Analytical Techniques for Cell Fractions

XII. A Multiple-Cuvet Rotor for a New Microarialytical System1

Molecular Anatomy Program, Oak Ridge National Laboratory t2

Fast automatic analytical devices are required to handle the increasing number of different analyses desired on an ever-increasing number of samples in both biochemical research and clinical

Mechanized analytical systems3 have been divided into two major groups (1). Class I systems analyze a large number of samples for a single substance or activity, whereas those of class II analyze single samples for a number of different compounds or elements. Class I systems may be further subdivided into IA, in which many samples are analyzed simultaneously, and IB, in which they are done in sequence, A variety of hybrid systems is possible, some of which have been constructed. In this paper the problem of developing analyzers of class IA is considered, and the basic studies on the development of a new analytical system are presented.

1. Simultaneity. If a series of reactions done in parallel is carried to completion, it is not important that they all be started at exactly the same time. However, many colorimetric reactions are time-

1 Research conducted under the joint NIH-AEC Molecular Anatomy (MAN) Program supported by the National Institute of General Medical Sciences, the National Cancer Institute, the National Institute of Allergy and Infectious Diseases, and the U. S. Atomic Energy Commission.

2 Operated by Union Carbide Corporation Nuclear Division for the U. S.

3 Automation involves feedback control of a procss. Strictly speaking, nearly all analytical systems in present use are mechanized and use electronic control, sensing, and recording devices, but do not use feedback control. Hence, the words "automated" and "automation" do not apply to them despite previous dependent, e.g., with the color density increasing slowly, perhaps plateauing, and subsequently decreasing. In such instances, and where reaction rates are measured, as is the case with enzymatic reactions, it is of advantage to have all of a set of reactions start at as nearly the same time as possible, provided that all optical measurements can also be carried out over a short time span. If both the initiation of the reaction (mixing) and reading can be done in a few seconds or less, the reading may often be done very early in the course of a reaction. This is especially true when the set of samples is large enough to include a number of standards.

2. Scaling. The number of samples which may be analyzed with class IA analyzers should, in principle, be continuously variable. Some existing analyzers are adequate for larger loads but are not efficient for only a few (say, 1-10) samples. Small numbers of samples are usually done by manual methods. Hence, clinical and research analysts must often learn and use two different procedures for the same type of analysis, the choice depending on the volume of the analytical load. Obviously, methods that apply efficiently to both large and small sets of sample are desirable.

3. Engagement. If a sufficiently large number of analyses can be done in one apparatus and in a short time, the analyst can give full attention to the entire process. However, when data output is slow or a number of analyzers must be operated at the same time, the analyst's time is divided between them, with related tasks interleaved throughout the day. Because data output rate is slow, realtime computing is not economical, and data reduction is often postponed. Ideally, a set of analyses should be completed in the time required for the operator to determine whether a system of class IB is working properly. While truly automated variants which require no operator attention may ultimately be constructed, the potential for complete engagement (i.e., direct observation of reactions in real time) should still exist so that errors in dilution, reagent preparation, or procedure, or instrumental failure may be readily

4. Sample and reagent volumes. Many analytical methods use enzymes and other expensive reagents. To minimize the reagent requirement, and to allow relatively small samples to be used, the techniques developed should be in the microliter range with sample and reagent volumes summing to between 0.2 and 3 ml. (It will be evident that the methods to be described are applicable to smaller

5. Sample and reagent measurement. Since discrete measurement is employed (as opposed to flow-ratio measurement and control),

ROTOR FOR A MICROANALYTICAL SYSTEM 547

the system developed should be adaptable to use with a variety of

6. Sequencing of analytical elements. If one considers each step in an analysis as an element, the following are the most common

(b) Measured volume transfer (two or more).

(d) Incubation for a measured time at a known temperature.

(e) Precipitation (filtration or centrifugation).

(/) Extraction or washing of a precipitate.

(g) Extraction of an aqueous solution with an organic solvent.

(h) Absorbancy or other physical measurement, including radio-

The elements employed and their sequence may differ in different analyses. To be generally useful, an analytical system must be able to include most or all of these elements.

7. Reaction monitoring. The progress of all reactions being carried out simultaneously should be visible, A permanent chart record is not needed when reactions are carried out rapidly, if all results are monitored as they occur, and where data is reduced in real time.

8. Data reduction. Complete data reduction should be possible in a matter of seconds to minutes, and the computing system used should be programmed to handle both rate determinations and

9. Simplicity„ To be generally useful, the basic techniques must be simple, but should be applicable to a series of systems of varying complexity, ranging from a minimal-cost unit for teaching purposes to large units for handling the analytical load of a large hospital,

In this and subsequent papers, the development of systems designed to fulfil these requirements is described. The present paper is concerned with demonstrating the feasibility of using a multiple-cuvet rotor to measure the absorbancies of a multiplicity of solutions simultaneously by using centrifugal force to load the cuvets, and variation in angular momentum to mix solutions.

