Secondary metabolites produced by primary producers have a wide range of functions as well as indirect effects outside the scope of their direct target. Research suggests that protease inhibitors produced by cyanobacteria influence grazing by herbivores and may also protect against parasites of cyanobacteria. In this study we asked whether those same protease inhibitors produced by cyanobacteria also can influence interactions of herbivores with their parasites. 

We used the Daphnia-Metschnikowia zooplankton host-fungal parasite system to address this question because it is well-documented that cyanobacteria protease inhibitors suppress trypsin and chymotrypsin in the gut of Daphnia, and because it is known that Metschnikowia infects via the gut. We tested the hypothesis that Daphnia gut proteases are necessary for Metschnikowia spores to be released from their asci. We then also tested whether diets that decrease trypsin and chymotrypsin activity in the guts of Daphnia lead to lower levels of infection.

Our results show that chymotrypsin promotes release of the fungal spores from their asci. Moreover, a diet that strongly inhibited chymotrypsin activity in Daphnia decreased infection levels, particularly in the most susceptible Daphnia clones.

Our results support the growing literature that cyanobacterial diets can be beneficial to zooplankton hosts when challenged by parasites and uncover a mechanism that contributes to the protective effect of cyanobacterial diets. Specifically, we demonstrate that host chymotrypsin enzymes promote dehiscence of Metschnikowia spores; when cyanobacteria inhibit activity of chymotrypsin in hosts, this most likely traps the spore inside the ascus, preventing the parasite from puncturing the gut and beginning the infection process, and reduced the proportion of Daphnia infected.

This study illustrates how secondary metabolites of phytoplankton can protect herbivores against their own enemies.

Host-parasite system

Our study used 11 clones of Daphnia magna, which is a common species in ponds and lakes, especially in Europe. These clones come from different populations and have been used in studies focusing on the impact of cyanobacterial diets on Daphnia gut digestive enzymes (e.g., Schwarzenberger et al., 2021; see supplementary material and Table S1 for additional information about these clones). We used the common fungal parasite Metschnikowia bicuspidata (“Standard” isolate, originally isolated from Baker Lake in Barry County, Michigan). Daphnia become infected after consuming transmission spores they encounter in the water column when feeding. Infection takes place when the needle-shaped spore crosses the gut barrier and is not fought off by a host hemocyte response (Metschnikoff, 1884; Stewart Merrill and Cáceres, 2018). Once infection has taken hold, the fungus replicates within the hemolymph of the host (Stewart Merrill and Cáceres, 2018). The parasite reduces the fecundity and lifespan of infected hosts (Auld, Hall, and Duffy, 2012). Metschnikowia is an obligate killer, meaning it must kill its host in order to transmit to a new host (Ebert, 2005); transmission spores are released into the environment after host death, after which they can be consumed by a new host, completing the parasite’s life cycle.

Cultivation and preparation of phytoplankton food

We cultivated five strains of phytoplankton: one green alga and four cyanobacteria. We obtained the green alga Scenedesmus obliquus UTEX 3155 from the Culture Collection of Algae at the University of Texas at Austin. The cyanobacteria Microcystis aeruginosa strain CYA160/1 and strain CYA43 came from the Norwegian Institute for Water Research. Finally, the Microcystis aeruginosa strains wild type (WT) PCC7806 and mutant (MT) PCC7806 -mycB were obtained from the Pasteur Culture Collection of Cyanobacteria at the Institute Pasteur. The Microcystis strains used here vary in their production of microcystin and also their protease inhibitors (some effective at inhibiting chymotrypsin and others at inhibiting trypsin in the gut of Daphnia (Agrawal et al., 2005; Schwarzenberger, Ilić and Von Elert, 2021)). We used the CYA43 strain of Microcystis because it strongly inhibits chymotrypsin activity in D. magna (von Elert, Zitt and Schwarzenberger, 2012), but does not produce microcystins. We used the CYA160 strain because a prior study found it protected Daphnia dentifera against infection by Metschnikowia, yielding much lower levels of infection than Scenedesmus (Sánchez et al., 2019). We used the PCC7806 and PCC7806-mycB strains to allow us to explore variation among Microcystis strains in their effects on the infection process and to isolate whether there is an effect of microcystin, since PCC7806-mycB is a microcystin deficient mutant of the wild type strain (Dittmann et al., 1997); hereafter, we refer to these strains as PCC7806-WT (for wild type) and PCC7806-MT (for the microcystin-deficient mutant type).

