Your browser is out-of-date!

Update your browser to view this website correctly. Update my browser now

×

Discipline
Biological
Keywords
Plankton
High Throughput Screening
Observation Type
Standalone
Nature
Orphan Data
Submitted
Jan 21st, 2020
Published
Jun 24th, 2020
  • Abstract

    Plankton provides an essential foundation for life on Earth, supplying most of the breathable oxygen, carbon sequestration, and larvae nutrition. Toxic chemicals introduced into the environment pose a potential danger to plankton and the ecosystem. Traditional plankton chemical toxicity assays measure the concentration required to cause death. Sublethal doses that affect plankton activities like foraging and escaping predators can have a cascading impact on ecosystems. We have developed a high-throughput device to measure the sublethal effect of chemicals on the plankton rate of movement. The device automatically creates a series of chemical dilutions, subjects each concentration to a group of plankton, and calculates the average group movement speed. We tested the device on groups of Stentor coeruleus (n=3 to 26, mean=10) with acetic acid dilutions (43 to 25,000 ppm) and measured a declining trend in average speed with increasing sublethal concentrations (170 to 502 ppm).

  • Figure
  • Introduction

    Plankton are omnipresent in freshwater, seas, and oceans. They form the backbone of aquatic ecosystems and food chains, providing most of the planets' breathable oxygen, carbon sequestration, and larval nutrition. Plankton are any organism that lives in water and are unable to swim against the current. This broad definition includes bacteria, phytoplankton, and zooplankton. Many plankton have cilia or flagella to propel themselves to forage for food and avoid predators. Due to the interconnectedness of organisms in an ecosystem, disturbing one plankton species can cause a major disruption in an ecosystem, for example, harmful algae blooms.

    Anthropogenic activities, including agricultural runoff, consumer, and industrial waste, introduce toxic chemicals into waterways. Only a small portion of a large number of chemicals in use are tested for plankton toxicity (https://cen.acs.org/articles/95/i9/chemicals-use-today.html). The standard acute toxicity reference is LC50, the chemical concentration required to kill 50% of a group of test organism after a short exposure, typically 24 to 96 h. Chemicals that interfere with a biological process tend to have a LC50 response that is strongly species-specific. For example, cadmium has a LC50 toxicity from 2 ug/L to 4.6 g/L depending on the aquatic species.

    Chronic toxicity tests observe sublethal impacts over a long duration (>10% of the organism’s life span) to allow the toxic effect to accumulate. Chronic exposure may result in alterations in reproduction, sex ratio, morphology, feeding and predator response, effects that can impact the community and ecosystem. Many of these symptoms appear at concentrations lower than LC50. For example, for Daphnia, the carbaryl insecticide has a LC50 of 12.76 ppb for a 24 h exposure, while a 50 min exposure at 5 ppb reduces the hop rate by 25% and a 21-day exposure at 0.4 ppb reduces molts by 11%.

    Recently microfluidic devices have been developed to perform high-throughput plankton research. By fabricating gradient generators, multiple chambers with different chemical concentrations can be constructed to test for acute toxicity. Multiple zones have been fabricated to accommodate plankton of different sizes. To accommodate the 4 most common plankton cited in toxicity studies, the chambers must accommodate a size range from 0.2 mm to 30 mm (Daphnia, a cladoceran, and Hydra, a cnidarian, respectively), a dimensional span outside of the range of typical microfluid devices.

  • Objective

    The objective is to design and build a high-throughput device to measure chemical toxicity on a variety of aquatic organisms over a wide range of concentrations by detecting sublethal movement speed during a short-exposure time (one-minute). To demonstrate the operation, the device tests the effect of acetic acid (from household vinegar) on the movement speed of groups of Stentor, one of the largest single-cell plankton.

  • Results & Discussion

    The average speed of Stentor (n=3 to 26, mean=10), exposed to a concentration of acetic acid for 1 min, is presented in figure 1E. A group is exposed to one concentration then discarded, then a new group is subjected to a lower concentration. Acetic acid concentrations of 624 ppm and higher are fatal (all group members died). Concentrations of 211 ppm and lower were sublethal (all group members survived). At concentrations higher than 170 ppm speed decreases until a fatal concentration is reached. LC50 occurred at 502 ppm. There is considerable speed variation in sublethal concentrations below 170 ppm, with several confounding factors to consider. Stentor exhibit several modes of motion (Fig. 1D) including traveling long and fast in a straight line, slower travel with a pronounced corkscrew trail, precessing about an anchor point and moving in the Z direction (towards the camera), an axis not captured. The number of Stentor in a group can impact the average speed; small groups are more sensitive to intra-member variation while Stentor in large groups may slow down and change direction to avoid collisions.

