Microzooplanktonic grazers – a potentially devastating threat to the commercial 1 success of microalgal mass culture 2 3

13 Eukaryotic microalgae and prokaryotic cyanobacteria are globally the most important primary producers, 14 forming the base of food web in aquatic ecosystems. As such, they are eaten by a huge diversity of 15 protistan taxa (e.g., amoeba, flagellates and ciliates), as well as zooplanktonic and larger metazoan grazers. 16 As in terrestrial agriculture, grazing has the potential to devastate the microalgal “crop” and this has 17 obvious implications to the commercial success of the developing microalgal industry. Whilst in 18 conventional agriculture thousands of years of exploitation of a relatively small number of crop plants, has 19 resulted in tools, knowledge and strategies that can manage this issue, in the case of microalgal mass 20 culture this is relatively undeveloped. This review explores our current understanding of the issue and 21 where further research is needed, focusing on the diversity of grazers and how microalgae under various 22 environmental regimes and culture conditions avoid being annihilated. In addition, the implications of 23 algal mass culture, where the objective is to maintain a virtual monoculture, are discussed in the context of 24 how infection could be prevented/ minimised and if infection occurs, how this may be managed to prevent 25


7
Phaeocystis globosa which may be induced to form colonies by the heterotrophic dinoflagellates Noctiluca 162 scintilans and Gyrodinium dominans [75]. The poor nutritional quality of the colonies can result in low or 163 inefficient feeding by copepods [76]. To date, there has been a limited amount of work on the specificity 164 of microzooplanktonic grazer-induced phenotypic plasticity, although the study by Luo et al. [77] indicates 165 that there is a specific interaction, at least at the genus level, although it was observed that different strains 166 of the same species demonstrated different levels of response. In field samples of Micractinium pusillum, 167 the alga forms characteristic cuboidal or tetrahedral colonies with long and strong bristles, much more 168 frequently than in cultured samples, which on long-term routine maintenance in culture collections lose 169 their bristles/ spines [78]. Medium from cultures of B. calyciflorus was demonstrated to induce long 170 bristles (50-100 µm) in M. pusillum; however, medium from Daphnia cultures failed to induce bristle 171 formation [77]. On testing a related chlorophyte, Chlorella vulgaris, no effect was observed on challenging 172 cultures with samples of either grazer. Further work is needed to gain greater understanding of the nature 173 and function of info chemicals and their capacity to induce protective phenotypic change in algae, but the 174 implications to their use to protect algal crops has obvious potential. 175 176 3.2 Chemical antipredator defence 177 It has been known for many years that microalgae are capable of producing bioactive metabolites, 178 with one of the first reports being on the production of an antibacterial fatty acid, Chlorellin, isolated from 179 cultures of C. vulgaris [79]. The bioactive metabolites produced by algae encompass a wide range of 180 chemical entities, with activities against viruses, bacteria, fungi, protozoa, as well as potential use in the 181 treatment of cancers and other health problems [80,81,82]. It may be speculated that algae produce these 182 chemicals to give them an ecological advantage and it is known that in natural ecosystems substances can 183 be released by cyanobacteria into the waterbody that influence feeding