Impedance spectroscopy flow cytometry: on-chip label-free cell differentiation.

BACKGROUND: The microfabricated impedance spectroscopy flow cytometer used in this study permits rapid dielectric characterization of a cell population with a simple microfluidic channel. Impedance measurements over a wide frequency range provide information on cell size, membrane capacitance, and cytoplasm conductivity as a function of frequency. The amplitude, opacity, and phase information can be used for discrimination between different cell populations without the use of cell markers. METHODS: Polystyrene beads, red blood cells (RBCs), ghosts, and RBCs fixed in glutaraldehyde were passed through a microfabricated flow cytometer and measured individually by using two simultaneously applied discrete frequencies. The cells were characterized at 1,000 per minute in the frequency range of 350 kHz to 20 MHz. RESULTS: Cell size was easily measured with submicron accuracy. Polystyrene beads and RBCs were differentiated using opacity. RBCs and ghosts were differentiated using phase information, whereas RBCs and fixed RBCs were differentiated using opacity. RBCs fixed using increasing concentrations of glutaraldehyde showed increasing opacity. This increased opacity was linked to decreased cytoplasm conductivity and decreased membrane capacitance, both resulting from protein cross-linking. CONCLUSIONS: This work presents label-free differentiation of cells in an on-chip flow cytometer based on impedance spectroscopy, which will be a powerful tool for cell characterization.     

A high buoyant density fraction in mammalian DNA: Characterization and impact on the detection of heteroduplex DNA

DNA from both Chinese hamster ovary (CHO) cells and human fibroblasts contains a high buoyant density fraction. This fraction of DNA was purified from CHO cells and characterized. Compared with mainband CHO DNA, this high buoyant density DNA binds more of a GC-specific dye, actinomycin D (actD), and less of an AT-specific dye, netropsin, which suggests that its increased density is due to an increase in clusters of GC base pairs. The detection of heteroduplex DNA which has been hypothesized to occur during sister chromatid exchange formation is considerably complicated by the presence of this high density DNA. Experiments to detect heteroduplex DNA in CHO cells, using actD to shift the position of the high density DNA, did not reveal any underlying heteroduplex material, thus placing an upper limit on the size of the hypothesized heteroduplex regions. Experiments with both CHO cells and with human fibroblasts indicated that the amount of the high buoyant density DNA did not consistently increase when the sister chromatid exchange frequency increased.     

THE SEPARATION OF DIFFERENT CELL CLASSES FROM LYMPHOID ORGANS : III. The Purification of Erythroid Cells by pH-Induced Density Changes

1. Mammalian erythrocytes swell as the pH of the isotonic suspending medium is lowered, as a direct consequence of the specialized permeability properties of the erythrocyte membrane. Lymphocytes and granulocytes from a variety of sources did not exhibit this property. 2. The behaviour of mouse bone marrow erythroid cells at various stages of differentiation was studied by using a change in buoyant density with pH as an index of swelling. The ability to swell with a pH drop was acquired while the cell was still nucleated. All non-nucleated cells showed swelling. Most small erythroblasts shared this property, whereas most large erythroblasts did not. 3. The density shift with pH was used to provide a purification scheme specific for erythroid cells. The bone marrow cells were first centrifuged to equilibrium in an isotonic albumin density gradient at neutral pH. Regions of the gradient containing the erythroid cells were collected, and the cells were recovered and redistributed in an albumin gradient at acid pH. The erythroid cells showed a specific density shift which removed them from contaminants. Preparations containing 90–97% erythroblasts were obtained by this technique. 4. Differentiation within the erythroid series was accompanied by a general increase in cell buoyant density at neutral pH. This density increase may have been a discontinuous process, since erythroid cells appeared to form a number of density peaks. 5. The pH shift technique, in association with established density distribution and sedimentation velocity procedures, provides a range of cell separation techniques for biological or biochemical studies of erythroid cell differentiation in the complex cell mixtures in bone marrow or spleen.     

Changes in buoyant density and cell size of Escherichia coli in response to osmotic shocks.