    For each chemical concentration, 1 mL of Stentor was injected into the chamber, providing Stentor and the diluent (spring water). If an injection did not contain any Stentor, there would be no reading for that concentration, as the dilution would have been performed, and injecting more Stentor would create a new dilution. We relied on the pre-filtering to achieve a minimum concentration of 3 Stentor per mL and bubbling to mix the Stentor before each injection. The minimum target was reached, however, there was a significant range of membership (3 to 26) per injection.

    The device is intended to work with any plankton between 0.2 mm and 3 mm, constrained by the chamber diameter and tubing inner diameter (3 mm). We use the device for three purposes; species selection (i.e. Stentor purification), chemical toxicity assay (i.e. reporting average motion speed at sublethal concentrations), and investigating cell functions (ongoing research not reported here). We selected Stentor to test this device based on our interest in evaluating Stentor as a biosensor and as a model organism to investigate how cells function and interact with their environment. It is large plankton (>0.5 mm), simplifying observation and handling, and a single cell, so there is no multi-cellular interaction to consider. Instead, the functional complexity is in its subcellular features and organelle. As a biosensor, Stentor exhibit visually recognizable responses to aversion, including changes in swimming speed (demonstrated in this paper), direction, shape (contracting into a ball), and releasing stentorin, a blue-green pigment that is toxic to some of its adversaries.

    In this paper, our intention was to use movement speed as a general sublethal metric for any plankton, note limited to Stentor. The device was tested on freshwater plankton, but all the surfaces that contact liquids are glass or plastic (Teflon, silicone, and vinyl), hence the system can be used for saltwater plankton. When doing so, the protocol script can be modified to include an additional wash cycle using fresh water to clean the system in between chemical runs to prevent salt deposit build-ups.

  • Conclusions

    The device automatically measured Stentor speed averaged over 1 min which exhibited an inverse relationship with sublethal concentrations. Inexpensive peristaltic pumps under computer control accurately dispensed liquids. The device was also used to automatically remove contamination from the Stentor stock solution.

  • Limitations

    The test was conducted once on a variable number of Stentor per concentration. Several trials, preferably with one Stentor per concentration, are required to measure precision. Only one chemical and one species of plankton were tested. Calculating group speed does not capture the variety of Stentor movement that may be affected by chemical exposure, nor does the camera capture motion in the Z-axis.

  • Conjectures

    Several modifications are suggested to improve and extend the performance and value of the device. Controlling the number of plankton that are injected into the chamber would distinguish chemical response in isolation and in groups. Incorporating plankton purification with plankton injection could speed up the assay. Extracting more movement features (e.g. path length and speed, turn angle, rest time, imaging, and tracking in 3D) would provide a better characterization of sublethal motion effects and insight into the toxic mechanism. Measuring morphology would add another set of potential sublethal toxic effects. Testing chemicals and plankton with published LC50 scores can be used to compare and calibrate system accuracy. Repeating a series several times would provide precision metrics. Automatically loading test chemicals would add another level of automation. Performing an initial low-resolution LC50 scan (several large dilution steps) followed by a high-resolution scan in the sublethal range would conserve plankton stock, cover a wider concentration range, and reduce assay time. Creating a 3D printable chamber design would enable more rapid production, modification, and distribution of the system.

  • Methods

    Device

    The device is based on an open-top mixing and viewing chamber machined out of a 57×81×19 mm tall block of Teflon (Fig. 1B). The chamber was created by drilling a 15.5 mm diameter by 15 mm deep hole into the center of the block (Fig. 1A). The hole’s conical bottom terminates in a 4 mm diameter by 4 mm deep cylinder where four 5 mm diameter orthogonal ports meet (p1 to p4, Fig. 1A, inset). Three of the ports (p1, p2, p3) are connected to computer-controlled peristaltic pumps (Pump 1–3) that connect to Stentor, water and waste containers, respectively. A fifth port (p5) is located 11.6 mm above the lower port (p4) to support a filtering process. Ports p4 and p5 are connected to a pump (Pump 4) with a tube that contains a fine nylon mesh (Rotifer Floss, Reed Mariculture, Campbell, CA) to trap Stentor as the pump circulated water from the bottom to the top of the chamber.

    The water container and its associated pump (Pump 2) were used to clean the chamber after a full sequence of dilutions were completed. The filter was backwashed by running Pump 4 in reverse. During Stentor purification, the water container was replaced with an empty container to collect Stentor. The chamber and filter tubing created a total working volume of 5 mL. A fifth pump (Pump 5) was used to inject air to mix the Stentor container. A camera mounted above the chamber captured video of Stentor (Fig. 1C) backlit by an array of white LEDs diffused by the white Teflon block.

