OLI Grant: New Tools for Ocean Biology: Imaging In Flow to Improve the Resolution of Pump-During-Probe Measurements of Phytoplankton Photosynthetic Characteristics
Grant Funded: 2001
The phytoplankton comprise thousands of species of single-celled
plants ranging from cells just visible to the naked eye (a millimeter
or so) down to species a million times smaller. Together these microscopic
cells comprise the base of aquatic food chains and produce nearly
half the Earth's oxygen. Since the species composition of the phytoplankton
can determine the fate of higher trophic levels (such as the amount
and kinds of fish produced) and because phenomena such as harmful
algal blooms are associated with particular species, we need to
better understand how different kinds of cells respond to environmental
changes such as increased temperature or nutrient supply.
To investigate the regulation of phytoplankton species and cell types within the phytoplankton community, we have been developing assays and instrumentation to monitor the physiological condition of phytoplankton on a cell-by-cell basis. These assays are based on the fact that light energy absorbed by the chlorophyll in a phytoplankton cell can have different kinds of fates; it can be used in photosynthesis to make new plant material, or it can be "wasted" by being re-emitted (fluorescence). Our "Pump-during-probe" (PDP) fluorescence assays are based on the inverse relationship between photosynthesis and fluorescence at short time scales; they provide information about several photosynthetic parameters (which in turn can be related to nutritional status of the cells).
We have constructed a PDP flow cytometer (which assays cells as they pass through a laser beam focused on a flow cell), and a PDP microfluorometer (with which an operator manually locates and assays a cell under a microscope). With the microfluorometer, the operator can visually identify the phytoplankton cell being assayed (often to genus or even species), but the procedure is slow (~ 1 min per cell), so it is difficult to accumulate data on statistically significant numbers of cells from different groups in a sample. The flow cytometer can assay thousands of cells in a few minutes, but the resolution of the cell identification is crude; only light scattering (related to cell size) and pigment characteristics are available for classifying cells.
We therefore propose to add imaging capability to the PDP flow cytometer, with the goal of combining the speed of flow cytometry with the resolution of manual microscopy. In the PDP flow cytometer, imaging will take place immediately after a cell has been assayed. Cells will be categorized into taxonomic groups based on both flow cytometric signatures and automated image analysis, and a PDP induction curve will be constructed for each group. The addition of image analysis to PDP flow cytometry will open up a new avenue in studies of phytoplankton community regulation by allowing us to rapidly assess the physiological state of different taxa of phytoplankton (as opposed to simply cells of different size); such an assay could greatly facilitate investigations of the factors affecting species composition in the sea.
Our goal is to use imaging-in-flow to enhance the cell classification possible in the Pump-During-Probe (PDP) flow cytometer, combining the speed of flow cytometry with the resolution of manual microscopy. We have now added to the PDP flow cytometer an imaging system consisting of a CCD camera, a strobed LED illuminator, and a PC-based frame grabber. We began by purchasing a commercially-available imaging-in-flow system, but to meet our special requirements we found we had to replace or modify nearly all aspects of the system: We must use blue or infrared illumination (rather than the red light provided by the most common high-intensity LEDs) to avoid compromising the flow cytometric chlorophyll fluorescence measurements, and due to the high velocity of the cells through the flow cytometer, we are limited to exposures of 1 microsecond (an order of magnitude shorter than in the commercial instrument). This short exposure necessitated a more powerful LED driver and independent timing electronics. We also modified the image acquisition software to allow us to limit acquisition of images to those with defined scattering and fluorescence properties. The imaging system in various stages of development has been tested at sea during 2 cruises, and is now nearly optimized. To automate the classification of images, we are adapting the Video Plankton Recorder neural network software (which was kindly supplied by Cabell Davis), and we have written Matlab routines to integrate the images with the PDP flow cytometric measurements. We are just now beginning to test classification abilities and have successfully discriminated between a centric diatom, a dinoflagellate, and plastic beads.
Originally published: January 1, 2001