NMR magnetization-transfer analysis of rapid membrane transport in human erythrocytes

Abstract

Nuclear magnetic resonance (NMR) magnetization-transfer (MT) experiments provide a convenient tool for studying rapid sub-second membrane-transport processes in situ in metabolically active cells. These experiments are used with membrane-permeable substances when separate (resolved) NMR signals are observed from their populations inside and outside the cells. Here, we provide a description of the theory and practice of the most common NMR MT experiments that have been used to study membrane-transport processes in human erythrocytes (red blood cells; RBCs). The procedures, involved in the analysis of the experimental data for defining mechanisms of transport, and for estimating values of kinetic parameters in the corresponding mathematical models, are given special attention.

Keywords: Equilibrium exchange; Human erythrocyte; Magnetization transfer; Membrane transport; Michaelis-Menten kinetics; NMR.

Fig. 1

Kinetic scheme illustrating transmembrane exchange of a solute in a red blood cell. The semi-permeable plasma membrane of the cell separates the solution into extra- and intracellular compartments. S_o and S_i represent populations of a membrane-permeable hydrophilic solute S outside and inside the cells, respectively. k _oi and k _io are the unidirectional rate constants for the influx and efflux of S, respectively

Fig. 2

NMR pulse sequences used in magnetization-transfer experiments. The black rectangles represent high-power 90° radio-frequency (RF) pulses; the acquisition of the free induction decay is shown as a decaying oscillation at the end of the pulse sequence. The mixing delay is denoted by t _m. a Conventional 1D pulse-and-acquire. b Saturation-transfer experiment: the selective saturating irradiation, denoted by the open rectangle, can be achieved using a DANTE pulse train (Morris and Freeman 1978). c Inversion-transfer experiment: the rounded open rectangle represents a selective inversion (180°) RF pulse. d 1D EXSY: when the carrier frequency is at one of the exchanging resonances and the free-precession delay δ is set to $\frac{1}{2\mathrm{\Delta }\nu }$, where Δν is the difference in the chemical shifts of the two magnetizations (in Hz), a selective inversion is achieved, as shown in Fig. 3 (Robinson et al. 1985); the recommended phase cycling: third pulse: x, y, -x, -y; receiver: x, y, -x, -y. e 2D EXSY: in comparison with the 1D EXSY, the first free-precession delay is now substituted by the evolution time t ₁, which is incremented in the usual 2D-NMR fashion. It is advisable to suppress the axial peaks by a two-step phase cycle: first pulse: x, -x; receiver: x, -x (Claridge 2009)

Fig. 3

Vector-model illustration of the ‘δ-ordered inversion’ procedure, which is used in the 1D EXSY pulse sequence shown in Fig. 2d. a The two magnetization vectors are aligned along the z-axis (direction of the external magnetic field) at the beginning of the pulse sequence. b Both vectors are flipped onto the y-axis of the rotating frame of reference after the first 90°_x pulse. c Evolution of the vectors in the transverse plane during the free-precession delay δ; the black vector does not move because the carrier frequency is set at its chemical shift. d After delay δ = $\frac{1}{2\mathrm{\Delta }\nu }$, where Δν is the difference in the chemical shifts of the two magnetizations (in Hz), the vectors are pointing in opposite directions along the y-axis. e The magnetization of one of the vectors is selectively inverted after the application of the second 90°_x pulse. Adapted with permission from Robinson et al. (1985)

Fig. 4

‘Overdetermined’ set of one-dimensional ³¹P (161.5 MHz) 1D EXSY NMR spectra of 40 mM hypophosphite in an RBC suspension at 37 °C. a The ‘fully relaxed’ equilibrium spectrum: the high and low frequency peaks are from the extra-(o) and intracellular (i) species, respectively. b–g The upper spectra were acquired using the 1D EXSY pulse sequence of Fig. 2d with zero mixing time (t _m = 0) and δ equal to 3/16Δν, 4/16Δν … 8/16Δν s, respectively; the corresponding lower spectra were acquired using the same parameters, except t _m = 1 s. Adapted with permission from Price and Kuchel (1990)

Fig. 5

Michaelis–Menten plot demonstrating the dependence of the fluoride efflux rate from RBCs on the intracellular fluoride concentration, under equilibrium-exchange conditions. The solid line is the fit of the Michaelis–Menten model that allowed estimation of the K _m and V _max values at 27 ± 3 mM and 18.0 ± 0.5 fmol cell⁻¹ s⁻¹, respectively. Reproduced with permission from Chapman and Kuchel (1990)

Fig. 6

¹⁹F (470.4 MHz) NMR spectra of a fluorine-labelled analogue of glucose. a 1D pulse-and-acquire; and b 2D EXSY (t _m = 0.5 s) of 2,3,4-trifluoro-2,3,4-trideoxy-D-glucose (FDG-234) in 123 mM NaCl, 15 mM Tris/HEPES, 5 mM ascorbate, recorded at 37 °C. The EXSY experiment was conducted using 10 mM FDG-234 in the presence of RBCs (Ht 0.5). All six resonances, visible in (a) were split into two components: a broader peak, shifted to higher frequency, which corresponds to the intracellular resonance; and a sharper peak, corresponding to the extracellular resonance. The cross-peaks, visible in the 2D EXSY spectrum, indicate the exchange between the intra- and extracellular pools. Reproduced with permission from Bresciani et al. (2010)

Fig. 7

²H (61.422 MHz) 2D EXSY spectra of HDO, recorded at 15 °C in the presence of RBCs in a gelatin gel, stretched to twice its original length. a Oblique view of a spectrum acquired using t _m = 1 ms. The two most prominent resonances were assigned to the extracellular HDO while the central resonance was assigned to HDO inside the RBCs. The two remaining peaks are the cross-peaks between the two extracellular spin-populations. b Oblique view of a spectrum acquired using t _m = 75 ms. The spectrum was scaled to give the largest diagonal signal with the same amplitude as in (a); the additional four cross-peaks indicate the transmembrane exchange of HDO. The insets show the corresponding contour plots. Adapted with permission from Kuchel and Naumann (2008)