Thermal Limits Determination for Zooplankton Using a Heat Block

Summary

The present protocol illustrates the use of commercially available components to generate a stable and linear thermal gradient. Such gradient can then be used to determine the upper thermal limit of planktonic organisms, particularly invertebrate larvae.

Abstract

Thermal limits and breadth have been widely used to predict species distribution. As the global temperature continues to rise, understanding how thermal limit changes with acclimation and how it varies between life stages and populations are vital for determining the vulnerability of species to future warming. Most marine organisms have complex life cycles that include early planktonic stages. While quantifying the thermal limit of these small early developmental stages (tens to hundreds of microns) helps identify developmental bottlenecks, this process can be challenging due to the small size of target organisms, large bench space requirement, and high initial fabrication cost. Here, a setup that is geared toward small volumes (mL to tens of mL) is presented. This setup combines commercially available components to generate a stable and linear thermal gradient. Production specifications of the setup, as well as procedures to introduce and enumerate live versus dead individuals and compute lethal temperature, are also presented.

Introduction

Thermal tolerance is key to organisms' survival and function^1,2. As the planet continues to warm due to anthropogenic carbon emissions, increasing attention is being paid to the determination and application of thermal limits³. Various endpoints, such as mortality, failure to develop, and loss of mobility, have been used to determine both upper and lower thermal limits⁴. These thermal limits are often considered a proxy for an organism's thermal niche. This information is in turn used to identify species that are more vulnerable to global warming, as well as predict future species distribution and the resulting species interactions^3,5,6,7. However, determining thermal limits, especially for small planktonic organisms, can be challenging.

For planktonic organisms, particularly the larval stages of marine invertebrates, the thermal limit can be determined through chronic exposure. Chronic exposure is achieved by rearing larvae at several temperatures over days to weeks and determining the temperature at which larval survivorship and/or developmental rate reduces^8,9,10. However, this approach is rather time-consuming and requires large incubators and experience in larval husbandry (see reference¹¹ for a good introduction to culturing marine invertebrate larvae).

Alternatively, acute exposure to thermal stress can be used to determine thermal limits. Often, this determination approach involves placing small vials with larvae in temperature-controlled dry baths^12,13,14, leveraging thermal gradient functions in PCR thermal cyclers^15,16, or putting glass vials/microcentrifuge tubes along a thermal gradient generated by applied heating and cooling on the ends of large aluminum blocks with holes in which the vials fit snuggly^17,18,19. Typical dry baths generate a single temperature; hence, multiple units must be operated simultaneously to assess performance across a range of temperatures. Thermal cyclers generate a gradient but only accommodate a small sample volume (120 µL) and require careful manipulations. Similar to thermal cyclers, large aluminum blocks create linear and stable temperature gradients. Both approaches can be coupled with logistic or probit regression to compute the lethal temperature for 50% percent of the population (LT₅₀)^12,20,21. However, the aluminum blocks used were ~100 cm long; this size demands a large lab space and access to specialized computer numerical control milling machines to drill the holes. Together with using two research-grade water baths to maintain the target temperature, the financial cost of assembling the setup is high.

Therefore, this work aims to develop an alternative means to generate a stable, linear temperature gradient with commercially available parts. Such a product must have a small footprint and should be able to be easily used for acute thermal stress exposure experiments for planktonic organisms. This protocol is developed with zooplankton that is <1 mm in size as target organisms, and thus, it was optimized for the use of a 1.5 or 2 mL microcentrifuge tube. Larger study organisms will require containers larger than the 1.5 mL microcentrifuge tubes used and enlarged holes in the aluminum blocks.

In addition to making the experimental apparatus more accessible, this work aims to simplify the data processing pipeline. While commercial statistical software provides routines to compute LT₅₀ using logistic or probit regression, the licensing cost is non-trivial. Therefore, an easy-to-use script that relies on the open-source statistical program R²² would make data analysis more accessible.

This protocol shows how a compact heat block can be fabricated with commercially available parts and be applied to exposing zooplankton (larvae of the sand dollar Dendraster excentricus) to acute heat stress to determine their upper thermal limit.