The buoyant density of Escherichia coli was shown to be related to the osmolarity of the growth medium. This was true whether the osmolarity was adjusted with either NaCl or sucrose. When cells were grown at one osmolarity and shocked to another osmolarity, their buoyant density adjusted to nearly suit the new osmolarity. When cells were subjected to hyperosmotic shock, they became denser than expected. When cells were subjected to hypoosmotic shock they occasionally undershot the new projected density, but the undershoot was not as dramatic as the overshoot seen with hyperosmotic shocks. Shrinkage and swelling of the cells in response to osmotic shocks could account for the change in their buoyant density. The changes in cell size after osmotic shocks were measured by two independent methods. The first method measured cell size with a Coulter Counter, and the second method measured cell size by stereologic analysis of Nomarski light micrographs. Both methods gave qualitatively similar results and showed the cells to be flexible. The maximum swelling recorded was 23% of the original cell volume, while the maximum shrinkage observed was 33%.     

Buoyant density constancy during the cell cycle of Escherichia coli

Cell buoyant densities were determined in exponentially growing cultures of Escherichia coli B/r NC32 and E. coli K-12 PAT84 by equilibrium centrifugation in Percoll gradients. Distributions within density bands were measured as viable cells or total numbers of cells. At all growth rates, buoyant densities had narrow normal distributions with essentially the same value for the coefficient of variation, 0.15%. When the density distributions were determined in Ficoll gradients, they were more than twice as broad, but this increased variability was associated with the binding of Ficoll to the bacteria. Mean cell volumes and cell lengths were independent of cell densities in Percoll bands, within experimental errors, both in slowly and in rapidly growing cultures. Buoyant densities of cells separated by size, and therefore by age, in sucrose gradients also were observed to be independent of age. The results make unlikely any stepwise change in mean buoyant density of 0.1% or more during the cycle. These results also make it unlikely that signaling functions for cell division or for other cell cycle events are provided by density variations.     

Increase in cell mass during the division cycle of Escherichia coli B/rA.

Increase in the mean cell mass of undivided cells was determined during the division cycle of Escherichia coli B/rA. Cell buoyant densities during the division cycle were determined after cells from an exponentially growing culture were separated by size. The buoyant densities of these cells were essentially independent of cell age, with a mean value of 1.094 g ml-1. Mean cell volume and buoyant density were also determined during synchronous growth in two different media, which provided doubling times of 40 and 25 min. Cell volume and mass increased linearly at both growth rates, as buoyant density did not vary significantly. The results are consistent with only one of the three major models of cell growth, linear growth, which specifies that the rate of increase in cell mass is constant throughout the division cycle.     

Buoyant density, growth rate, and the cell cycle in Streptococcus faecium.

The buoyant density in rapidly growing Streptococcus faecium 9790 cells varies over the cell cycle, in contrast to the density in Escherichia coli. Buoyant density in S. faecium was measured by using Percoll (Pharmacia Fine Chemicals, Piscataway, N.J.) density gradients. We found that the mean and coefficient of variation of the population density increased with growth rate; and within a population, the mean cell volume, which was measured electronically, increased with density. These results were compared with electron microscopic measurements of the size distributions of cell wall growth sites within each fraction of the density gradient. As the density increased within a population, the frequency of large cells increased and the frequency of newly initiated cell wall growth sites increased. These effects were more marked as the growth rate increased. Next, these data were regrouped by cell size by using the size of the central growth site as an index of cell cycle stage. Each frequency value was weighted by the proportion of the population represented by that density fraction. Then, the average buoyant density was calculated for each value of cell size. In all cell populations, the density decreased and then increased as the central site enlarged. Peripheral growth sites were initiated as density reached a maximum. At faster growth rates, density increased more steeply, and new peripheral growth sites opened up at a higher frequency. We suggest that the rate at which density increases during the cell cycle correlates with the initiation of new cell wall growth sites.     

Relationship between changes in buoyant density and formation of new sites of cell wall growth in cultures of streptococci (Enterococcus hirae ATCC 9790) undergoing a nutritional shift-up.