    The device was controlled by a Raspberry Pi 3 single-board computer running the Raspbian operating system (www.raspberrypi.org), providing a user interface to program and execute Python protocol scripts and video recording. An Arduino single-board microcontroller (www.arduino.cc), controlled by the Raspberry Pi over a USB connection, generated pulses to control the peristaltic pumps to accurately dispense liquids. A laptop (Intel i7-6600U @ 2.6 GHz, 16 G RAM) was used to post-process chamber videos, calculating the average Stentor group speed for each dilution. The image processing code was written in Python (Version 3.0) calling OpenCV library functions.

    Pump Operation

    Computer-controlled syringe pumps are often used to accurately dispense liquids in biological and chemical research. However, they are relatively expensive (>$200), large, often uni-directions (dispense only) and have a fixed volumetric capacity. Inexpensive (~$12) compact peristaltic pumps are available (Gikfun Model AE1207) but are designed for continuous low-precision operation. By applying a sequence of pulses to the peristaltic pump’s DC motors, we were able to dispense small increments of liquid under computer control. By switching the voltage polarity with an H-bridge circuit (Texas Instruments L293D), we could selectively dispense and withdraw liquids or air. We found experimentally that apply 10 pulses of ~110 ms duration would reliably administer or remove 100 uL +/- 5%. To achieve this accuracy, the pulse duration for each individual pump was calibrated for each direction.

    Stentor Purification

    We established a design target of injecting 1 mL of spring water containing at least 3 Stentor into the chamber for each chemical concentration. The Stentor stock (Carolina Biological Supply, Burlington, NC) was contaminated with detritus and algae in suspension. We were concerned the Stentor might interact with the contamination (feeding, anchoring, colliding, avoiding), introducing an uncontrolled variable in the toxicology assay. To eliminate this variable, we used the device to remove the contamination. The water container was replaced with an empty collection container and the following protocol script implemented;

    1. Mix Stentor by pumping bubbles into the container;

    2. Pump 0.2 mL of Stentor into the chamber;

    3. Record 20 frames of video and analyze for movement;

    4. If the movement detected, pump the chamber content into the collection container, else pump the chamber content into the waste container;

    5. Repeat steps #1 to #4 until the Stentor container is empty.

    The purification process reduced the initial stock volume by 45%, producing 47.5 mL of purified Stentor stock in under 2 h. Based on visual analysis of the 15 concentration videos used to make figure 1E, 3 of the 148 objects injected into the chamber were identified as algae (2%) and the balance were the Stentor. The level of contamination of the original stock, however, was not quantified.

    Chemical Dilutions

    Household vinegar (Heinz distilled white vinegar from grain 5% acetic acid, 50k ppm) was selected as a benign chemical for prototype testing. The device was initialized to create the first dilution of 25k ppm. The following dilution protocol script was then invoked;

    1. Mix Stentor by pumping bubbles into the container;

    2. Pump 1 mL of Stentor into the chamber;

    3. Mix Stentor and chemical by pumping bubbles into the chamber;

    4. Record 60 seconds of video;

    5. Remove Stentor from the chamber by running the filter pump;

    6. Remove 1 mL of chemical solution from the chamber;

    7. Repeat steps #1 to #7, 29 times then stop;

    Each dilution iteration mixed 4 parts stock with 1 part dilatant (S:D=4:1), to produce a concentration range of 25k ppm to 46 ppm. Each group of Stentor (step #2) were exposed to only one concentration then expelled (step #5). To conserve Stentor stock, spring water was used for the first 15 dilution iterations, as these high concentrations were previously determined to be lethal to Stentor. The assay (initialization, chemical dilution sequence, and image processing) required less than 1 h to complete.

    Video Analysis

    A dilution series produced a collection of 30 videos (1920×[email protected], 60 s duration), one for each dilution. The average speed of all the moving Stentor in each video was determined by accumulating the movement of each Stentor (in pixels) and converting to microns by the optical gain (1 pix = 14.16 um). The Stentor length ranged from 0.64–1.1 mm, corresponding to 45 to 78 image pixels.

    The video analysis algorithm comprised the following steps;

    1. Convert color video to grayscale;

    2. Create a median frame by calculating the median intensity of 25 randomly selected frames;

    3. For each frame detect objects using the following method;

    a. Create a difference frame by taking the absolute difference of the current frame and the median frame;

    b. Create a binary quantized frame by applying a fixed threshold to the difference frame;

    c. Find objects in the binary quantized frame using the OpenCV function cv2.findContours;

    d. Save all objects within an acceptable area range.

