Nanometre-Scale Visualization of Chemical Parameter Changes by T 1 -Weighted ODMR Imaging Using a Fluorescent Nanodiamond

: The dynamics of physical parameters in cells is strongly related to life phenomena; thus, a method to monitor and visualize them on a single-organelle scale would be useful to reveal unknown biological processes. We demonstrate real-time nanometre-scale T 1 -weighted imaging using a ﬂuorescent nanodiamond. We explored optically detected magnetic resonance (ODMR) contrast at various values of interval laser pulse ( τ ), showing that su ﬃ cient contrast is obtained by appropriate selection of τ . By this method, we visualized nanometre-scale pH changes using a functionalized nanodiamond whose T 1 has a dependence on pH conditions.


Introduction
Physical parameters in cells have a heterogeneous distribution in space and time. For example, temperature near the mitochondria is considered to be comparatively higher than that in the cytosol due to the heat generation by mitochondrial respiration [1,2]. In addition, an intracellular pH gradient is thought to regulate signal transduction during actin polymerization, which plays an important role in protoplasmic flow, cell division and cell migration [3,4]. Changes in local protein concentration and viscosity to form membrane-less organelles are thought to regulate RNA transcription and enable the cell to endure temporal stresses [5,6]. As described herein, the dynamics of these parameters is strongly related to several cell events; thus, a method to monitor and visualize them on a single-organelle scale, with a spatial resolution on the order of nanometres, would be useful to reveal unknown biological processes [7].
Magnetic resonance imaging (MRI) is one of the best-known methods to visualize simultaneously both physical and locational changes in various parameters. In particular, longitudinal relaxation time (T 1 )-weighted imaging is a powerful method to visualize tissue structure, molecular composition and protein dynamics inside a living body [8][9][10][11]. However, the general coil detection mechanism restricts its spatial resolution to the milli-to micrometre order, and thus it cannot provide insight about sub-micrometre events that are involved in many biological processes.
In recent years, nitrogen-vacancy (NV) centres in diamond have been attracting much attention as atomic-size MR detectors [12][13][14]. Fluorescence from an NV centre does not show any photo-bleaching, making it possible to obtain an optically detected magnetic resonance (ODMR) spectrum from only one NV centre [15][16][17]. In addition, the T 1 value obtained from the ODMR experiment reflects the nanometre-scale environment around the diamond surface. Previous studies have shown that T 1 reflects physical parameters such as redox potential [18], protein concentration [19], and pH [18,20]. Therefore, if a T 1 -weighted imaging technique based on NV centres can be established, it would be possible to achieve real-time monitoring of physical parameters in nano-space. Although many studies have described methodologies to measure T 1 from the fluorescence of NV centres, as well as the experimental results of numerical measurements, a procedure to carry out T 1 -weighted imaging to visualize simultaneously both the dynamics and location of a parameter has not been reported.
Here we describe a method to construct T 1 -weighted ODMR images reflecting the nano-spatial environment around a nanodiamond containing NV centres. First, we optimized the measurement parameters used for T 1 -weighted imaging to obtain an adequate contrast. To conduct this, the contrasts of T 1 -weighted imaging were calculated for various values of laser pulse interval (τ) by simulation. As a result, it was suggested that an appropriately fixed value of τ can give an enough contrast in the T 1 range typical of changes in a 100-nm nanodiamond previously reported [20]. Next, we applied the method to experimentally obtained fluorescence images of a nanodiamond whose T 1 has a dependence on nano-spatial pH. This showed that the reconstructed T 1 -weighted images can visualize nano-spatial pH conditions at sufficient contrast, in agreement with the results from simulation. By extending this method to other parameters, we expect that previously unknown sub-micrometre biological processes will be revealed.