When the glutamate concentration of cultures of Enterococcus hirae was raised from 20 to 300 micrograms/ml, the mass doubling time decreased from ca. 85 to 45 min in 9 min, but balanced growth was not reestablished for 30 to 40 min. During the unbalanced period of growth, RNA and protein synthesis proceeded more rapidly than did peptidoglycan synthesis, buoyant density increased from ca. 1.1024 to 1.1075 g/ml, and the rate of formation of new cell wall growth sites transitorily accelerated above the new growth rate. When studied as a function of cell size, all cultures showed buoyant density to decrease around cell separation, increase as cells increased in size, and then plateau when cells reached large volumes. Greater increases in buoyant density as a function of cell size were seen after shift-up, with the greatest increases observed at 15 to 20 min after shift-up, when the rate of formation of new sites was also maximal. In a population of cells examined by electron microscopy 15 min after shift-up, buoyant density and the frequency of cells with new sites increased as old sites approached the size of two poles. These data were consistent with a model whereby buoyant density increases in the terminal stages of the cell cycle when the surface grows slower than the cytoplasm. The greater the difference in the rates of inside to outside growth, the greater the increase in buoyant density and the more frequently new sites will be initiated.     

Buoyant density studies of several mecillinam-resistant and division mutants of Escherichia coli.

The buoyant density of wild-type Escherichia coli cells has previously been reported not to vary with growth rate and cell size or age. In the present report we confirm these findings, using Percoll gradients, and analyze the recently described lov mutant, which was selected for its resistance to mecillinam and has been suggested to be affected in the coordination between mass growth and envelope synthesis. The average buoyant density of lov mutant cells was significantly lower than that of wild-type cells. Similarly, the buoyant density of wild-type cells decreased in the presence of mecillinam. The density of the lov mutant, like that of the wild type, was invariant over a 2.8-fold range in growth rate. In this range, however, the average cell volume was also constant. Analysis of buoyant density as a function of cell volume in individual cultures revealed that smaller (newborn) lov mutant cells had higher density than larger (old) cells; however, the density of the small cells never approached that of the wild-type cells, whose density was independent of cell size (age). A pattern similar to that of lov mutant cells was observed in cells carrying the mecillinam-resistant mutations pbpA(Ts) and rodA(Ts) and the division mutation ftsI(Ts) at nonpermissive temperatures as well as in wild-type cells treated with mecillinam, but not in mecillinam-resistant crp or cya mutants.     

Density fluctuation during the cell cycle in the defective vacuolar morphology mutants of Saccharomyces cerevisiae.

The buoyant densities of the yeast cells of defective vacuolar morphology mutants were examined by equilibrium sedimentation centrifugation in a Percoll density gradient. These vacuoleless mutants also show density fluctuation as wild-type cells during the cell cycle. This suggests that morphological changes of the vacuole are not related to cyclic density fluctuation in Saccharomyces cerevisiae.     

Buoyant density constancy of Schizosaccharomyces pombe cells.

Buoyant densities of cells from exponentially growing cultures of the fission yeast Schizosaccharomyces pombe 972h- with division rates from 0.14 to 0.5 per h were determined by equilibrium centrifugation in Percoll gradients. Buoyant densities were independent of growth rate, with an average value (+/- standard error) of 1.0945 (+/- 0.00037) g/ml. When cells from these cultures were separated by size, mean cell volumes were independent of buoyant density, indicating that buoyant densities also were independent of cell age during the division cycle. These results support the suggestion that most or all kinds of cells that divide by equatorial fission may have similar, evolutionarily conserved mechanisms for regulation of buoyant density.     

Density-dependent sorting of physiologically different cells of Vibrio parahaemolyticus

A pure bacterial culture is composed of clonal cells in different physiological states. Separation of those subpopulations is critical for further characterization and for understanding various processes in the cultured cells. We used density-dependent cell sorting with Percoll to separate subpopulations from cultures of a marine bacterium, Vibrio parahaemolyticus. Cells from cultures in the exponential and stationary phases were fractionated according to their buoyant density, and their culturability and ability to maintain culturability under low-temperature and low-nutrient stress (stress resistance) were determined. The buoyant density of the major portion of the cells decreased with culture age. The culturability of stationary-phase cells increased with increasing buoyant density, but that of exponential-phase cells did not. Stress resistance decreased with increasing buoyant density regardless of the growth phase. The results indicate that density-dependent cell sorting is useful for separating subpopulations of different culturabilities and stress resistances. We expect that this method will be a powerful tool for analyzing cells in various physiological states, such as the viable but nonculturable state.     