    4. Track all saved objects using the following method;

    a. For the first frame, assign every object a unique ID;

    b. For each subsequent frame, match object in the current frame with the closest object in the previous frame, carry forward the ID, and save the object ID and location.

    5. For all objects tracked, calculate the average speed using the following method;

    a. Remove objects that do not substantially move during the duration of the video;

    b. Accumulate the distance of all moving objects in the video;

    c. Average the accumulated distances by the number of objects and the duration of the video.

  • Funding statement

    This work is funded by the National Science Foundation (NSF) grant No. DBI-1548297.

    Disclaimer: Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.

  • Ethics statement

    Not Applicable.

  • References
  • 1
    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum
    2
    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum
    3
    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum
    4
    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum
    5
    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum Lorem ipsum Lorem ipsum Lorem ipsum Lorem ipsum ipsum

    Lorem ipsum Lorem ipsum Lorem ipsum
    Matters14/20

    High Throughput Device for Screening Chemical Sensitivit­­y of Plankton Samples

    Affiliation listing not available.
    Abstractlink

    Plankton provides an essential foundation for life on Earth, supplying most of the breathable oxygen, carbon sequestration, and larvae nutrition. Toxic chemicals introduced into the environment pose a potential danger to plankton and the ecosystem. Traditional plankton chemical toxicity assays measure the concentration required to cause death. Sublethal doses that affect plankton activities like foraging and escaping predators can have a cascading impact on ecosystems. We have developed a high-throughput device to measure the sublethal effect of chemicals on the plankton rate of movement. The device automatically creates a series of chemical dilutions, subjects each concentration to a group of plankton, and calculates the average group movement speed. We tested the device on groups of Stentor coeruleus (n=3 to 26, mean=10) with acetic acid dilutions (43 to 25,000 ppm) and measured a declining trend in average speed with increasing sublethal concentrations (170 to 502 ppm).

    Figurelink

    Figure 1. High throughput device for screening chemical sensitivit­­y of plankton samples.

    (A) Cross-section of a mixing and observation chamber and assorted plumbing. The chamber is initially filled with a 4 mL diluted test chemical. The automatic chemical dilution protocol script is as follows: Pump 5 injects bubbles to mix the Stentor container. Pump 1 injects 1 mL of Stentor into the chamber. Pump 3 (not shown) bubbles the chamber to mix the chemical and Stentor. The camera records 1 min of video. Pump 4 passes the chamber content through a filter, removing the Stentor. Pump 3 (not shown) removes 1 mL of diluted chemical and the process repeats. The entire process is performed in less than 1 h.

    (B) Image of the device (Stentor and waste containers are not shown). The system automatically performs a series of chemical dilutions and captures 1 min videos of Stentor in each dilution. The finished set of videos are transferred to a laptop where an image processing program calculates the average speed of the group members for each dilution.

    (C) A cropped frame from a chemical dilution video of 9 Stentor members (the little black dots) swimming in a chamber of 5 mL of acetic acid (71 ppm). The Stentor is backlit by a panel of LEDs below the chamber. The bright circle in the center is the drain, the deepest section of the block, closest to the LED panel.

    (D) Trail of 9 plankton swimming in acetic acid (71 ppm) integrated over 1 min, calculated by the image processing algorithm. Note the different Stentor swimming patterns including long and straight, corkscrew, and rapid turns.

    (E) The average group speed of Stentor (n=3 to 26, mean=10) in 12 concentrations of acetic acid (71 to 502 ppm) exposed for 1 min. Each point is one group at one chemical concentration. After the group has been exposed and recorded, it is discarded and a new group is introduced at a lower concentration. Concentrations of 624 ppm and higher are fatal. Concentrations of 211 ppm and lower were sublethal. LC50 occurred at 502 ppm.

    Introductionlink

    Plankton are omnipresent in freshwater, seas, and oceans. They form the backbone of aquatic ecosystems and food chains, providing most of the planets' breathable oxygen, carbon sequestration, and larval nutrition. Plankton are any organism that lives in water and are unable to swim against the current. This broad definition includes bacteria, phytoplankton, and zooplankton. Many plankton have cilia or flagella to propel themselves to forage for food and avoid predators. Due to the interconnectedness of organisms in an ecosystem, disturbing one plankton species can cause a major disruption in an ecosystem, for example, harmful algae blooms[1].

    Anthropogenic activities, including agricultural runoff, consumer, and industrial waste, introduce toxic chemicals into waterways[2]. Only a small portion of a large number of chemicals in use are tested for plankton toxicity (https://cen.acs.org/articles/95/i9/chemicals-use-today.html). The standard acute toxicity reference is LC50, the chemical concentration required to kill 50% of a group of test organism after a short exposure, typically 24 to 96 h[3][4]. Chemicals that interfere with a biological process tend to have a LC50 response that is strongly species-specific. For example, cadmium has a LC50 toxicity from 2 ug/L to 4.6 g/L depending on the aquatic species[5].

