A method for studying the metabolic activity of individual tardigrades by measuring oxygen uptake using microrespirometry

Semi-terrestrial and tidal tardigrades have attracted considerable attention as in the dormant state called cryptobiosis they can survive extremely unfavorable environmental conditions such as drying, freezing, high salinity and even space conditions (Czerneková et al., 2017; Møbjerg et al., 2011; Wright et al., 1992). However, despite the great interest in the mechanisms behind this resilience, a reproducible experimental approach to study the metabolism of tardigrades on an individual level has been lacking (Brown et al., 2004; Nichols, 2005). Using individual specimens instead of cohorts to study metabolism has several advantages: (1) parameters influencing the single animal can be controlled to a higher degree, (2) the process is more specifically described using single specimens and (3) heterogeneity in the individual response to changes in the environment may exist, which will not be discovered when using a cohort for measurements (Drinkwater and Clegg, 1991; Szela and Marsh, 2005). Hitherto, studies addressing tardigrade respiration have used Cartesian divers. This method had limited success in measuring respiration rate as it produced inconsistent and unreliable results (Jennings, 1975; Klekowski and Opalinski, 1989; Pigon and Weglarska, 1953). More recently, microrespiratory methods using O₂ microelectrodes in microdiffusion tubes have been used successfully to determine the respiratory rates of a cohort of 3–5 specimen of foraminifera (Høgslund et al., 2008), copepod eggs (Nielsen et al., 2007) and individual zebrafish embryos (Bang et al., 2004).

In this study, we demonstrate that this method based on O₂ microsensors and applying Fick's first law of diffusion to microgradients in capillary chambers is also an excellent tool to study the metabolism of individual tardigrades. We used specimens of the species Richtersius coronifer (Richters 1903) and Macrobiotus macrocalix Bertolani and Rebecchi 1993 to demonstrate the power and the experimental potential of this approach.

Mosses containing the animals used in this study were collected from a limestone fence at Öland, Sweden, in July 2016 and kept in coffee filters under dry conditions at room temperature until use. The moss primarily contained tardigrades of the species R. coronifer, but also some individuals of the species M. macrocalix. The species were identified under a stereomicroscope, during the extraction phase. To extract tardigrades, moss was ground through a parsley cutter into the upper fraction of a sieve system with decreasing mesh size. The moss in the upper fraction was then flushed with demineralized water to facilitate the descent of the tardigrades through the mesh and to remove soil particles and plant fragments from their bodies. The highest density of tardigrades was typically found in the second-lowest fraction (mesh 120), and the material of this fraction was transferred to a Petri dish containing tap water and the tardigrades were collected using an Irwing sling and transferred to a salt jar with tap water.

All tardigrades were hydrated and left in tap water with access to moss overnight. Afterwards, they were incubated in tap water without access to moss overnight, before being placed in the capillary chambers for measurement of O₂ respiration rate. This procedure ensured the selection of tardigrades that remained intact after exiting cryptobiosis and which had approximately the same amount of gut content. In total, 28 live specimens of R. coronifer and a control group of 6 dead R. coronifer, as well as 17 specimens of M. macrocalix were measured at room temperature (22°C) in MilliQ water (<0.01 ppt). The dead R. coronifer were collected from the water surface, and death was confirmed based on the lack of movement under a stereomicroscope 24 h before being measured.

The capillary chambers used for the measurements were constructed by pulling 5 mm glass pipes over a Bunsen burner to produce an elongated segment at the center of the pipe with a diameter of approximately 660 µm. This segment was then cut from the larger pipe with a glass cutter and further cut into fragments of approximately 2.5 mm in length. The fragments were then sealed at one end by exposure to a horizontal Bunsen burner flame. The resulting capillary chambers were observed through a light microscope equipped with a Si CETi camera (Medline Scientific, Chalgrove, UK). Chambers were discarded if (1) they deviated significantly from the standard dimensions of 2.5×0.66 mm, (2) the sides of the chamber were not parallel, thus leading to a change in diameter along the length of the chamber, or (3) the opening of the chamber was cracked, therefore distorting the assumed diffusion area. An example of the characterization of the chamber labeled a1 can be seen in Fig. 1A,B, showing the two cross-sections used to calculate the radius of the chamber using its mean diameter. Characterizations of all chambers used in this study can be seen in Fig. S1. To transfer medium to and from the chambers, pipettes were constructed from 1 ml Luer plastic pipettes (Chirana, Stará Turá, Slovakia). The tip was heated using a Bunsen burner until it collapsed. Subsequently, the pipette was removed from the flame, and held with the tip downward, allowing gravity to elongate it. After solidification, the elongated tip was cut to fit the inside of the capillary chambers.