Protocol

1. Fabrication of the heat block

1. Wire the 120 V, 100 W strip heater to the rheostat (see Table of Materials).
2. Prepare the 20.3 cm x 15.2 cm x 5 cm (8 in x 5 in x 2 in) block of aluminum by drilling 60 holes in a 6 x 10 grid (see Table of Materials). Ensure that the holes are spaced 2 cm from center to center in both directions. Each should be 1.1 cm in diameter and 4.2 cm deep (Figure 1).
   NOTE: Perform the drilling on a milling machine or drill press with high-speed steel drill bits. The heating element and cooling element were both chosen to cover as much of the contact surface of the 15.2 cm x 5 cm surfaces as possible.
3. Drill two additional holes on one of the 20.3 cm x 5 cm surfaces between the 1^st and 2^nd column and the 9^th and 10^th column, matching the size of the temperature controller probes (see Table of Materials).
4. Construct a case from 1.2 cm (0.5 in) clear acrylic sheets (see Table of Materials) to both hold the elements in place and insulate the completed heat block. Use two layers of acrylic to insulate the back side of the heating element (Figure 1).
5. In the final assembly, apply thermal paste (see Table of Materials) to maximize heat conductance from the heating element into the block and from the block to the cooling element.

2. Determination of thermal gradient settings

1. Connect the water bath/aquarium chiller with Tygon tubing (see Table of Materials). Insulate the tubing with foam pipe insulation as needed.
2. Insert the thermostat probe into the holes on the side of the aluminum block. Ensure that probe 1 is positioned near the heating element.
3. Place microcentrifuge tubes filled to the brim (1.5 mL) with tap water in all the milled holes (60 tubes total).
4. Turn on the temperature controller and set the stop heating temperature of probe 1 to 35-37 °C and probe 2 to 21.5-22.5 °C.
   NOTE: The proposed thermostat has two outlets that operate independently; only probe 1 is used for regulating warm temperature in this particular use case. Therefore, set the temperature of probe 2 to that of the low-end temperature.
5. Rotate the rheostat to turn on the heating element and set it to medium.
6. Turn on the water bath/aquarium chiller and set the chiller temperature to 15 °C.
7. Check that the block is warm on one end and cool on the other after 10 min.
   CAUTION: The exposed ends of the heating element can be hot; do not touch them.
8. Check the temperature inside each microcentrifuge tube using a thermocouple with a K-type electrode (see Table of Materials) every 10 min afterward. The temperature will stabilize after ~60 min and appear linear (Figure 2).
9. Adjust the values of the endpoints by changing the settings of the temperature controller and the water bath as needed.

3. Thermal exposure and live:dead enumeration

NOTE: Step 2 can be omitted once the desired settings for the temperature gradient are determined.

