Temperature Response of ${}^{129}\mathrm{Xe}$ Depolarization Transfer and Its Application for Ultrasensitive NMR Detection

Trapping xenon in functionalized cryptophane cages makes the sensitivity of hyperpolarized (HP) ${}^{129}\mathrm{Xe}$ available for specific NMR detection of biomolecules. Here, we study the signal transfer onto a reservoir of unbound HP xenon by gating the residence time of the nuclei in the cage through the temperature-dependant exchange rate. Temperature changes larger than $\sim 0.6\mathrm{K}$ are detectable as an altered reservoir signal. The temperature response is adjustable with lower concentrations of caged xenon providing more sensitivity at higher temperatures. Ultrasensitive detection of functionalized cryptophane at 310 K is demonstrated with a concentration of 10 nM, corresponding to a $\sim 4000\mathrm{\text{−}}\mathrm{\text{fold}}$ sensitivity enhancement compared to conventional detection. This makes HPNMR capable of detecting such constructs in concentrations far below the detection limit of benchtop uv-visible light absorbance.

Figure 1

(a) ${}^{129}\mathrm{Xe}$ NMR spectrum illustrating that only a small fraction of xenon is associated with a functionalized cryptophane cage at $33\mu \mathrm{M}$ cage concentration. Increasing $T$ from 295 K (black spectrum) to 313 K (red spectrum) shifts the signal of encapsulated xenon by ca. 5.5 ppm and induces accelerated chemical exchange that causes line broadening and decreases $S/N$. (b) Chemical structure of the functionalized cryptophane cage used in this study.

Figure 2

(a) Input function $T(t)$ (black solid line) for hyper-CEST signal transfer using the construct shown in Fig. 1. The saturation frequency offset ${\omega }_{\mathrm{sat}}$ (gray dashed line) is adjusted according to the resonance shift of $26.6\mathrm{Hz}/\mathrm{K}$. Bars indicate the bandwidth of the saturation pulse of $\Omega =76.7\mathrm{Hz}$ (0.925 ppm). (b) Response $I_{\mathrm{res}}(t)$ to $T(t)$. Varying the temperature changes ${}^{129}\mathrm{Xe}$ signal transfer through the cage that is present at $27.2\mu \mathrm{M}$. The reference signal for no saturation transfer is illustrated by the dashed line.

Figure 3

(a) Calibration for temperature sensitivity of $I_{\mathrm{res}}$ shown in Fig. 2b at $27.2\mu \mathrm{M}$ cage concentration (×) and for $11.2\mu \mathrm{M}$ cage concentration (+). The signal intensity is normalized to a control experiment without saturation pulse. Data are fitted to Eq. (1) and shown with a 95% confidence band (dashed lines). (b) The differential sensitivity of ${}^{129}\mathrm{Xe}$ signal, intensity change per kelvin, is given by the first derivative of the fit results in (a). Reducing the concentration of the cage construct shifts the maximum sensitivity to higher temperatures. Since the experimental SD in determining $I_{\mathrm{res}}$ is $\sim 3.6\%$ (see below), the temperature resolution is limited to ca. 0.6 K for the $27.2\mu \mathrm{M}$ solution.

Figure 4

High-sensitivity detection of 1.4 nM encapsulated xenon signal (10 nM cage present) in ${\mathrm{D}}_2\mathrm{O}$ at 310 K. The SD in the control data set is 3.6% and is systematic noise due to imperfections in the xenon delivery. Maximum signal depletion is 16% of a resonance that was detected with a $S/N$ of 110. However, the limiting factor for the hyper-CEST detection is the variability of 3.6% in the off-resonance experiment, so $S/N$ for indirect sensor detection is $16.5/3.6=4.6$.