A schematic drawing of the set-up is shown in Fig. 1C and a typical image of the set-up containing an R. coronifer can be seen in Fig. 1D. To ensure a constant O₂ supply and to avoid evaporation of the medium during measurements, an aquarium set-up was established. This set-up consisted of a large cuvette, constructed by gluing five cover-slides together. A smaller cuvette that had been modified with air-holes near the bottom was glued upside-down to the bottom of the larger cuvette with aquarium-grade silicone. The aquarium was then filled with the same water as the capillary chambers. All measurements were done using MilliQ water that had been passed through a Q-Max PES 0.22 µm mesh. This was done both to minimize contamination and to reduce the risk of unknown substrates in the medium that could affect the metabolism of the tardigrade. The capillary chambers were fixated with modeling wax and also filled with water, using the customized pipettes. A single tardigrade was transferred to the chamber using an Irwing sling whilst being observed under a stereomicroscope. The chamber was then placed on top of the small cuvette in the aquarium, using tweezers. The entire system was placed on an adjustable platform under a microsensor that was fixed to a clamp with motor-control. Next to the platform was a cold-lamp for illumination during preparations for the measurements and a thermometer, to monitor the temperature during each experiment. All individual elements of this set-up, except for the tardigrades, were washed with 70% ethanol and rinsed with demineralized water, before being introduced into the rest of the system. The chamber was observed with a horizontal stereomicroscope to make sure the tardigrade was still present and that any air bubbles had been removed before measuring.

The microsensors were connected to an Ampere-meter with an A/D converter that transmitted the signal to a computer, through the program SensorTrace Pro (Unisense, Aarhus, Denmark). The sensors were calibrated in O₂-saturated MilliQ water and O₂-free ascorbic acid solution (∼0.5 mol l⁻¹) following the Unisense MicroRespiration System User Manual (https://www.unisense.com/files/PDF/Manualer/MicorRespiration%20System%20Manual.pdf; date accessed: 1 April 2020). Sensors were washed with MilliQ water between exposure to ascorbic acid and measurements. The sensor was observed under the horizontal microscope, whilst being lowered via a motorized micromanipulator into the chamber. The lowest possible depth within the parallel-sided part of the chamber that did not risk interfering with the tardigrade was identified and set as the depth of measurement in the SensorTrace Pro software. Profiling began at this point and a new measurement was taken for every step when moving the sensor up through the capillary chamber using the motorized micromanipulator. The step length was 100 µm. The measuring time was 1 s at each depth. For steady-state measurements, the system was left for 45 min after the tardigrade had been transferred and before measurements began; thus, a steady-state O₂ gradient could be fully established. Five to ten O₂ profiles were measured for each animal, depending on whether the gradient appeared fully established during the initial profiles. During profiling, the exact temperature and time of transfer for the tardigrade were noted in the SensorTrace Pro software. Before measuring the O₂ profile with tardigrades present, 5–10 O₂ profiles of the empty chamber were measured to ensure that nothing but the tardigrade was responsible for the development and maintenance of an O₂ gradient through the chamber. Fig. 1E,F shows screen captures of the profiles seen in SensorTrace Pro when measuring either in a tardigrade-free chamber or in one with a respiring R. coronifer at the bottom.

After completion of the measurements, images were taken of the chamber containing the tardigrade using a Leitz Biomed light microscope (Leica, Wetzlar, Germany) connected to the Si CETi camera (Medline Scientific) to determine the length and width of the tardigrade. Of the images recorded, the one that best showed the tardigrade dimensions was selected. To minimize distortion of proportionality resulting from the curvature of the glass, images with the tardigrade centered along both axes of the chamber were preferred. Images where the tardigrades longitudinal axis was perpendicular to the viewing angle of the microscope were also preferred. To better match the assumption that tardigrades have a cylindrical shape, images where the tardigrade body was fully extended were selected above images where the body was in a crouched position. A typical image for size determination can be seen in Fig. S2. The images were analyzed in the open source program Fiji (ImageJ). The scale was set to fit the scale bar of the image. Dimensions were determined using the measure function. Because of the imperfect cylindrical shape of the animal, 5 measurements of width were done at various points of the body and a mean value determined, whereas only one measurement of length was required.