rates of zooplankton, as 184 demonstrated by Ostrofsky et al. [83] who reported inhibition of feeding rates by compounds released by 185 Anabaena flos-aquae under both lab and field conditions. This phenomenon appears to be relatively widely 186 spread across both cyanobacterial and microalgal taxa and it would appear logical to assume that toxin 187 production is specifically related to grazing inhibition. However, published evidence suggests otherwise, 188 for example the systematic study undertaken by Wang et al. [84]. This study investigated the effects on 189 the rotifer B. plicatilis of 10 strains of Alexandrium including eight toxic (PSP-toxin producers) and three 190 non-toxic isolates, their media and material from lysed cells. The feeding experiments demonstrated that 191 the rotifers readily ingested cells of a non-toxin containing isolate of A. tamarense, as well as A. 192 lusitanicum, and A. minutum, both of which contain PSP-toxin, but grazed relatively less, or not at all, on 193 the other seven Alexandrium strains tested. Ingestion of PSP toxin containing cells from these two known 194 toxic species had no impact on rotifer mortality; however, even though no PSP toxins were found in 195 Alexandrium sp1 and Alexandrium sp2, these resulted in a collapse of the rotifer populations, as did the 196 other five toxin containing isolates tested. It thus appears likely that toxic mechanism(s) other than PSP 197 were responsible for causing the lethal effects on the rotifers. 198 To date, in most cases, there is little published on either the structure or functional mode of algal-199 produced antipredator chemicals. However, a significant body of work has been undertaken on the model 200 grazer, the dinoflagellate Oxyrrhis marina, which on ingestion and subsequent lysis of the alga Emiliania 201 huxleyi, converts dimethylsulphoniopropionate (DMSP) from the alga, to dimethyl sulphide (DMS) via the 202 enzyme DMSP lyase [85]. In a subsequent study [86] it was demonstrated that this reaction deters 203 protozoan herbivores, presumably through the production of highly concentrated acrylate, and that O. 204 marina differed in its ability to ingest and survive on algae with high-activity DMSP lyase, but 205 preferentially select strains with low enzyme activity when offered prey mixtures. This response was not 206 restricted to a single grazer, as addition of DMSP has also been reported to reduce grazing on E. huxleyi by 207 the dinoflagellates Amphidinium longum and Gymnodinium sp., as well as the ciliate Coxliella sp. [87]. 208 An alternative protective strategy involves the release of chemicals that do not prevent, or reduce, grazing 209 on the alga, but reduce fecundity of the grazer. Diatoms are major primary producers at the base of the 210 marine food web and their main predators include the herbivorous copepods. Secondary metabolites, 211 including defensive oxylipins, released by these algae immediately after grazing-induced cell damage, are 212 targeted not against the predators themselves, but rather at interfering with their reproductive success [88]. 213 This strategy is obviously slower acting than directly killing the grazer, or deterring it from ingestion of 214 algal cells, but has the potential to facilitate managed coexistence of the algae and grazers. 215 9 mode(s) of action. The possibility of harness the potential of the natural "chemical warfare" between the 218 algae and their grazers has commercial implications and could be a crucial component in the management 219 of future algal production facilities. Clearly, this is a future area of research that warrants further 220 investigation. 221 222