1. Turn on the recirculating water bath and heater and set them to 15 °C and 37 °C, respectively, to generate a temperature gradient from 19.5 °C to 37 °C.
2. To ensure the thermal gradient is linear, place microcentrifuge tubes filled to the brim (1.5 mL) with tap water in all the milled holes (60 tubes total).
3. Let the heat block reach the set temperature by waiting 45-60 min. Check the temperature inside each microcentrifuge tube using a thermocouple with a K-type electrode to see if it has reached the expected temperature. Note these temperatures.
4. If the study organisms are >500 µm in size and can be easily transferred from one container to another (e.g., a copepod), fill a 1.5 mL microcentrifuge tube with 750 µL of 0.45 µm filtered seawater. Alternatively, if the study organisms are small, fill a 1.5 mL microcentrifuge tube with 250 µL of 0.45 µm filtered seawater.
   NOTE: For the representative data, larvae of the sand dollar Dendraster excentrics, which are 2 , 4, and 6 days post-fertilization, were used (see Table of Materials). The average (± S.D., n = 15 for each age) size of these individuals was 152 ± 7 µm, 260 ± 17 µm, and 292 ± 14 µm, respectively. Given these larvae can be easily concentrated (step 3.5), the microcentrifuge tubes were filled with 750 µL of filtered seawater.
5. Concentrate the study organisms' culture with reverse filtering (i.e., place the mesh in the container holding the study organisms and remove water through the top of the mesh), so that the study organisms remain in the bottom of the beaker¹¹.
   NOTE: A 30 µm nylon mesh was used for the larval sand dollars studied (see Table of Materials).
6. Rinse the concentrated animal sample with filtered seawater (e.g., when culturing with algal food items or other chemicals). Repeat the reverse filtering once more to concentrate the animal sample.
7. Place a known number of individual organisms into the half-filled microcentrifuge tubes. Count the small planktonic organisms under a dissecting microscope (see Table of Materials) and transfer them with glass Pasteur pipettes.
   NOTE: The number of organisms to place is size dependent; for larval sand dollars that were ~200 µm in size, 20 individuals per microcentrifuge tube was appropriate.
   CAUTION: Glass pipettes are more desirable than plastic pipettes as some planktonic organisms are hydrophobic and will stick to plastic surfaces.
8. Add 0.45 µm filtered seawater to the microcentrifuge tubes containing animals until the final volume is 1 mL.
9. To allow the organisms to gradually warm up to the desired experimental temperature, place the microcentrifuge tubes with animals, prepared in step 3.7, into the heat block starting from the cold end. Place pairs of microcentrifuge tubes on each row (12 tubes total).
10. Wait 10 min.
11. Move the pairs of microcentrifuge tubes inserted at step 3.9 to the adjacent drilled holes with warmer temperatures. Place additional pairs of microcentrifuge tubes in each row at the cold end. Each row will now have four tubes. Wait another 10 min.
12. Continue to add microcentrifuge tubes with animals by shifting their positions from the colder end to the warmer end in pairs. Wait 10 min between each shift until the heat block is completely filled.
    NOTE: Steps 3.9-3.12 are considered a ramping-up phase to increase the temperature experienced by the study organisms gradually.
13. Let the animals incubate at the designated temperature for 2 h. This step is the constant temperature exposure phase of the experiment.
    1. Check the temperature of the microcentrifuge tubes with a thermocouple every hour if the incubation period exceeds 2 h.
       NOTE: Adjust the incubation time based on the experimental needs. If the incubation is longer than 2 h, check the temperature of the tubes at regular time intervals with a thermocouple in case of unforeseen equipment failure. To minimize disturbance to the study organisms, randomly place six or more microcentrifuge tubes filled only with filtered seawater into the block for temperature monitoring.
14. At the end of the incubation period, measure the temperature inside each microcentrifuge tube using a thermocouple with a K-type electrode. Note these temperatures.
15. Remove all 60 microcentrifuge tubes with animals and place them in pre-labeled holders.
16. Incubate the tubes (step 3.14) at the predetermined temperature, such as the rearing temperature, for 1 h, which is the recovery period.
    NOTE: The recovery period can be species-specific. For the larval sand dollar, the rearing temperature was 18 °C, and thus the sample was placed in an environmental chamber. Consult relevant literature and/or conduct a trial experiment to ensure the live:dead count was not affected by the length of the recovery period. In the representative data, the number of animals alive after 1 h was the same as after 12 or 24 h of recovery.
17. To enumerate the proportion of study organism that is alive after the thermal exposure, transfer the contents of an individual microcentrifuge tube onto a 35 mm Petri dish using a glass pipette.
18. Observe and note the relative number of individuals that are still active (alive) and those that have seized swimming or dissolved (dead) under a dissecting microscope. Ensure that the total number of individuals observed equals the number of individuals placed into the tubes in step 3.7. Check the side of the microcentrifuge tubes and Petri dish for individuals if the numbers do not match.

4. Computation of LT₅₀

1. Generate a data table in CSV format with at least the following headers: grouping variable of interest, temperature of the tube in °C, number of individuals alive, and number of individuals dead.
   NOTE: For the representative data, the grouping variable of interest is replaced by age since the goal is to compare between age groups.
2. To fit the data with logistic regression, use a generalized linear model with a binomial distribution. Supplementary Coding File 1 shows an example sample script using the open-source software R²².
3. To determine the median upper thermal limit (LT₅₀), compute the predictor value (i.e., temperature) at which 50% of the individuals survived. Supplementary Coding File 2 shows an example script using the function dose.p from the MASS²³ in R²².