The boxplots in Fig. 2 show the distribution of R_i, R_m and R_A in cohorts of the two species. A two-tailed heteroscedastic t-test was performed on the results to determine the significance of variance. The only significant difference between species was shown for R_m, where the mean (±s.d.) mass-specific O₂ respiration rate was 10.8±1.84 and 13.4±2.19 nmol O₂ mg⁻¹ h⁻¹ for R. coronifer and M. macrocalix, respectively. We also observed that the standard deviation was ∼33% for R_A, but only ∼18% for R_m.

From individual size determinations, the mean±s.d. length, width, mass and surface area was determined for both species. The species varied significantly (P<0.001) for all four measures of size. The values are summarized in Table 1, with similar size determinations published in Ramløv (1989) and Czerneková and Jonsson (2016). The values determined for R. coronifer in this study (length: 565.7±64.5 µm, mass: 14.30±2.77 µg), are consistent with the values reported in Ramløv (1989) (length: 570±56 µm, mass: 16±3.8 µg), while the study by Czerneková and Jonsson (2016) reported a greater average length of 653 µm. To determine how size affects the O₂ respiration rate, both R_i and R_m of each R. coronifer and M. macrocalix were plotted against their respective mass (see Fig. S3). For both species, R_i increased linearly with mass (R²≈0.7), whereas R_m was independent of mass (R²<0.1).

From the average surface area and mass determined in this study, the surface to mass ratio (hereafter A:m) was calculated for both species to determine whether the difference in size was in itself sufficient to explain the lack of significant differences in R_A compared with the significant difference seen for R_m between species. Because of their smaller size, the A:m ratio of M. macrocalix was 12.6% higher than that of R. coronifer. In comparison, the difference in R_m between the two species was 13.3% higher than the difference in R_A – leaving only 0.7% of the difference unexplained by the A:m ratio. Assuming that respiration occurs in all living cells of the organism, mass is generally a better metric to normalize respiration against than surface area, as long as the O₂ supply is not limited. After all, the surface area only includes 2 dimensions. As mentioned above, our interest in R_A as an expression of the O₂ respiration rate stems from the fact that tardigrades take up O₂ by passive diffusion across the cuticle, i.e. the surface area. If the cuticle either functions as a barrier or actively facilitates O₂ uptake, the R_A:R_m ratio should differ from the A:m ratio. This was not the case as a difference of 0.7% between the two ratios was not significant. Therefore, we suggest that O₂ uptake is driven by passive diffusion across the cuticle and depends on the permeability of the cuticle and the difference in O₂ partial pressure between the inside and outside of the animal. Together with the fact that R_m does not appear to vary with size within either species (Fig. S3B), we conclude that scaling is not a relevant factor for interpreting results of the O₂ respiration rate of tardigrades. Because of this, hereafter we will refer only to R_m as O₂ respiration rate.

Fig. S4 shows the variation inherent in the population against variation introduced by the applied technique to measure O₂ concentration. The two largest bars show the standard deviation as a percentage of the mean value (s.d.%) for R. coronifer as a population, with regard to O₂ respiration rate and mass. s.d.% within the population was 17% for O₂ respiration rate and 19% for mass. The two smallest bars show s.d.% of repeated measurements of the same individuals under similar conditions. The mean s.d.% of the 5 case studies was 1.4% for O₂ respiration rate and 8.7% for mass. Error bars represent the standard deviation of that mean s.d.% between the 5 individuals that were studied. To evaluate the reliability of the methods, 5 R. coronifer were randomly selected for case studies where multiple determinations of the size and O₂ respiration rate were analyzed. Standard deviations of O₂ respiration rate (s.d.) were calculated from 3–8 profiles in steady state for the 5 respective tardigrades. The standard deviation between these 5 s.d. values was calculated as a percentage of the mean s.d. (s.d.%). In each case study, size determination was carried out and the mass calculated from 4–6 different images of the same tardigrade at varying angles or positions. Mean mass and s.d.% between the different images were calculated for each of the 5 tardigrades. Finally, the mean s.d.% (±s.d.) for the mass determinations was calculated. The s.d.% of both O₂ respiration rate and mass was compared with the respective standard deviations calculated for the whole population (Fig. S4).