Microalgal mass culture 223
In the context of commercial exploitation the scale of production in algal biotechnology can vary 224 hugely with 10's or 100's of litres being appropriate in the context of the needs for small-scale aquaculture 225 applications, but for other products considerably larger volumes/ foot-prints are needed. For example, the 226 βcarotene production facility in Hutt Lagoon Australia is ~740 hectares and future algal biofuel facilities, 227 in order to be economically viable, will need to be significantly larger again. Irrespective of the 228 volume/size of production plant, or whether the algae are being cultivated in closed photobioreactors, or 229 open pond-based systems, infection by microzooplanktonic grazers poses a threat that could potentially 230 devastate the algal crop, or at best reduce productivity and hence commercial viability. 231 232

Detection 233
By the time an algal culture is being heavily grazed the effects may be obvious by eye as a gross 234 change in colour, with a reduction in the green colouration as the algal cell density reduces and in extreme 235 cases an increase in brown colouration as the grazers "bloom" and the algae die and lyse. Whilst changes 236 in colouration, optical density, or turbidity could all form the basis of an automated monitoring system, 237 these are unlikely to be sufficiently sensitive to detect the early stages of infection. Therefore, by the time 238 the grazers have been detected it may be too late to apply any strategies that may be available to control or 239 manage the infection. It is, therefore, important to be able to have an "early warning system" one can 240 potentially prevent grazing becoming a problem. 241 Microscopy is the most obvious approach to detection and enumeration. Sample size may vary, but is 242 restricted in the case of Haemocytometers to <1.0 µl, or if a larger chamber such as a Sedgwick Rafter cell 243 [89] is employed to 1.0 ml. Additionally, the size and configuration of the Haemocytometer may preclude 244 its use in detecting or enumerating large grazers including rotifers and copepods. An alternative to examine 245 larger volumes (50-100 ml) is to fix samples and then employ sedimentation chambers, where the samples 246 can be observed using an inverted microscope [90]. This approach has the advantage that much lower 247 densities of grazers may be detectable directly from the sample. However, the samples need a settlement 248 phase, generally 12 -20 h, so there is a significant delay in being able to detect, or enumerate, any grazers 249 present. Furthermore, from the authors experience it is problematic when algal cell densities are high (> 1 250 x 10 6 cells ml -1 ) as the algae may "bury" the grazers and although it works well for larger zooplankton, 251 many protozoans will lyse on being treated with some fixatives. However, in many cases employing  An alternative strategy that could be employed is to use a targeted molecular approach whereby 269 known individual grazers, or groups of grazers could be identified employing multiple polymerase chain 270 reaction (PCR)-based tools. On using allele-specific probes and monitoring contaminants using 271 Quantitative PRC (QPCR) "weedy" invader algae may be detected at levels as low as one in 10 8 cells in a 272 culture [93]. Furthermore, this study demonstrated that the QPCR method developed was 10 4 times more 273 sensitive than flow cytometry in the detection of Tetraselmis striata cells serially diluted in 274 Nannochloropsis salina culture [93]. The approach has also been employed to detect the parasitic ciliate 275 Cryptocaryon irritans, which causes "white spot disease" in marine fishes, from the natural environment at the authors that ultimately molecular approaches will be the most probable way forward for the early 284 detection of grazers in algal crops. 285 Irrespective of the detection method that may be employed, unless the option for on-line continuous 286 monitoring is available, the key to early detection will be the instigation of a suitable sampling regime. 287 This will need to be tailored to the requirements of the individual production system, but it is suggested 288 that a minimum of daily sampling at a number of points in each production unit (pond, bioreactor, etc.) is 289 needed. Furthermore, sampling of sediments and/or any biofilms, if present, should be included, as these 290 may be potential sources of inocula that could result in infection. In addition, effective grazer DNA/RNA 291 sample preparation is critical, particularly in high-density microalgal cultures where the cell density of 292 microalgae may be much greater than that of the invader grazers at early stages of infection. This makes it 293 difficult to get enough grazer DNA/RNA for an effective gene-level analysis, as the grazer DNA/RNA 294 materials may be lost or masked by the overwhelming amount of microalgal DNA/RNA isolated from an 295 infested sample. A method for selectively separating or enriching grazer DNA/RNA from an algae-296 dominant background needs to be developed. 297 298