    Chronic toxicity tests observe sublethal impacts over a long duration (>10% of the organism’s life span) to allow the toxic effect to accumulate[6]. Chronic exposure may result in alterations in reproduction, sex ratio, morphology, feeding and predator response, effects that can impact the community and ecosystem[7]. Many of these symptoms appear at concentrations lower than LC50[8]. For example, for Daphnia, the carbaryl insecticide has a LC50 of 12.76 ppb for a 24 h exposure, while a 50 min exposure at 5 ppb reduces the hop rate by 25% and a 21-day exposure at 0.4 ppb reduces molts by 11%[9][10].

    Recently microfluidic devices have been developed to perform high-throughput plankton research[11]. By fabricating gradient generators, multiple chambers with different chemical concentrations can be constructed to test for acute toxicity[12]. Multiple zones have been fabricated to accommodate plankton of different sizes[13]. To accommodate the 4 most common plankton cited in toxicity studies, the chambers must accommodate a size range from 0.2 mm to 30 mm (Daphnia, a cladoceran, and Hydra, a cnidarian, respectively)[8], a dimensional span outside of the range of typical microfluid devices[14].

    Objectivelink

    The objective is to design and build a high-throughput device to measure chemical toxicity on a variety of aquatic organisms over a wide range of concentrations by detecting sublethal movement speed during a short-exposure time (one-minute). To demonstrate the operation, the device tests the effect of acetic acid (from household vinegar) on the movement speed of groups of Stentor, one of the largest single-cell plankton.

    Results & Discussionlink

    The average speed of Stentor (n=3 to 26, mean=10), exposed to a concentration of acetic acid for 1 min, is presented in figure 1E. A group is exposed to one concentration then discarded, then a new group is subjected to a lower concentration. Acetic acid concentrations of 624 ppm and higher are fatal (all group members died). Concentrations of 211 ppm and lower were sublethal (all group members survived). At concentrations higher than 170 ppm speed decreases until a fatal concentration is reached. LC50 occurred at 502 ppm. There is considerable speed variation in sublethal concentrations below 170 ppm, with several confounding factors to consider. Stentor exhibit several modes of motion (Fig. 1D) including traveling long and fast in a straight line, slower travel with a pronounced corkscrew trail, precessing about an anchor point and moving in the Z direction (towards the camera), an axis not captured. The number of Stentor in a group can impact the average speed; small groups are more sensitive to intra-member variation while Stentor in large groups may slow down and change direction to avoid collisions.

    For each chemical concentration, 1 mL of Stentor was injected into the chamber, providing Stentor and the diluent (spring water). If an injection did not contain any Stentor, there would be no reading for that concentration, as the dilution would have been performed, and injecting more Stentor would create a new dilution. We relied on the pre-filtering to achieve a minimum concentration of 3 Stentor per mL and bubbling to mix the Stentor before each injection. The minimum target was reached, however, there was a significant range of membership (3 to 26) per injection.

    The device is intended to work with any plankton between 0.2 mm and 3 mm, constrained by the chamber diameter and tubing inner diameter (3 mm). We use the device for three purposes; species selection (i.e. Stentor purification), chemical toxicity assay (i.e. reporting average motion speed at sublethal concentrations), and investigating cell functions (ongoing research not reported here). We selected Stentor to test this device based on our interest in evaluating Stentor as a biosensor[15] and as a model organism to investigate how cells function and interact with their environment. It is large plankton (>0.5 mm), simplifying observation and handling, and a single cell, so there is no multi-cellular interaction to consider. Instead, the functional complexity is in its subcellular features and organelle[16]. As a biosensor, Stentor exhibit visually recognizable responses to aversion, including changes in swimming speed (demonstrated in this paper), direction, shape (contracting into a ball), and releasing stentorin, a blue-green pigment that is toxic to some of its adversaries[17].

    In this paper, our intention was to use movement speed as a general sublethal metric for any plankton, note limited to Stentor. The device was tested on freshwater plankton, but all the surfaces that contact liquids are glass or plastic (Teflon, silicone, and vinyl), hence the system can be used for saltwater plankton. When doing so, the protocol script can be modified to include an additional wash cycle using fresh water to clean the system in between chemical runs to prevent salt deposit build-ups.

    Conclusionslink

    The device automatically measured Stentor speed averaged over 1 min which exhibited an inverse relationship with sublethal concentrations. Inexpensive peristaltic pumps under computer control accurately dispensed liquids. The device was also used to automatically remove contamination from the Stentor stock solution.