We grew all cultures in chemostats with 24-hour light. Scenedesmus was grown in standard COMBO media, Microcystis CYA43 and CYA160 were grown in Z8 media, Microcystis PCC7806-WT in BG11 media, and Microcystis PCC7806-MT was grown in BG11 media with 5µg/mL chloramphenicol (which is necessary to keep microcystin genes inactivated; (Dittmann et al., 1997)). Food for experiments was prepared weekly. We harvested cells in 50 mL Falcon tubes and centrifuged (Sorvall St 16, ThermoScientific) for 10 min at 3000 rpm. After, we decanted media from the tubes and resuspended cells in ADaM (Klüttgen et al., 1994). The volume was adjusted to a concentration of 100 mg C/L and this solution was used to feed animals a final concentration of 2 mg C/L. Microcystis PCC7806-MT was rinsed and spun down twice with milliQ water before resuspending in ADaM to rinse out any chloramphenicol that may have been left from growth media.

Experiment 1. Dehiscence assay: isolating the impact of gut extracts and chymotrypsin on spore dehiscence.

To test whether diet impacted dehiscence of Metschnikowia spores, we exposed them to extracts from the guts of animals that were fed Scenedesmus or from the guts of animals that were fed Microcystis diets that are known to inhibit chymotrypsin activity in the gut. If chymotrypsin promotes spore dehiscence, we expected higher levels of dehiscence for spores exposed to gut extracts from Scenedesmus (which should have normal chymotrypsin activity) than from Microcystis. To further test the impact of chymotrypsin, we also added pure chymotrypsin (CAS 9004-07-3, Sigma-Aldrich) to the gut extracts of animals fed Microcystis, to see if we could recover the levels of dehiscence seen for spores exposed to gut extracts from animals fed green algae. This assay was done using extracts from the guts of the ‘May 20’ Daphnia magna clone, because prior work indicated the susceptibility of this clone differed greatly between Scenedesmus and Microcystis diets.  

We reared 5 individuals of clone ‘May 20’ per 100 mL beaker filled with 50 mL of filtered lake water, feeding them 2mg C/L Ankistrodesmus falcatus AJT strain (Tessier et al., 1983; Schomaker and Dudycha, 2021) until they were 5-6 days old. At this time, we moved the animals to clean filtered lake water, keeping 5 individuals per beaker, and fed them 2 mg C/L of either Scenedesmus or Microcystis aeruginosa CYA43; we had 20 replicate beakers for the Scenedesmus treatment and 40 replicate beakers for the Microcystis treatment. We had twice the number of Microcystis treatment replicates because we needed extra animals to make gut extracts – 100 of these animals would be used for the regular gut extract treatment, and 100 would be used for the gut extract + chymotrypsin treatment, as described below. On the next day, when animals were 6-7 days old, we again fed them 2 mg C/L of their treatment diet. On the following day we placed animals in 2mL Eppendorf tubes, placing 20 animals per tube (combining four of the beakers into a single tube) for a total of 5 tubes with 100 animals total for the Scenedesmus diet treatment, and 10 tubes with 200 animals total for the Microcystis treatment, and stored the tubes with animals at -20 °C. We later dissected the entire guts of animals and placed 20 guts in a tube, again for a total of 5 tubes with 20 guts each for the Scenedesmus diet treatment and 10 tubes with 20 guts each for the Microcystis diet treatment. We added 100 µL of 0.1 M potassium-phosphate (P-P) buffer pH 7.5 to each of the tubes and homogenized them using a pestle. We centrifuged the tubes at 14,000 g for 3 minutes, then transferred the supernatant into a 2mL Eppendorf tube. We used bovine chymotrypsin (Sigma Aldrich) as a positive control, resuspending it in P-P buffer to a concentration of 10% w/v. Finally, we used P-P buffer as our negative control.

We tested the impact of these gut extracts (or controls) on parasite transmission spores. Specifically, we measured how many spores lost their ascus over time. The extracts and the concentrations we tested are the following: 5% w/v Scenedesmus gut extract, 5 % w/v Microcystis gut extract, 5% w/v Microcystis gut extract with 1% w/v bovine chymotrypsin, 1% w/v bovine chymotrypsin, and finally, P-P buffer as a negative control. We predicted that, if chymotrypsin is important for dehiscence, the highest level of dehiscence (that is, most spores without asci) would occur in treatments with high levels of chymotrypsin activity, and the lowest levels would be in treatments with low activity levels of chymotrypsin (inhibited due to diet). More specifically, we predicted that spores incubated in gut extracts of animals fed Microcystis would have lower levels of dehiscence than those incubated in gut extracts of animals fed Scenedesmus, and that there would be high levels of dehiscence of spores incubated in pure chymotrypsin or in those where chymotrypsin was added to gut extracts of animals fed Microcystis.

To test this, we placed 250 µL of spore slurry containing ~ 100,000 spores of the fungal parasite Metschnikowia bicuspidata in wells of a 96 well plate. We did this for 20 wells placed randomly throughout the plate. One person would place a treatment picked haphazardly to one of the wells in the plate and note the time at which the treatment was placed. After 30 and 210 minutes another person would count the number of spores with and without asci within a given well without knowing which treatment was placed in that well. The counts were done under a compound microscope at 400X with a Neubauer counting chamber. This was repeated for every treatment for a total of 4 replicates per treatment.