Preparation of Carboxylated Nanodiamond
We followed the protocol reported previously [20]. Nanodiamond powder was purchased from Element Six (Micron + MDA 0-0.10 µm). First, the nanodiamond powder was electron-irradiated (2 MeV, 1.0 × 10 18 e -/cm 2 ) and thermally annealed at 800 • C for 2 h under vacuum. Next, the nanodiamonds were oxidized at 550 • C for 2 h. The residue was heat-treated with a mixture of H 2 SO 4 /HNO 3 (9:1 v/v) at 70 • C for 3 days, followed by 0.1 M NaOH at 90 • C for 2 h and 0.1 M HCl at 90 • C for 2 h. The residue was washed with Milli-Q to give carboxylated nanodiamonds, which were characterized by FT-IR, XPS and zeta potential measurements.
Functionalized nanodiamonds dispersed in water were injected into a flow chamber assembled from glass slides (Matsunami micro cover glass). A single-crystal nanodiamond nonspecifically adsorbed on the bottom of the chamber was used for T 1 measurements. Each pH condition was achieved by injecting 200 µL (20-fold larger than the chamber volume) of standard buffer solution.
To replace the buffer solution inside the flow chamber, we injected 500 µL of Milli-Q water into the chamber to wash out the old solution, followed by 200 µL of the appropriate standard buffer solution to achieve the next target pH condition.

T 1 -Weighted Imaging
Following the protocols described in Section 2.2., data sets of fluorescence images and T 1 values in various pH conditions were obtained. From these T 1 values, the delay time τ that gives the highest contrast (τ max ; see also Figure 1a) was estimated. To construct T 1 -weighted images, fluorescence intensity changes from τ = 0 to τ = τ max were obtained pixel by pixel and these were plotted as a colour map. Simulations and experimental data analysis were performed by Wolfram Mathematica.

Simulation of T 1 -Weighted Images
In our method, polarization and readout laser pulses are applied to nanodiamonds with an interval time τ to perform T 1 -weighted imaging (Figure 1a). The fluorescence intensity, which decreases exponentially with τ, can be recorded by a highly sensitive photodetector such as an EMCCD camera. Here, we define a signal at τ as the difference between fluorescence intensity at interval time = 0 and that at interval time = τ (Figure 1b). Pixel signals are mapped with false colours to obtain T 1 -weighted images of a nanodiamond. However, T 1 can take several values depending on the microenvironment surrounding the nanodiamond. Therefore, external stimulation can alter the T 1 value-for instance, from T 1a to T 1a , as shown in Figure 2b-resulting in a contrast between T 1 -weighted image changes before and after stimulation. However, this contrast also depends on τ (Figure 1b), which means that an appropriate τ needs to be selected. To search for the optimal τ, we simulated a simplified model in which T 1 changes between two values, T 1a and T 1b , within a typical range of T 1 for NV centres in a 100-nm nanodiamond(400-2500 µs). We then plotted the value of τ giving the highest contrast between each pair of T 1a and T 1b values ( Figure 2, black dot). As shown, the highest contrast shifted, depending on both T 1a and T 1b , indicating that τ must be selected in accordance with the T 1 range of most interest.
Next, we considered more detailed conditions in which T 1 takes several values. Simulated T 1 -weighted images were constructed for the 400-1000 µs range, within which T 1 of a carboxylated 100 nm nanodiamond changes depending on pH (Figure 3) [20]. As a result, comparatively short τ (300 µs) gave relatively high contrast within a narrow T 1 range (400-550 µs). By contrast, comparatively long τ (2100 µs) gave relatively high contrast within a wide T 1 range (850-1000 µs). Intermediate values of τ (600 or 1200 µs) were found to give good contrast over a wider range of T 1 . This shows that, by selecting an appropriate value of τ, we contain obtain T 1 -weighted images with sufficient contrast.