    Limitationslink

    The test was conducted once on a variable number of Stentor per concentration. Several trials, preferably with one Stentor per concentration, are required to measure precision. Only one chemical and one species of plankton were tested. Calculating group speed does not capture the variety of Stentor movement that may be affected by chemical exposure, nor does the camera capture motion in the Z-axis.

    Conjectureslink

    Several modifications are suggested to improve and extend the performance and value of the device. Controlling the number of plankton that are injected into the chamber would distinguish chemical response in isolation and in groups. Incorporating plankton purification with plankton injection could speed up the assay. Extracting more movement features (e.g. path length and speed, turn angle, rest time, imaging, and tracking in 3D) would provide a better characterization of sublethal motion effects and insight into the toxic mechanism. Measuring morphology would add another set of potential sublethal toxic effects. Testing chemicals and plankton with published LC50 scores can be used to compare and calibrate system accuracy. Repeating a series several times would provide precision metrics. Automatically loading test chemicals would add another level of automation. Performing an initial low-resolution LC50 scan (several large dilution steps) followed by a high-resolution scan in the sublethal range would conserve plankton stock, cover a wider concentration range, and reduce assay time. Creating a 3D printable chamber design would enable more rapid production, modification, and distribution of the system.

    Methodslink

    Device

    The device is based on an open-top mixing and viewing chamber machined out of a 57×81×19 mm tall block of Teflon (Fig. 1B). The chamber was created by drilling a 15.5 mm diameter by 15 mm deep hole into the center of the block (Fig. 1A). The hole’s conical bottom terminates in a 4 mm diameter by 4 mm deep cylinder where four 5 mm diameter orthogonal ports meet (p1 to p4, Fig. 1A, inset). Three of the ports (p1, p2, p3) are connected to computer-controlled peristaltic pumps (Pump 1–3) that connect to Stentor, water and waste containers, respectively. A fifth port (p5) is located 11.6 mm above the lower port (p4) to support a filtering process. Ports p4 and p5 are connected to a pump (Pump 4) with a tube that contains a fine nylon mesh (Rotifer Floss, Reed Mariculture, Campbell, CA) to trap Stentor as the pump circulated water from the bottom to the top of the chamber.

    The water container and its associated pump (Pump 2) were used to clean the chamber after a full sequence of dilutions were completed. The filter was backwashed by running Pump 4 in reverse. During Stentor purification, the water container was replaced with an empty container to collect Stentor. The chamber and filter tubing created a total working volume of 5 mL. A fifth pump (Pump 5) was used to inject air to mix the Stentor container. A camera mounted above the chamber captured video of Stentor (Fig. 1C) backlit by an array of white LEDs diffused by the white Teflon block.

    The device was controlled by a Raspberry Pi 3 single-board computer running the Raspbian operating system (www.raspberrypi.org), providing a user interface to program and execute Python protocol scripts and video recording. An Arduino single-board microcontroller (www.arduino.cc), controlled by the Raspberry Pi over a USB connection, generated pulses to control the peristaltic pumps to accurately dispense liquids. A laptop (Intel i7-6600U @ 2.6 GHz, 16 G RAM) was used to post-process chamber videos, calculating the average Stentor group speed for each dilution. The image processing code was written in Python (Version 3.0) calling OpenCV library functions.

    Pump Operation

    Computer-controlled syringe pumps are often used to accurately dispense liquids in biological and chemical research. However, they are relatively expensive (>$200), large, often uni-directions (dispense only) and have a fixed volumetric capacity. Inexpensive (~$12) compact peristaltic pumps are available (Gikfun Model AE1207) but are designed for continuous low-precision operation. By applying a sequence of pulses to the peristaltic pump’s DC motors, we were able to dispense small increments of liquid under computer control. By switching the voltage polarity with an H-bridge circuit (Texas Instruments L293D), we could selectively dispense and withdraw liquids or air. We found experimentally that apply 10 pulses of ~110 ms duration would reliably administer or remove 100 uL +/- 5%. To achieve this accuracy, the pulse duration for each individual pump was calibrated for each direction.