Representative Results

The goal of this protocol is to determine the upper thermal limit of zooplankton. To do so, a stable and linear thermal gradient is needed. The proposed setup was able to generate a thermal gradient ranging from 14 °C to 40 °C by setting the water bath temperature to 8 °C and the heater to 39 °C (Figure 2A). The temperature gradient can be narrowed and shifted by changing the endpoint values. A thermal gradient with a narrower range (19 °C to 37 °C) was also generated by setting the heater to 37 °C and the water bath to 15 °C. The temperature in the block stabilizes within 45 min to 1 h of setup (Figure 2B).

To illustrate the application of this protocol to zooplankton, the change in the upper thermal limit, indicated by LT₅₀, through ontogeny in the larvae of the sand dollars (Dendraster excentricus) was examined. Gravid sand dollars were obtained commercially (see Table of Materials). The release of gametes was induced by injecting 0.5-1 mL of 0.35 M potassium chloride. The eggs collected were rinsed through 63 µm nylon mesh with 0.45 µm filtered seawater. The sperm was collected dry and kept on ice. The eggs were fertilized at ~10⁴ sperms per mL. Common garden cultures were created with gametes from three males and three females at five individuals per mL. These larval cultures were kept in filtered seawater with a salinity of 32 psu at 18 °C under a 12:12 light:dark cycle with complete water change every other day.

As larval sand dollars developed, the upper thermal limit increased from 28.6 °C (± 0.02 °C S.E) at 2 days post-fertilization to 28.8 °C (± 0.02 °C S.E) at 4 days post-fertilization and 29.3 °C (± 0.02 °C S.E) at 6 days post-fertilization (Figure 3). These upper thermal limits suggest sand dollars live within their thermal limit during the average summer sea surface temperature of ~20 °C or lower along the Pacific coast. However, with increasing frequency and intensity of marine heatwaves, the maximum temperature continues to rise. A peak temperature of 26.4 °C was recorded in Southern California Bight in August 2018 (Fumo et al.²⁴). Given that this species reproduce in spring and summer, the survivorship of their early life stage is likely to diminish during these extreme events. The predicted survivorship would decrease by 10% when the temperature reaches 26.5 °C.

Pair-wise comparisons using the ratio test developed by Wheeler et al.²⁵ suggest the median lethal temperature was significantly different between the three age groups (p < 0.001). Earlier stages (gastrula and early prisms that were 2 days old) were more sensitive to thermal stress than older larvae. This observation suggests that the thermal limit deduced from a single time point of development is not representative of that species throughout its life history.

Figure 1: Labeled diagram of the heat block. (A) Top view of the setup with all components connected. (B,D) Placement and connections for the heater terminals. (C,E) Placement of the heat exchanger (cooling elemenet) and the associated tubings to the water bath. Please click here to view a larger version of this figure.

Figure 2: Temperature changes in the heat block over 1 h with endpoints set to 15 and 37 °C. (A) A linear gradient was achieved within 1 h. Changing in the endpoint settings varies the temperature range, and the largest range was from 14 °C to 40 °C. (B) The temperature difference between replicate rows was negligible (<0.8 °C); data from two replicate rows were plotted for each setting in (B). Please click here to view a larger version of this figure.

Figure 3: Survivorship of larval sand dollars (Dendraster excentricus) across a temperature range of 19 to 37 °C through ontogeny (2, 4, and 6 days post-fertilization [dpf]). Each datum represents the proportion of larvae that survived a 2 h incubation at the specific temperature followed by a 1 h recovery period. A logistic regression was performed using the generalized linear model with binomial distribution in the statistical software R. Please click here to view a larger version of this figure.

Supplementary Coding File 1: An R script to generate logistic curves for the data set with a step-by-step example. Please click here to download this File.

Supplementary Coding File 2: An R script to generate LT₅₀ estimates. Please click here to download this File.

