Ultra-rapid separation of an angiotensin mixture in nanochannels using shear-driven chromatography

Abstract

The present paper reports on the separation of a mixture of fluorescein isothiocyanate-labeled angiotensin I and II peptides in a shear-driven nanochannel with a C₁₈-coating and using an eluent consisting of 5% acetonitrile in 0.02 M aqueous phosphate buffer at pH 6.5. The flat-rectangular nanochannel in fused silica consisted of an etched structure in combination with a flat moving wall. The very fast separation kinetics that can be achieved in a nanochannel allowed to separate the angiotensin peptides in less then 0.2 s in a distance of only 1.8 mm. Plate heights as small as 0.4 μm were calculated after substraction of the injection effect.

Introduction

Started out in the early 1990s [1], [2], [3], the field of miniaturized microfluidic devices has grown rapidly over the last years. Miniaturization of analytical chemical methods offer many advantages, including the reduced requirements for solvents and reagents, short reaction times, the possibility to build portable instruments, low cost, low power consumption, versatility in design, and possibility for parallel operation and integration with other miniaturized devices [4]. Miniaturization has lead to lab-on-a-chip devices where one or several separations are performed in microfluidic channels on a silicon or glass chip [5], [6]. An example of these chip-based separations system is the combined nano high performance liquid chromatography (HPLC)/mass spectrometry (MS) polymer microfluidic chip that has recently been described and commercialized by researchers of the Agilent Laboratories [7]. The best liquid chromatography (LC) separation could be achieved when running the chip at flow rates between 200 and 300 nL/min. The column had a 75 μm × 50 μm cross-section and a length of 45 mm and the LC experiments took 40 min.

Although most systems have dimensions in the micrometer scale, more and more researchers start to look into the possibility to use sub-micron channels, entering the area of “nanofluidics” [8], [9]. One of the motivations for this work is that reducing the system dimensions offers large advantages in liquid chromatography. Considering that the time needed to diffuse across a given distance d is according to Einstein's diffusion law roughly equal to t = d²/D_mol, it follows immediately that performing an open-tubular LC (OT-LC) or capillary electrochromatography (CEC) separation in a channel that is only 0.1 μm deep instead of 10 μm would allow a 10 000-fold reduction of the separation time. Several methods for the fabrication of nanochannels with [10] or without nanoimprint lithography [11] have been investigated and pressure-driven as well as electrokinetic nanochannel flows have already been reported [12], [13], [14], [15], [16]. The nanochannel research also already focused on the selection of the most appropriate substrate materials, considering poly(dimethylsiloxane) (PDMS), silicon as well as glass substrates [17], [18]. Eijkel et al. studied drying-out effects caused by osmosis and pervaporation through polyimide nanochannels [19]. Using optimized surface machining techniques also made it possible to produce channels with lateral dimensions below 100 nm and to realize a pneumatically actuated, capillary-pressure-driven micropump capable of delivering picoliter amounts of liquid in a controlled way [20], [21]. Nanochannels also have the advantages that their dimensions are smaller than the size of biological macromolecules, with potentially useful consequences for separations; they also have a very high ratio of surface to volume, and thus allow the study of wall effects in biological separations [22], [23].

The two main methods for driving the flow of fluids in microchannels, pressure-driven and electrokinetic, however, each have their drawbacks. Pressure-driven flow performs poorly in assays requiring high-resolution separation because the velocity profile of a cross-section is parabolic, and samples in the form of plugs undergo axial dispersion and peak broadening. The main problem however is that extremely high inlet pressures are needed to drive the fluid flow through the narrow nanochannels. Electrokinetic flow also has important drawbacks for bioassays, including buffer incompatibility (only buffers of appropriate pH and ionic strength are compatible), the need for an off-chip power supply, electrolytic bubble formation, and evaporation of solvent due to heating [4]. The main problem with respect to nanochannel flow is that the electrical-double-layers that are formed at the channel walls start to overlap at the center of the channel. This leads to the formation of a more parabolic instead of a perfectly flat flow profile, and more importantly, also to a dramatic reduction of the achievable flow rate [24], [25].

In the past few years, our group has been developing a shear-driven flow method that uses channels composed of two independent longitudinal walls kept together by applying a small normal load to the assembled system [26], [27], [28], [29], [30]. The fluid flow in this system is caused by the dragging effect originating from the movement of one channel wall while the other stays stationary (Fig. 1(a)). As described in [26], this leads to a linear velocity profile with a mean velocity equaling:

$$u=\frac{u_{\text{wall}}}{2}$$

The flow velocity is thus independent of the fluid viscosity and the channel depth and length, which allows the use of channels with very small depths at high fluid flow velocities [27]. Applying a retentive coating on the stationary wall then in principle allows to perform all forms of liquid chromatography without any restriction on the applicable fluid velocity or channel depth. In previous work, the practical feasibility of this shear-driven chromatography (SDC) principle has been demonstrated by separating a mixture of four coumarin dyes in sub-second times [31]. Whereas coumarin dyes are neutral compounds and have a high autofluorescence, the present study is devoted to demonstrate that the SDC principle can also be applied to separate ionisable species. The motivation for this work is that whereas HPLC and OT-LC are two-phase systems, the SDC channels have a third phase, i.e., the surface of the moving wall. This surface is as large as the stationary wall surface and hence makes up an important part of the channel. To have a minimal interference with the actual separation, this third phase (i.e., the moving wall) should be as inert as possible. Since uncoated glass substrates currently are the single cost-effective alternative to be used as the moving wall, the presence of the silanol groups on this surface can certainly be expected to compromise the inertness requirement in the case of ionisable components. It is therefore a priori not obvious that ionisable components can be separated with a high resolution.

As a model system, we selected the separation of two biologically important molecules: angiotensin I (ANG I) and angiotensin II (ANG II). The renin-angiotensin system is well known to be the most important pressor system in the body. In the circulatory system, angiotensins (ANGs), are potent vasoconstrictor peptides that play an important role in normal physiology processes and in various disease states [32]. Angiotensinogen is cleaved by the protease renin to produce ANG I, a biologically inactive decapeptide. The angiotensin converting enzyme removes the carboxyl terminal His-Leu from ANG I to produce the octapeptide ANG II, which explains the lower hydrophobicity of ANG II. ANG II is a potent pressor agent and the determination of angiotensins in mixtures is of interest in connection with the vital pressor mechanism.