Density-Dependent Sorting of Physiologically Different Cells of Vibrio parahaemolyticus

A pure bacterial culture is composed of clonal cells in different physiological states. Separation of those subpopulations is critical for further characterization and for understanding various processes in the cultured cells. We used density-dependent cell sorting with Percoll to separate subpopulations from cultures of a marine bacterium, Vibrio parahaemolyticus. Cells from cultures in the exponential and stationary phases were fractionated according to their buoyant density, and their culturability and ability to maintain culturability under low-temperature and low-nutrient stress (stress resistance) were determined. The buoyant density of the major portion of the cells decreased with culture age. The culturability of stationary-phase cells increased with increasing buoyant density, but that of exponential-phase cells did not. Stress resistance decreased with increasing buoyant density regardless of the growth phase. The results indicate that density-dependent cell sorting is useful for separating subpopulations of different culturabilities and stress resistances. We expect that this method will be a powerful tool for analyzing cells in various physiological states, such as the viable but nonculturable state.     

Replication of bromodeoxyuridylate-substituted mitochondrial DNA in yeast.

The DNA of several strains of Saccharomyces cerevisiae was labeled by growing the culture in medium supplemented with thymidylate and bromodeoxyuridylate. It was thus possible to follow the course of mitochondrial DNA replication in density shift experiments by determining the buoyant density distribution of unreplicated and replicated DNAs in analytical CsCl gradients. DNA replication was followed for three generations after transfer of cultures from light medium to heavy medium and heavy medium to light medium. Under both conditions, the density shifts observed for mitochondrial DNA were those expected for semiconservative, nondispersive replication. This was further confirmed by analysis of the buoyant density of alkali-denatured hybrid mitochondrial DNA. With this method, no significant recombination between replicated and unreplicated DNA was detected after three generations of growth.     

Buoyant density variation during the cell cycle of Saccharomyces cerevisiae

Cell buoyant densities of the budding yeast Saccharomyces cerevisiae were determined for rapidly growing asynchronous and synchronous cultures by equilibrium sedimentation in Percoll gradients. The average cell density in exponentially growing cultures was 1.1126 g/ml, with a range of density variation of 0.010 g/ml. Densities were highest for cells with buds about one-fourth the diameter of their mother cells and lowest when bud diameters were about the same as their mother cells. In synchronous cultures inoculated from the least-dense cells, there was no observable perturbation of cell growth: cell numbers increased without lag, and the doubling time (66 min) was the same as that for the parent culture. Starting from a low value at the beginning of the cycle, cell buoyant density oscillated between a maximum density near midcycle (0.4 generations) and a minimum near the end of the cycle (0.9 generations). The pattern of cyclic variation of buoyant density was quantitatively determined from density measurements for five cell classes, which were categorized by bud diameter. The observed variation in buoyant density during the cell cycle of S. cerevisiae contrasts sharply with the constancy in buoyant density observed for cells of Escherichia coli, Chinese hamster cells, and three murine cell lines.     

Buoyant density analysis of myeloid colony-forming cells in germfree and conventional mice.

Granulocyte-macrophage colony-forming cells (CFUc), in the bone marrow of germfree and conventioal CBA mice, were compared quantitatively and qualitatively. Cells were separated on the basis of their buoyant density by equilibrium centrifugation in continuous albumin density gradients. CFUc in the density subpopulations were detected by culture in agar containing three different types of colony stimulating factor (CSF). The sources of the CSF were post-endotoxin mouse serum (CSFES), mouse lung conditioned medium (CSFMLCM) and human urine (CSFHU). Mice were removed from the germfree environment and the buoyant density status of their CFUc was examined 1, 4 and 8 weeks later. No difference was found between germfree and conventional mice in the number of nucleated cells per femur or in their modal density. Neither was the number of CFUc per femur different. The cell cycle status of CFUc, as determined by the thymidine suicide technique was not significantly different. Functional heterogeneity was found among the density subpopulations for both groups of mice. This depended on the type of CSF. The density distribution of CFUc was significantly different in germfree mice. There were proportionately more low density CFUc. The mean modal density of CFUc under CSFES stimulation was less by 0.0045 g/cm3 in germfree mice. The removal of mice from the germfree environment resulted in a shift of the distribution to higher densities. The trend was towards the conventional situation. The significance of the buoyant density status of CFUc is discussed.     

