2Department of Medical Pharmacology and Physiology, University of Missouri, Columbia, MO, USA
When separating two species with similar densities but differing sedimentation velocities (because of differences in size), centrifugal elutriation is generally the method of choice. However, a major drawback to this approach is the requirement for specialized equipment. Here, we present a new method that achieves similar separations using standard benchtop centrifuges by loading the seperands as a layer on top of a dense buffer of a specified length, and running the benchtop centrifugation process for a calculated amount of time, thereby ensuring that all faster moving species are collected at the bottom, while all slower moving species remain in the buffer. We demonstrate the use of our procedure to isolate bacteria from blood culture broth (a mixture of bacterial growth media, blood, and bacteria).
In many applications, separation of target cells from mixtures containing other cells with similar densities, but different sizes (either slightly or substantially), is desired. Examples include the generation of synchronized cell cultures where one isolates cells at a particular stage of their cycle (characterized by their size) (1,–2), isolation of corticotropes from other pituitary cells (3), and sorting of various sub-populations of monocytes (4). The method of choice for separating cell populations based on differences in sedimentation speed (especially when density differences are not significant) is a process called centrifugal elutriation (5). Here, the separands (usually cells) are subjected to two forces: (i) a centrifugal force that causes them to sediment with a velocity that is a function of their density and size, and (ii) a convective flow in the opposite direction that imparts a fixed velocity to all cells. The flow and/or the centrifugation speed can be adjusted to collect only those cells whose sedimentation velocity falls in a particular range. While this process has also been successfully used to separate bacterial cells from matrices in which they are present (6) (our application of interest), the need for specialized equipment serves as an impediment to many labs.
While there are a number of protocols that have been described for isolating bacteria from various matrices such as food (7,–8), soil (9), and microbial mats (10) which make use of readily available benchtop centrifuges to separate cells on the basis of differing sedimentation speeds, the recommended protocols often appear to be ad hoc. For instance, Wang et al. (8) recommended centrifugation at <1000×g for an unspecified amount of time to remove debris when isolating bacteria from soft cheese, followed by centrifugation at 9000×g for 3 min to collect bacteria. Similarly, for isolating bacteria from homogenized meat, Neiderhauser et al. (7) recommended centrifugation at 100×g (for an unspecified length of time) to remove food particles, followed by centrifugation at 3000×g (again, for an unspecified length of time) to collect bacteria. In other cases, recommended protocols involve multiple rounds of centrifugation. For instance, in order to isolate bacteria from microbial mats, Bey et al. (10) recommend not only centrifugation at 500×g for 15 min, but also resuspension of the sediment and a repeat of the procedure four times. Similarly, to isolate bacteria from soil, Bakken (9) recommends repeated (for a “number of times”) centrifugation of the sample homogenate at 630–1060×g (for an unspecified amount of time), followed by resuspension and rehomogenization. Similar approaches have also been proposed to isolate bacteria from “positive” blood culture broth. To obtain pure isolates of bacteria prior to determining their identity using MALDI-TOF mass spectrometry, Stevenson et al. (11) used a multi-step centrifugation procedure involving sequential centrifugation steps for “1-2 minutes” each at 8500, 1000, and 13,000 RPM (value in g not specified) with resuspension of the pellet in different buffers at each step.
These cited protocols are by no means the only examples of ad hoc centrifugation protocols characterized by apparently arbitrary choice of centrifugation speeds and/or times. An important consequence of such ad hoc protocols is that they do not provide any rational basis with which others can devise protocols to isolate different target cells from other matrices/mixtures of cells. For instance, it is intuitively obvious that prolonged centrifugation will result in all particles settling, whereas centrifugation for a short duration will only sedment the fastest particles, leaving most slower particles in the suspension. Determining the optimum speed/duration for the centrifugation to isolate the target particle while eliminating others, however, remains more of an art. In other words, currently reported protocols do not provide a rational basis for modification (determining optimum centrifugation speeds/times) for other related separation/isolation objectives.Materials and methods Theory
Our proposed method to separate two particle-types that have similar densities, but different sedimentation velocities because of size differences, is illustrated in Figure 1. At the beginning of the process (Figure 1(1)), the suspension (blood culture broth, in our case) containing the separands (white blood cells or WBCs, red blood cells or RBCs, platelets, and bacteria) is loaded as a separate layer on top of a clear dense buffer (Histopaque 1083, in our case). The density of the buffer is greater than that of WBCs (ρ = 1.06 – 1.08 g/mL (12)) and platelets (ρ = 1.05 – 1.07 g/mL (13)), but lower than that of RBCs (ρ = 1.1 g/mL (14)) and bacteria (ρ = 1.105 g/mL for Escherichia coli (15)). As centrifugation proceeds, all particles are propelled downward, first through the top layer (blood culture media), and then into the dense buffer (the Histopaque). While particles with densities greater than that of the buffer (RBCs and bacteria) sediment through both the blood culture broth and the dense buffer, less dense separands (WBCs and platelets) are excluded from the latter. This process is similar to the protocol recommended for separating WBCs and platelets from whole blood (16). For our purpose, however, this process alone is insufficient because while some undesired particles (WBCs and platelets) are stripped from the target particles (bacteria) on the basis of density differences, other particles having similar densities (RBCs) remain. To achieve our goal of collecting all target cells (bacteria) while eliminating all the remaining undesired particles (RBCs), we propose to take the system to the state depicted in Figure 1(4), and halt the centrifugation process there. In this state, all the RBCs, which, because of their larger size, have a higher settling velocity, are deposited at the bottom of the tube as a pellet, whereas all the bacteria cells are dispersed in the dense buffer.