The principle of double-beam spectrophotometry, in which the absorbancies of a reference solution and a sample are inter-compared, either continuously or over time intervals short relative to the rate of ehange in intensity of the light source, is well known

To apply the same principle effectively to a larger number of samples requires that either they, or the light beam, move rapidly, one with respect to the other. We have chosen to move the samples rapidly past the light beam by using cuvet rotors. The centrifugal field inherent in this design has the advantage of providing the force for moving samples and reagents into the cuvets, for mixing them, for removing air bubbles, for sedimenting particulate matter, and for separating liquid phases. The electronic signal generated photo-electrically may be conveniently and continuously displayed on a cathode-ray tube and recorded photographically. In addition, the individual cuvet readings are made at a rate compatible with rapid electronic averaging of readings obtained over a short time period

Cuvette Rotors. A flat 40 cuvet rotor (rotor G-I), with small, cylindrical test tubes as cuvets, was used in orienting studies. For more definitive studies, a 15 cuvet rotor (G-II) was designed. The G-II cuvets were formed by compressing a ring of — i*1- Teflon sheet between 2 discs of V2 in. thick Pyrex. The Teflon section contains 15 round-bottom slots opening toward the center; the Teflon and upper Pyrex discs have a large hole in the center. Stainless-steel flanges above and below were connected with bolts to compress all parts together. A completed G-II rotor is shown in Figure 1, and a disassembled one in Figure 2. The circular apertures over the cuvets are 14 in. in diameter, and the light beam is arranged to give a short, flat region on the end of the peaks as shown in Figure 3. When the rotor is spinning, 200 X of solution is sufficient to fill the portion of the cuvet visible through the circular apertures in the stainless-steel end plates. The cuvet dimensions and volumes are chosen so that the rotor may be carefully brought to rest and re-accelerated without having the contents of the cuvets mix with each other when a total volume of 200-250 A per cuvet is used. If larger volumes are to be decelerated to rest, longer cuvets must be

Photoelectric sensors mounted next to the rotor edge provide synchronizing signals for the sweep circuits.

Sample and Reagent Addition. Samples and reagents are moved into the cuvets by centrifugal force. Since this may be done over a short period of time, all reactions start essentially together and may be followed continuously on the oscilloscope. Sample and reagent discs (Fig. 4) allow the sample and two reagents to be loaded into separate depressions designed not to allow fluids to mix at rest, but which all drain to the edge into the proper cuvets in a

ROTOR FOR A MICROANALYTICAL SYSTEM 549

ROTOR FOR A MICROANALYTICAL SYSTEM 549

Fig. 1. G-II cuvet rotor: (A) euvet rotor; (B) photomultiplier housing; (C) filter holder; (D) light source with diaphragm; (E) drive motor from I EC clinical centrifuge.

centrifugal field. Connections to the edge are through small capillary holes or past sloping surfaces which prevent mixing before spinning, but allow free horizontal drainage during rotation. The transfer discs may also be adapted to hold the transfer tubes previously described (1), or small, commercially available disposable microliter pipets. It is evident that these devices allow single or multiple addition reactions to be used, or reactions in which an incubation period occurs between two additions. It is also evident that precipitates formed during a reaction may be moved out of the optical path by centrifugal force, allowing the absorbancies of a clear supernatant to be measured.

Mixing. In many instances, especially where zonal centrifuge fractions are being analyzed, the reagents and the sample may differ considerably in density and viscosity. An effective means for achieving rapid mixing is therefore required.

In the cuvet rotor described, the radii to the top and the bottom of a 250 A reaction volume differ by approximately 7 mm, giving,a ratio of tangential velocities of 1.05. If the rotor is rapidly accel-

550 NORMAN G. ANDERSON

Fie. 2. Disassembled G-II rotor: (A) lower rotor housing; (B) lower gasket; (C) lower Pyrex plate; (D) Teflon cuvet spacer; (E) upper Pyrex window ring; (F) upper Teflon gasket; (G) upper stainless-steel end plate.

Fig. 3. Oscilloscope tracing of two peaks indicating flat tip. Baseline at top indicates 0% transmission (infinite optical density) line.

ROTOR FOR A MICROANALYTICAL SYSTEM 551

ROTOR FOR A MICROANALYTICAL SYSTEM 551

Fig. 4. Sample and reagent transfer and loading discs: (A, B) dises with cavities with top tilting toward the axis hut connected by capillaries draining the bottom of the cavities to the edge; (C) disc having two concentric sets of fluid-holding cavities tilting outward so that fluid rises along cavity wall to drain through peripheral holes; (D) Teflon center piece of G-II rotor showing how loading disc fits cuvets in rotor; (E) cover plate used to cover disc and expose only cavities being filled.