T 1 -Weighted Images from a Nanodiamond under Various pH Conditions
To demonstrate the method, we constructed T 1 -weighted images obtained from a 100-nm nanodiamond under different pH conditions. As reported previously, T 1 of a carboxylated nanodiamond changes depending on the local pH conditions [20]. The surface of a carboxylated nanodiamond becomes negatively charged in alkaline conditions due to ionization of the carboxyl groups. These charges act as electrical noise to shorten T 1 . In acidic conditions, by contrast, the surface becomes neutral and T 1 gets longer. Thus, T 1 -weighted images are considered to visualize the nanometric spatial pH condition around a nanodiamond.
Here, carboxylated nanodiamonds were placed in buffers of various pH, and fluorescence intensities in each of pixels were recorded during irradiation with laser pulse trains (Figure 1a). T 1 -weighted images were constructed by using τ = 525 µs when the pH was changed from pH 3 (T 1 = 738 µs) to pH 9 (T 1 = 412 µs), resulting in the T 1 -weighted images changing in accordance with pH (Figure 4a, top). The imaging contrasts were sufficient to recognize the difference visually, as predicted in the simulation. The height of a gaussian function fitted to the signal intensities on a Y cross-section increased with increasing pH (Figure 4a, bottom). The value for the definite integral of the fitting function also increased with increasing pH (Figure 4b). These results show that our method can visualize nanometric spatial pH conditions with sufficient contrast. To quantify the pH dependence of the contrast obtained from the carboxylated nanodiamond statistically, we calculated the correlation coefficient and compared it with one obtained from a non-carboxylated nanodiamond whose T 1 has no dependence on pH. The maximal information coefficient (MIC) was used to measure the strength of the non-linear correlations. As a result, MIC from the carboxylated nanodiamond (0.99) was much higher than that from the non-carboxylated nanodiamond (0.52) ( Figure S1). Lastly, we evaluated the reversibility of the method. pH is a parameter that changes in both directions. Therefore, reversibility is critically required for pH probes. T 1 -weighted images were constructed by using τ = 875 µs when the pH was alternated between pH 3 (average T 1 = 1139 µs) and pH 9 (average T 1 = 656 µs), with the result that the T 1 -weighted images and their signal intensities on a cross-section Y changed reversibly depending on the pH conditions (Figure 5a,b). Therefore, these results suggest that repeatedly changing pH can be visualized in nanometric space by our method.

Discussion
It is known that nanodiamonds work as a nanometre-sized quantum sensor inside a living cell and have low toxicity to cells [15]. So, through the development of our method, pH conditions in organelles (3-10 µm) such as lysosomes (pH 4.5-4.7) and mitochondria (pH~8) may be visualized [4] (see also Supplementary Materials Section S2). It should be noted, however, that T 1 may be sensitive not only to pH but also to other biological parameters such as temperature, electric field, and magnetic field. Thus, further optimizations and improvements will be needed before bio-application. In the future, time resolution, sensitivity, and selectivity are likely to be improved by peripheral technology, such as nanodiamond manufacturing methods, mechanical device performance, and decoupling methods, among others. In addition, the possibility of measuring other physical parameters will be explored by examining various surface modifications of nanodiamond. These studies will increase our understanding of global life phenomena by characterizing local physical quantities in cells, providing important insight into disease diagnosis, pathological mechanisms, and drug discovery.

Conclusions
In summary, we have developed a T 1 -weighted imaging technique based on nanodiamond detection that can visualize nano-spatial pH conditions. First, we investigated the relationship between τ and its contrast in the T 1 range typical of 100-nm nanodiamonds. Our simulations showed that pH could be visualized with sufficient contrast by selecting an appropriate τ. We then showed that T 1 -weighted images constructed from a 100-nm diamond could indeed visualize nano-spatial pH with sufficient contrast within 3 minutes. The reversibility was investigated, which showed that our method can be used even when the pH changes repeatedly.
Supplementary Materials: The following are available online at http://www.mdpi.com/2227-9040/8/3/68/s1, Figure S1: (a) T 1 -weighted images of a non-carboxylated 100-nm nanodiamond in various pH conditions (top) and its signal intensity on a cross section Y = 5.1 mm plotted against X (bottom). The red line represents the best fit function consisting of seven Gaussian curves. (b) Definite integrals of the best fit Gaussian curves of each pH condition. Error bars represent the standard error for the fitting.