Discussion

This protocol provides an accessible and customizable approach to determine the thermal limits of small plankton organisms through acute thermal exposure. The 10-hole design and flexible temperature endpoints, controlled by the water bath at the lower end and the heater at the upper end, enable one to determine LT₅₀ with precision. Using this approach, a difference in the thermal limit that is <1 °C could be detected (Figure 3). This approach provides a rapid determination of thermal limits (in hours) for a variety of species, and the resulting values have been applied to multiple species distribution models^2,21. However, it is important to note that acute exposure likely provides a different thermal tolerance estimate when compared to chronic exposure^8,26.

One major advantage of the current design is that 10 temperature treatments and six replicates are included within a small footprint (20.3 cm x 15.2 cm x 5 cm). Previous publications using a similar thermal gradient approach to determine thermal limits used aluminum bars that were bigger (180 cm x 10 cm x 6 cm in²⁷, 91 cm × 25 cm × 15 cm in¹⁰, and 60 cm x 20 cm in¹⁷). While dry baths that hold a single temperature are smaller (e.g., 18.5 cm x 18.5 cm x 2.5 cm) and offer multiple replicates, several units (more than four) are required to generate a performance curve that includes multiple temperatures, or the experiments need to be repeated over time which could introduce confounding factors. The heat block design reduces both the fabrication cost and space requirements. The fabrication can be completed with a drill press, or researchers without immediate access to a milling machine could opt for commercial CNC machining services. The use of commercially available parts further controls the fabrication cost. If one can use an existing heating/cooling water bath or aquarium chillers, the remaining cost of the parts totals to less than $350. Otherwise, aquarium chillers for a 10 gallon (~35 L) fish tank can be purchased for <$150.

The current design may be modified to fit the researcher's needs. If the target organisms are bigger in size, scintillation vials are good alternative containers, and larger holes would be required. That said, the aluminum block is removable in the current design, so multiple blocks can be made and swapped out to fit the experimental needs. If the goal of the experiment is to determine a lower thermal limit or focus on polar organisms, placing chilling water blocks on both ends of the main aluminum block is more appropriate.

Similar to other studies on zooplankton, the current protocol does not include a gradual cooling phase^20,27. Researchers can consider removing the microcentrifuge tubes in pairs and shifting them down the temperature gradient (i.e., reversing steps 3.9-3.12) to achieve gradual cooling if their study organisms are sensitive to a sudden temperature decrease.

The utility of this setup can be diminished by several factors, namely the choice of (1) the endpoint temperature settings, (2) the exposure and recovery duration, and 3) the metric used to determine the binomial state (live vs. dead; developed vs. non-developed). To address these potential limitations, preliminary testing is highly recommended.

Since the logistic regression assumes a binomial distribution, endpoints with 100% survival and mortality are preferred. For marine organisms, a reasonable starting range would be the mean annual sea surface temperature of the collection site plus 10-15 °C. One can then narrow the temperature range investigated after such an initial trial, as the smaller the temperature difference between holes, the more fine-tuned the LT₅₀ estimate.

The duration of exposure and recovery are species-specific. For instance, Kuo et al.²⁷ allowed juvenile whelks (Nucella canaliculata) to recover for 24 h, while Hammond et al.²⁸ allowed larval purple urchins (Stronglylocentrotus purprtaus) 1 h for recovery. One could perform a short experiment to determine if the live:dead count differs between recovery periods. Depending on the definition of the binomial state chosen (e.g., live vs. dead), recovery time may not be necessary. If the goal of the experiment is to test if developmental processes, such as cleavage and gastrulation, occur across a range of temperatures. In other words, the binomial state used in the model would be developed versus not developed^8,19,21. Fixatives such as 4% paraformaldehyde must be added to the samples at the thermal exposure period without any recovery time.

To ensure accurate count and determination of binomial state (live vs. dead; developed vs. non-developed), it is advisable to count the samples after the recovery time in random order to avoid potential observer biases. If there is sufficient personnel, different researchers could count replicate rows and compare their results. Alternatively, individuals can repeatedly count a small subset of the samples and check if the numbers are consistent.