    Stentor Purification

    We established a design target of injecting 1 mL of spring water containing at least 3 Stentor into the chamber for each chemical concentration. The Stentor stock (Carolina Biological Supply, Burlington, NC) was contaminated with detritus and algae in suspension. We were concerned the Stentor might interact with the contamination (feeding, anchoring, colliding, avoiding), introducing an uncontrolled variable in the toxicology assay. To eliminate this variable, we used the device to remove the contamination. The water container was replaced with an empty collection container and the following protocol script implemented;

    1. Mix Stentor by pumping bubbles into the container;

    2. Pump 0.2 mL of Stentor into the chamber;

    3. Record 20 frames of video and analyze for movement;

    4. If the movement detected, pump the chamber content into the collection container, else pump the chamber content into the waste container;

    5. Repeat steps #1 to #4 until the Stentor container is empty.

    The purification process reduced the initial stock volume by 45%, producing 47.5 mL of purified Stentor stock in under 2 h. Based on visual analysis of the 15 concentration videos used to make figure 1E, 3 of the 148 objects injected into the chamber were identified as algae (2%) and the balance were the Stentor. The level of contamination of the original stock, however, was not quantified.

    Chemical Dilutions

    Household vinegar (Heinz distilled white vinegar from grain 5% acetic acid, 50k ppm) was selected as a benign chemical for prototype testing. The device was initialized to create the first dilution of 25k ppm. The following dilution protocol script was then invoked;

    1. Mix Stentor by pumping bubbles into the container;

    2. Pump 1 mL of Stentor into the chamber;

    3. Mix Stentor and chemical by pumping bubbles into the chamber;

    4. Record 60 seconds of video;

    5. Remove Stentor from the chamber by running the filter pump;

    6. Remove 1 mL of chemical solution from the chamber;

    7. Repeat steps #1 to #7, 29 times then stop;

    Each dilution iteration mixed 4 parts stock with 1 part dilatant (S:D=4:1), to produce a concentration range of 25k ppm to 46 ppm. Each group of Stentor (step #2) were exposed to only one concentration then expelled (step #5). To conserve Stentor stock, spring water was used for the first 15 dilution iterations, as these high concentrations were previously determined to be lethal to Stentor. The assay (initialization, chemical dilution sequence, and image processing) required less than 1 h to complete.

    Video Analysis

    A dilution series produced a collection of 30 videos (1920×[email protected], 60 s duration), one for each dilution. The average speed of all the moving Stentor in each video was determined by accumulating the movement of each Stentor (in pixels) and converting to microns by the optical gain (1 pix = 14.16 um). The Stentor length ranged from 0.64–1.1 mm, corresponding to 45 to 78 image pixels.

    The video analysis algorithm comprised the following steps;

    1. Convert color video to grayscale;

    2. Create a median frame by calculating the median intensity of 25 randomly selected frames;

    3. For each frame detect objects using the following method;

    a. Create a difference frame by taking the absolute difference of the current frame and the median frame;

    b. Create a binary quantized frame by applying a fixed threshold to the difference frame;

    c. Find objects in the binary quantized frame using the OpenCV function cv2.findContours;

    d. Save all objects within an acceptable area range.

    4. Track all saved objects using the following method;

    a. For the first frame, assign every object a unique ID;

    b. For each subsequent frame, match object in the current frame with the closest object in the previous frame, carry forward the ID, and save the object ID and location.

    5. For all objects tracked, calculate the average speed using the following method;

    a. Remove objects that do not substantially move during the duration of the video;

    b. Accumulate the distance of all moving objects in the video;

    c. Average the accumulated distances by the number of objects and the duration of the video.

    Funding Statementlink

    This work is funded by the National Science Foundation (NSF) grant No. DBI-1548297.

    Disclaimer: Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.

    Conflict of interestlink

    The authors declare no conflicts of interest.

    Ethics Statementlink

    Not Applicable.

    No fraudulence is committed in performing these experiments or during processing of the data. We understand that in the case of fraudulence, the study can be retracted by ScienceMatters.