Buoyant density of Escherichia coli is determined solely by the osmolarity of the culture medium

In previous studies, we had shown that the buoyant density ofEscherichia coli is determined by the osmolarity of the growth medium by varying the osmolarity of the medium with NaCl or sucrose. However, the buoyant density of the cells always exceeded that of the growth medium. Here we determined the effect of medium with a buoyant density greater than the expected buoyant density of cells by adding Nycodenz to Luria broth. Percoll gradients of cells were analyzed by laser light scattering. The buoyant density for 125- and 375-mOsM-grown cells was 0.002 g/ml and 0.003 g/ml more, respectively, for cells grown in the presence of Nycodenz than those grown without Nycodenz, while the buoyant density of 250-mOsM-grown cells was 0.005 g/ml less for cells grown in the presence of Nycodenz than those grown without Nycodenz. Cells grown in 500-mOsM medium with or without Nycodenz had the same buoyant density. the buoyant density of cultures grown in defined medium was the same as those grown in rich medium, with only the medium osmolarity correlating to buoyant density. We conclude from these experiments that neither buoyant density nor chemical make-up of the medium determines the buoyant density of cells grown in that medium. Only the medium osmolarity determines cell buoyant density, suggesting thatE. coli has no mechanisms to sense buoyant density.     

Water and ion balance in rat thymocytes under apoptosis induced with dexamethasone or etoposide. Ion-osmotic model of cell volume decrease

Cell ion and water balance was studied with respect to analysis of the osmotic model of apoptotic volume decrease (AVD) in rat thymocytes under dexamethasone (1 microM, 4-6 h) or etoposide (50 microM, 5 h) treatment. Intracellular water content was determined by measurement of cell buoyant density in continuous Percoll gradient, while intracellular potassium and sodium contents were determined by flame emission analysis. Apoptosis was verified by an increase in cell buoyant density, fluorescence of cells stained with Acridine orange and Ethidium bromide (flow cytometry), by changes in the cell cycle and the appearance of sub-diploid peak in the DNA histogram (flow cytometry), and by a decrease in cell size examined with light microscope. A separate fraction of dense cells with reduced size was found to appear after dexamethasone or etoposide treatment. This fraction was considered as apoptotic. An increase in buoyant density of apoptotic cells corresponded to a decrease in cell water content. In apoptotic cells vs. cells with normal buoyant density, the intracellular potassium content was lower, but sodium content was higher. The sum of potassium and sodium contents was lower in apoptotic cells. Taken into account the loss of anions, associated with the loss of cations, the bulk decrease in ions content has been sufficient to be accounted for cell volume decrease on the basis of the ion-osmotic model.     

BUOYANT DENSITY ANALYSIS OF MYELOID COLONY-FORMING CELLS IN GERMFREE AND CONVENTIONAL MICE

Granulocyte-macrophage colony-forming cells (CFUc), in the bone marrow of germfree and conventional CBA mice, were compared quantitatively and qualitatively. Cells were separated on the basis of their buoyant density by equilibrium centrifugation in continuous albumin density gradients. CFUc in the density subpopulations were detected by culture in agar containing three different types of colony stimulating factor (CSF). The sources of the CSF were post-endotoxin mouse serum (CSF_ES), mouse lung conditioned medium (CSF_MLCM) and human urine (CSF_HU). Mice were removed from the germfree environment and the buoyant density status of their CFUc was examined 1, 4 and 8 weeks later. No difference was found between germfree and conventional mice in the number of nucleated cells per femur or in their modal density. Neither was the number of CFUc per femur different. The cell cycle status of CFUc, as determined by the thymidine suicide technique was not significantly different. Functional heterogeneity was found among the density subpopulations for both groups of mice. This depended on the type of CSF. The density distribution of CFUc was significantly different in germfree mice. There were proportionately more low density CFUc. The mean modal density of CFUc under CSF_ES stimulation was less by 0.0045 g/cm³ in germfree mice. The removal of mice from the germfree environment resulted in a shift of the distribution to higher densities. The trend was towards the conventional situation. The significance of the buoyant density status of CFUc is discussed.