Experiment 2. Quantification of effects of diet treatments on gut digestive enzymes and infection levels

Because susceptibility and tolerance to protease inhibitors are Daphnia clone and diet dependent, we wanted to measure variation in protease activity on different diets and test how these gut enzyme levels corresponded with host susceptibility and parasite fitness. To quantify the impact of diet on gut proteases, we measured trypsin and chymotrypsin activity of gut extracts from animals fed different diets. We also measured variation in infection outcomes for animals fed on these different diets. We were interested in both the variation in gut protease activity and in whether reduced protease activity correlated with decreased infection.

We ran a factorial experiment with 11 Daphnia clones and 5 diets. We ran this experiment in 7 blocks aiming for 5 replicates per clone x diet treatment per block. We interspersed blocks that were used to measure enzyme activity and those used for infection assays. Blocks 1-3 were used entirely for infection assays, blocks 4 and 6 entirely for enzyme assays, and blocks 5 and 7 were used for both infection assays and enzyme activity; for blocks 5 and 7 we aimed for 10 replicates per treatment and half of those replicates were used for the enzyme activity assay. While we aimed for 5 replicates per block, in some cases, we lost animals, especially due to cyanobacteria toxicity; in these cases, we tried to replace the replicate in a future block. This was particularly problematic in blocks 1 and 2. For the two PCC7806 diet treatments, very high mortality means that the dataset does not include any animals from those two blocks for these two diet treatments in our analyses.

We first reared individuals of each clone under standardized lab conditions for multiple generations prior to the experiment. For all blocks, neonates (0-1 day old) were harvested from mothers and placed 5 each in 150 mL beaker with 100 mL of filtered lake water. Neonates were fed Ankistrodesmus falcatus daily for 5 days. When juveniles were 5-6 days old, each juvenile was placed individually in a 50 mL beaker with 30 mL of filtered lake water. Juveniles were fed their corresponding treatment diet as follows: 2 mg/L C of Scenedesmus, 2 mg/L C of Microcystis CYA160, 2 mg/L C of Microcystis CYA43, 1 mg/L C Microcystis PCC7806-WT with 1 mg/L C Scenedesmus, and 1 mg/L C Microcystis PCC7806-MT with 1 mg/L C Scenedesmus. For the two PCC7806 treatments, we needed to use a 50:50 mix of the treatment diet:Scenedesmus to promote survival. The first two blocks of the experiment used 100% of the Microcystis strains, but the very high mortality meant we could not use those animals in the experiment. Therefore, for blocks 3-7, animals in these two treatments were fed a 50:50 mix of the Microcystis strain and Scenedesmus.

For animals in blocks 4 and 6 and for the half of the block 5 & 7 animals that were used to measure protease activity, on the next day (that is, when animals were 6-7 days old), we transferred animals to clean beakers filled with 30mL of fresh filtered lake water and fed them half the amount from the previous day of their corresponding treatment food; this was done to be consistent with the treatment of animals for the infection assays (see below). After 24 hours, we sacrificed these animals and preserved them in Eppendorf tubes with no water and stored them at -20 °C. At a later date, animals were grouped to have 20 animals per clone x diet treatment (for a given clone ⅹ diet treatment combination, animals from blocks 4-7 were combined into one tube) in a given tube and used to measure proteolytic activity of trypsin and chymotrypsin. Trypsin activity was assayed using the substrate N-benzoyl- DL-arginine p-nitroanilide (BapNA), while chymotrypsin activity was assayed using N-succinyl-L-alanyl-L-alanyl-L-propyl-L-phenylalanine 4-nitroanilide (SucpNA). Detailed methods used to quantify potential enzyme activity (as ΔmAU/min/µg protein content) are given in the supplement.

For the five infection assay blocks, when animals were 6-7 days old, we transferred them individually to clean beakers filled with 30mL of fresh filtered lake water. On this day all animals were exposed to 500 spores/mL of Metschnikowia and fed half the amount from the previous day of their corresponding treatment food; prior work has shown that reducing phytoplankton during exposure promotes spore uptake by hosts. After 24 hours, we placed each animal singly in a 150 mL beaker with 100 mL of filtered lake water and fed them their corresponding treatment diet with the original amount of their treatment diet food. We checked for mortality daily and counted offspring twice a week during water changes, removing offspring from the experiment. Animals were fed their treatment diets daily. Animals that died throughout the experiment were preserved in 100 µL of MilliQ water and stored at 4° C for later spore counts. At the end of 20 days, any remaining animals were preserved in 100 µL of MilliQ water and stored at 4° C for later spore counts. We determined the final abundance of transmission spores in hosts by counting spores under a compound microscope at 400X using a Neubauer counting chamber. Animals were diagnosed as infected if they contained transmission stages (equivalent to “terminal infection” in Stewart Merrill et al., 2019).