    Referenceslink
    1. Rachel S. Lasley-Rasher, Kathryn Nagel, Aakanksha Angra, Jeannette Yen
      Intoxicated copepods: ingesting toxic phytoplankton leads to risky behaviour
      Proceedings of the Royal Society B: Biological Sciences, 283/2016, page 20160176 DOI: 10.1098/rspb.2016.0176chrome_reader_mode
    2. Priya Mallika D’costa, Maria Shamina D’silva, Ravidas Krishna Naik
      Impact of Pollution on Phytoplankton and Implications for Marine Econiches
      Marine Pollution and Microbial Remediation, 2016, pages 205-222 DOI: 10.1007/978-981-10-1044-6_13chrome_reader_mode
    3. S. K. Baliarsingh, S. Srichandan, T. Padmavati, S. Naik, K. C. Sahu
      Accumulation of heavy metals by freshwater zooplankton - a toxicological study
      Recent Advances in Space Technology Services and Climate Change 2010 (RSTS & CC-2010), 2010, pages 7-14 DOI: 10.1109/rstscc.2010.5712789chrome_reader_mode
    4. Zhibing Jiang, Yijun Huang, Xiaoqun Xu,more_horiz, Jiangning Zeng
      Advance in the toxic effects of petroleum water accommodated fraction on marine plankton
      Acta Ecologica Sinica, 30/2010, pages 8-15 DOI: 10.1016/j.chnaes.2009.12.002chrome_reader_mode
    5. J. A. Hoekstra, M. A. Vaal, J. Notenboom, W. Slooff
      Variation in the sensitivity of aquatic species to toxicants
      Bulletin of Environmental Contamination and Toxicology, 53/1994, pages 98-105 DOI: 10.1007/bf00205145chrome_reader_mode
    6. Michael C. Newman
      Fundamentals of Ecotoxicology
      CRC Press, 2010, page chrome_reader_mode
    7. T Hanazato
      Pesticide effects on freshwater zooplankton: an ecological perspective
      Environmental Pollution, 112/2001, pages 1-10 DOI: 10.1016/s0269-7491(00)00110-xchrome_reader_mode
    8. Roberto Rico-Martínez, Mario Alberto Arzate-Cárdenas, Daniel Robles-Vargas, Ignacio Alejandro Pérez-Legaspi, Alvarado-Flores Jesús, Gustavo Emilio Santos-Medrano
      Rotifers as Models in Toxicity Screening of Chemicals and Environmental Samples
      Invertebrates - Experimental Models in Toxicity Screening, 2016 DOI: 10.5772/61771chrome_reader_mode
    9. Stanley I. Dodson, Takayuki Hanazato, Patrick R. Gorski
      Behavioral responses of Daphnia pulex exposed to carbaryl and Chaoborus kairomone
      Environmental Toxicology and Chemistry, 14/1995, pages 43-50 DOI: 10.1002/etc.5620140106chrome_reader_mode
    10. Hela Toumi, Karen F. Burga-Perez, Jean-François Ferard
      Acute and chronic ecotoxicity of carbaryl with a battery of aquatic bioassays
      Journal of Environmental Science and Health, Part B, 51/2016, pages 57-62 DOI: 10.1080/03601234.2015.1080500chrome_reader_mode
    11. Mathias Girault, Thomas Beneyton, Yolanda Del Amo, Jean-Christophe Baret
      Microfluidic technology for plankton research
      Current Opinion in Biotechnology, 55/2019, pages 134-150 DOI: 10.1016/j.copbio.2018.09.010chrome_reader_mode
    12. Guo-Xia Zheng, Ya-Jie Li, Lin-Lin Qi,more_horiz, Yun-Hua Wang
      Marine phytoplankton motility sensor integrated into a microfluidic chip for high-throughput pollutant toxicity assessment
      Marine Pollution Bulletin, 84/2014, pages 147-154 DOI: 10.1016/j.marpolbul.2014.05.019chrome_reader_mode
    13. Myint Myint Maw, Xinxiang Pan, Zhen Peng, Yanjuan Wang, Long Zhao, Bowen Dai, Junsheng Wang
      A Changeable Lab-on-a-Chip Detector for Marine Nonindigenous Microorganisms in Ship’s Ballast Water
      Micromachines, 9/2018, page 20 DOI: 10.3390/mi9010020chrome_reader_mode
    14. Jessica Melin, Stephen R. Quake
      Microfluidic Large-Scale Integration: The Evolution of Design Rules for Biological Automation
      Annual Review of Biophysics and Biomolecular Structure, 36/2007, pages 213-231 DOI: 10.1146/annurev.biophys.36.040306.132646chrome_reader_mode
    15. Trishala K. Parmar, Deepak Rawtani, Y. K. Agrawal
      Bioindicators: the natural indicator of environmental pollution
      Frontiers in Life Science, 9/2016, pages 110-118 DOI: 10.1080/21553769.2016.1162753chrome_reader_mode
    16. James J. Russell, Julie A. Theriot, Pranidhi Sood, Wallace F. Marshall, Laura F. Landweber, Lillian Fritz-Laylin, Jessica K. Polka, Snezhana Oliferenko, Therese Gerbich, Amy Gladfelter, James Umen, Magdalena Bezanilla, Madeline A. Lancaster, Shuonan He, Matthew C. Gibson, Bob Goldstein, Elly M. Tanaka, Chi-Kuo Hu, Anne Brunet
      Non-model model organisms
    17. Akio Miyake Terue Harumoto Hideo Iio
      Defence function of pigment granules in Stentor coeruleus
      European Journal of Protistology, 37/2001, pages 77-88 DOI: 10.1078/0932-4739-00809chrome_reader_mode
    Commentslink

    Create a Matters account to leave a comment.