Another potential limitation is the lack of error estimation of the LT₅₀ from independent samples²⁹. The current data analysis method provides a 95% confidence interval along the fitted logistic curve (Supplementary Coding File 1) and a standard error of the LT₅₀ (Supplementary Coding File 2). These error bounds are generated from the curve fitting process, not through multiple measurements of individuals from the sample population. Given the current heatblock design has six rows, one can fit data from each row to generate six LT₅₀ estimates and obtain the observation-based error estimates.

In summary, an accessible approach to determining acute thermal limits that can be applied to a wide variety of zooplankton is presented. This setup can be used to determine the thermal limits of various organisms and to pinpoint development stages that are vulnerable. This information can help improve the prediction of organismal performance and potential community interactions in the face of global climate change.

Materials

Name	Company	Catalog Number	Comments
0.45 µm membrane filter	VWR	74300-042
½” Acrylic sheet	McMaster-Carr	8560K266	Used to construct a ridged case with sufficient insulation.
1 mL syringe	VWR	76290-420
2 Channel 7 Thermocouple Types Datalogger	Omega Engineering	HH506A	Can be replaced with any thermometer that will fit inside a microcentrifuge tube
Automatic pipette	Ranin
Bolt- and Clamp-Mount Strip Heater with 430 Stainless Steel Sheath, 120V AC, 1-1/2" Wide, 100W	McMaster-Carr	3619K32
Crystal Sea Bioassay Mix	Pentair	CM2B	Use to make aritifical seawater
Denraster excentricus	M-Rep		Sand dollars from California
Dissecting microscope	Nikon	SMZ645
DIYhz Aluminum Water Cooling Block, Liquid Water Cooler Heat Sink System for PC Computer CPU Graphics Radiator Heatsink Endothermic Head Silver(40 mm x 120 mm x 12 mm)	Amazon		Connects to water bath and used to cool one end of the block.
Easy-to-Machine MIC6 Cast Aluminum Sheet 2" thick 8" x 8"	McMaster-Carr	86825K953	Machined to 2" x 6" x 8" with 60 equally spaced holes (11 mm dia., 42 mm depth) with two addition holes drilled in one side for thermostat probes.
Economical Flexible Polyethylene Foam Pipe Insulation	McMaster-Carr	4530K121	Covers the plastic tubing between chiller and block to reduce heat loss. Can be omitted if temperature range is close to room temperature
EVERSECU 72w 110-240v Aquarium Water Chiller Warmer/Cooler Temperature Controller for Fish Shrimp Tank Marine Coral Reef Tank Below 20 L/30 L Aquarium Chiller	Amazon		Can be used in place of the lab-grade water bath
Example with larval sand dollar
GENNEL 100 g Silver Silicone Thermal Conductive Compound Grease Paste For GPU CPU IC LED Ovens Cooling	Amazon		Improves the thermal conductance between the block and the heating and cooling elements.
Inkbird WiFi Reptile Thermostat Temperature Controller with 2 Probes and 2 Outlets, IPT-2CH Reptiles Heat Mat Thermostat (Max 250 W per Outlet)	Amazon		Monitors hot and cold ends. Maintains hot end in range
Lauda Ecoline Silver Air-Cooled Refrigerated Circulators	VWR	89202-386	Can be replaced with an aquarium chiller
Microcentrifuge Tubes	VWR	76019-014	If larger animals are used, scanilation vials (VWR 66022-004) is a good alternative
Nitex mesh filter	Self made		Used hot glue to attached Nitex mesh to 1/2" PVC tubing
Pasteur pipette	VWR	14673-010
Potassium Chloride (0.35 M)	Millpore-Sigma	P3911-500G
R statistical software.	The R Project for Statistical Computing
Syringe needle	VWR	89219-346	Depending on size of target organism gague 14 and 16 can be used
Tygon Tubing	McMaster-Carr	5233K65	Adjust to match the chiller and block used
Zoo Med Repti Temp Rheostat	Chewy.com		Rated to 150 W and rewired to feed directly into the heating element. Used to control rate of heat output