Here, we present and validate a method modified after Revsbech (1989) to study the O₂ uptake by single specimens of tardigrades during steady-state conditions. We showed that the mean O₂ respiration rate at 22°C was 10.8±1.84 and 13.4±2.19 nmol O₂ mg⁻¹ h⁻¹ for R. coronifer and M. macrocalix, respectively and that the O₂ respiration rate does not vary with size within either species. Measurements of O₂ respiration rate performed under conditions most similar to those of our study are those described by Jennings (1975) for Macrobiotus hufelandi obtained at 20 and 25°C, respectively, using Cartesian divers. In the Jennings study, the O₂ respiration rates were 8.15 and 9.24 nmol O₂ mg⁻¹ h⁻¹, respectively, which are similar to the values found for R. coronifer and M. macrocalix at 22°C in the present study. Barrionuevo and Burggren (1999) found that zebrafish (Danio rerio) embryos had an O₂ consumption rate of 2.4 nmol O₂ mg⁻¹ h⁻¹, which is approximately 4 times lower than the mass-based O₂ respiration rate determined for tardigrades in this study. However, it can be assumed that embryos have a somewhat lower O₂ respiration rate than adult animals, and according to Bang et al. (2004), the O₂ respiration rate reported by Barrionuevo and Burggren (1999) is probably somewhat underestimated. Thus, our results are in the same range as results reported from other ectothermic animals of comparable size.

The O₂ respiration rates determined for R. coronifer and M. macrocalix at 22°C in MilliQ water can be used as a reference when studying O₂ respiration rates under varying physiological conditions, i.e. the effects of temperature and osmolality on O₂ requirements in future studies. The fact that active R. coronifer have O₂ respiration rates 50 times higher than the rates of dead controls demonstrates that it is indeed the activity of the tardigrade body cells that drives the measured O₂ consumption and not microbial contamination. It is not clear that the higher O₂ respiration rate of M. macrocalix can be explained by its most obvious characteristics of smaller size and more rapid movement compared with R. coronifer.

We found no relationship between O₂ respiration rate and size within either species (see Fig. S3). We did, however, observe a selection bias toward certain sizes inherent in the sieve system commonly used for extraction of tardigrades from moss. In the study by Czerneková and Jonsson (2016), tardigrades were collected from fraction 40 of the sieve system, as opposed to fraction 120 used in this study and by Ramløv (1989). Thus, it is likely that the average size for the whole population differs from that found in any of the studies. However, as the size range of tardigrades used in the current study includes tardigrades larger than the average reported in Czerneková and Jönsson (2016) and there was still no correlation between size and O₂ respiration rate, we conclude that our results, despite the size bias, are representative for the in situ population as a whole. Furthermore, if the size can be regarded as a proxy of age and thus sexual maturity, it follows that those factors do not affect O₂ respiration rate significantly.

Measuring oxygen uptake using microrespirometry is a reliable method to study the respiration of a single specimen of tardigrades. The method as a strong potential for studying the metabolism of tardigrades under relevant physiological and ecological conditions including termination of cryptobiosis.

Thank you to the Aarhus University Centre for Water Technology (Watec); to Niels Peter Revsbech and Lars Riis Damgaard for their advice on the theory behind the microsensors; and to Lars Borregaard Pedersen for invaluable help on the practical challenges faced in the laboratory. Also, thanks to the Mars-group including Ebbe Norskov Bak, Per Nørnberg, Svend Knak Jensen and Jan Thøgersen, for interesting perspectives that kept the work engaging and exciting. Reinhardt Møbjerg Kristensen Zoological Museum, Denmark, and Roberto Guidetti, University of Modena and Reggio Emilia, Italy, are gratefully thanked for their help in determining the species Macrobiotus macrocalix.