexes and therefore, the corrected molar susceptibility  M' was calculated according to equation 4 using tabulated values of Pascal's constants 11 to correct for the diamagnetic contributions from the ligands, Fe(II) core electrons and counteranions.  M dia for cage 1 is -0.01319 cm 3 mol -1 .
The molar susceptibilities support the 1 H NMR variable temperature experiments that cage 1 is high spin between 268 K and 318 K (Table S3).