Prevention of infection 299
There is a limited number of publications specifically suggesting how to reduce, or prevent, infection 300 of algal crops by grazers. The most obvious approaches involve engineering orientated solutions and/or 301 having housekeeping procedures in place that reduce the risk of initial infection. For closed 302 photobioreactors these are more readily achievable. An appropriately high specification photobioreactor, 303 ideally with the possibility of sterilizing both the unit and medium, with all manipulations undertaken 304 aseptically will theoretically not become infected. However, such ideals are not practicable on scale-up, 305 due to their cost implications, unless the product is of high commercial value. In reality, cleaning and 306 sanitisation rather than sterilization is achievable and these procedures must be tailored to the unit with the 307 objective of minimising the possibility or introduction of cysts/eggs of microzooplanktonic grazers. At the 308 planning stage engineering options that minimise connections between separate production units to avoid  (Table 2). Direct engineering solutions to prevent or reduce algal grazing are 322 fraught with difficulties and physical separation is not always possible, as in many cases both microalgae 323 and their grazers are similar in size (preventing size-fractionation or centrifugation approaches). These 324 techniques are in addition relatively expensive and time consuming, so some tested engineering 325 approaches will not be practicable above lab-scale. However, cavitation induced by ultrasonication has 326 13 significant potential and has been demonstrated to be effective in inactivating large zooplankton, such as 327  The above treatments are not likely to be applicable at scale and alternative more readily scalable and 351 cost-effective approaches are required. Options that involve temporarily changing environmental 352 conditions such as osmotic potential, pH, or temperature, all have significant engineering and cost 353 implications. However, where systems have been installed to increase CO2 levels to boost algal 354 14 productivity these could in parallel be employed to control the grazer population. Increasing pond night-355 time CO2 concentration by gas injection of pure CO2, has been demonstrated to result in the rapid control 356 of a zooplankton bloom in a 1.5 m 3 microalgal cultures [104]. Furthermore, the authors [33] have recently 357 investigated the effect of culture pH (i.e., 6.0, 6.5, 7.0 and 7.5), maintained by supply of compressed air 358 bubbles containing various concentrations of CO2, on Chlorella sorokiniana culture stability in the 359 presence of a grazing threat by the flagellate P. malhamensis and several other protozoa. When CO2 levels 360 were low, culture crashes due to grazing were observed, and it was speculated that increased CO2 partial 361 pressure in the culture medium enhanced diffusion of CO2 into the cytoplasm of P. malhamensis, reducing 362 the intracellular pH leading to cell death [33]. This approach has been successfully trialled in the field in a 363 100 l raceway pond where increasing the CO2 supply to a C. sorokiniana culture reduced the pH from 6.8 364 to 6.0 leading to a significant reduction in P. malhamensis numbers and elimination of other grazers within 365

h of initiating the treatment [33]. 366
An alternative approach that is included here for completeness is to employ top-down biological 367 control, such as the introduction of higher trophic predators. This could include larger copepods or 368 predatory rotifers. This approach has been employed by Mitchell and Richmond [105] to use rotifers to 369 remove microalgal contaminants from Spirulina; however, for many potential algal crops using this 370 approach to target grazers is likely to be problematic as they will invariably consume both microalgae and 371 their grazers. The alternative bottom up biological control approach of introducing targeted grazer 372 pathogens is equally problematic. Furthermore, the authors are unaware of any known, cultivatable, target-373 specific protozoan or other microzooplanktonic pathogens, or viruses, which could be introduced to 374 instigate biological control. 375

Further approaches to crop protection 376
It is certain that there is no "one size fits all" solution(s) to the prevention and control of grazers in 377 microalgal mass culture. Furthermore, there is a great deal of work needed before effective strategies are 378 available and the likelihood is that bespoke solutions will be required for each alga and production plant. 379 Prevention of infection is invariably more likely to be effective than treatment, or the management of 380 contaminant grazers, but there are constraints with respect of the economic viability of many potential 381 products if high cost, sterilisable, bio-secure bioreactors have to be employed. A number of important gaps 382 in our current knowledge and opportunities have been identified above and these are synthesised in the 383 three paragraphs below. 384 Firstly, "know your enemy", although there is a considerable body of work on different grazer taxa, 385 there is still, to some extent, a lack of coordinated research that interconnects all aspects of this topic in the 386 context of algal biotechnology. Montagnes et al. [122] started this process but were specifically discussing 387 natural populations focussing on the needs for research to understand: searching, contact, capture, 388 processing, ingestion, digestion at community, population, and individual levels, and the need to develop 389 linkages in research focussing on food selection, food webs and modelling. Understanding the interactions 390 between the organisms involved and the capacity to manage mixed populations to the extent that grazing 391 of the crop is not at a level that results in either catastrophic crop loss, or a reduction in economic viability, 392 will be crucial to sustainability in the microalgal industry.