Fig. 4. Sample and reagent transfer and loading discs: (A, B) dises with cavities with top tilting toward the axis hut connected by capillaries draining the bottom of the cavities to the edge; (C) disc having two concentric sets of fluid-holding cavities tilting outward so that fluid rises along cavity wall to drain through peripheral holes; (D) Teflon center piece of G-II rotor showing how loading disc fits cuvets in rotor; (E) cover plate used to cover disc and expose only cavities being filled.

erated and decelerated during fluid addition, effective circular flow is established in the cuvet because of the velocity differences between the centripetal and centrifugal surfaces of the fluid. In practice, therefore, the rotor is accelerated rapidly to transfer fluid, is rapidly decelerated (but not brought to rest), and then reaccel-erated to the speed used for observation. The experiments described in subsequent sections demonstrate that adequate mixing may be obtained by this method within 10-15 sec after the rotor is started using protein samples in distilled water and a biuret reagent. With concentrated sucrose solutions, more rapid change in speed than could be effected with the present prototype will be required.4

4 In subsequent studies the cuvets have been designed as syphons with a restricted entrance. Mixing may be achieved by bubbling air back through the cuvet syphon during rotation at slow speed. Detailed studies on mixing rates are included in a subsequent paper.

Cleaning. The prototypes used in these experiments have been hand-cleaned by using a fine stream of distilled water continuously removed by suction, followed by high-pressure air. An automatic

The studies recorded here are concerned with proof-of-principle and with more precise definition of problems remaining to be solved.

Leakage. Leakage from the cuvets to the edge during rotation was checked by filling the cuvets completely with water during rotation and observing the position of the fluid surface stroboscopically during rotation at 2000 rpm over an extended period. No leakage was observed. To determine whether leakage occurred between cuvets, alternate cuvets were filled with water and similarly observed. No

The possibility exists that fluid from the loading disc does not pass directly and quantitatively into the proper cuvet, but leaks laterally. A heavily stained solution of bovine serum albumin (BSA), 250 A per cuvet, was placed in the even-numbered positions in the loading disc and moved by centrifugal force into the cuvets. The results, compared with a pattern obtained with all the cuvets full of water, are shown in Figure 5. It is seen that almost no light passed through the evennumbered cuvets, but full transmission was observed through the odd-numbered ones. The white Teflon-bottom inner surfaces of the latter cuvets were carefully examined by using stroboscopic illumination to verify that no leakage had occurred. (Note that in the oscilloscope patterns the lines connecting the peaks indicate 0% transmission or infinite absorbancy. The peaks therefore appear to be inverted when compared with those obtained with other analytical systems; however, the peak tips are all at the

Drainage from Loading Disc. Drainage from the loading disc was studied by using heavily stained BSA solutions. After spinning at 2000 rpm, only a very few traces of blue color could be observed in the disc. This was thought to be due to very small irregularities in the Teflon due to machining. The volumes involved appeared to be less than 1 A. Further quantitative studies on solution transfer as a function of angular velocity, radius, disc design, and solution density, viscosity, and surface tension are indicated.

Calibration. The path length of the cuvets was measured directly by using an electronically indicating micrometer. The rotor cuvet end plates are not exactly parallel, but vary in a sine wave pattern as shown in Figure 6. This was confirmed by using identical dyed

ROTOR FOR A MICROANALYTICAL SYSTEM 553

ROTOR FOR A MICROANALYTICAL SYSTEM 553

Fig. 5. Test for leakage between cuvets. (A) oscilloscope pattern with cuvets containing water. The ordinate for the tracing ranges from 0% transmission at the top to 100% transmission at the bottom of the trace, Cuvet numbers are in order from left to right. (B) Tracing seen when a solution of dyed BSA was added to even-numbered cuvets. A 550 nm interference filter was used.

Fig. 5. Test for leakage between cuvets. (A) oscilloscope pattern with cuvets containing water. The ordinate for the tracing ranges from 0% transmission at the top to 100% transmission at the bottom of the trace, Cuvet numbers are in order from left to right. (B) Tracing seen when a solution of dyed BSA was added to even-numbered cuvets. A 550 nm interference filter was used.

protein solutions in each cuvet except the first, which contained water (Fig. 7B).

It is not desirable to depend on precision construction of cuvets to define path length or to assume that the blank (water) absorb-ancies of all cuvets are equal or constant. Instead blank absorban-cies and absorbancies with standard solutions to determine path length should be redetermined at intervals when high precision is required. The values obtained are incorporated in the final calculations.

To see whether reporducible curves could be obtained with standard solutions, a solution containing 1.5 gm of crystalline BSA and 15 mg of bromphenol blue (BPB) in 100 ml was diluted to give

554 NORMAN G. ANDERSON

Was this article helpful?

0 0

Post a comment