Recycling biofloc waste as novel protein source for crayfish with special reference to crayfish nutritional standards and growth trajectory

Screening of novel feedstuffs, that too for data-deficient (nutritionally) animals, is somewhat ambiguous or problematic. Through systematic meta-analyses, the present study formulated most up-to-date crayfish nutritional standards, against which a recyclable waste (biofloc biomass, BM) from intensive aquaculture systems was assessed as a novel protein source. Growth trajectory dependencies and thermal growth coefficient qualifying for good growth in crayfish (TGC 0.5–0.64 units) were benchmarked. Using these standards and a 7-week growth trial, BM's suitability as a novel protein source for red swamp crayfish Procambarus clarkii was evaluated through its graded inclusions in a commercial feed. Results suggest that BM can elevate growth at 33–66% inclusion in existing feed formulations. Beyond 66% inclusion, BM can deteriorate growth in crayfish due to high ash content (exceeding physiological limit > 14%), arginine deficiency (~ 14–20% lower than an optimum requirement), and insufficient non-protein energy: protein ratio (3.7 cal mg−1). Arginine is perhaps the most critical amino acid in dietary protein for crayfish, and deficient in BM. Although no critical bioaccumulation levels of heavy metals were breached by feeding 100% BM to crayfish, a mineral and heavy metal (Hg) stress seemed plausible. Crayfish raised solely on biofloc may not realize full growth potential.

meta-analyses, crayfish' optimum dietary nutritional requirement is tabulated as crayfish standards in Table 1. It is also compared with established standards of penaeid shrimps, often assumed as a template for most crustacean diets. Detailed information in this regard can be found in the supplementary material. In terms of crayfish growth trajectory, their thermal growth coefficient (TGC) may vary from 0.07-1 unit (interquartile range, IR 0.32-0.64 units). Results suggest any TGC in the range of 0.5-0.64 units may be regarded as 'reasonably good growth' in crayfish. Further insights into crayfish growth trajectory and its nutritional dependencies are presented in detail in the supplementary material. The information synthesized and approach used may serve as a template for future researchers exploring three less-established or unknown dimensions simultaneously (as in the present study)-novel feedstuff, optimum nutrition, and data-deficient (nutritionally) animals.
Growth response of crayfish to biofloc protein. Following a 9-week growth trial with graded BM levels in the diet, differential growth response by crayfish was realized (Fig. 1). Except for control and BM 33 groups, crayfish' final body weight showed a significant deviation from the normal distribution. Further examining the skewness of final body weight distribution in BM 66 and BM 100 groups, it was apparent that these groups were dominated by runts (smaller sized individuals) with large size deviations from the handful of bigger individuals. The size heterogeneity showed a significant and negative correlation with BM inclusion in the diet (Pearson's 2-tailed r = − 0.63, p < 0.05). Size heterogeneity in crayfish may aggravate community aggression 25 . However, the diet-driven size heterogeneity was not significantly correlated with mortality. The dietary treatments did not cause significant differences (p > 0.05) in survivability, confirmed by post-hoc analyses. Overall, the survivability remained > 70% through the experimental period in all groups ( Table 2). It implies-BM does not pose a significant mortality risk to crayfish stocks irrespective of inclusion levels, but it has implications on growth (presented below).
The growth in terms of TGC, live-weight gain (LWG), and body weight (BW) were significantly depressed (p < 0.05) in the BM 100 fed group. In contrast, the growth in control, BM 33, and BM 66 groups were higher with insignificant differences among them (p > 0.05) ( Table 2, Fig. 1, 2). A statistically insignificant dampening of growth rate over time (p > 0.05) was observed in groups BM 66 and BM 100 (Fig. 2). At the end of culture (63 days), the realized TGC in crayfish fed on BM 100 was on an average two times lower (p < 0.05) than the growth exhibited on control, BM 33, Table 2, and the relationship of feed utilization parameters in response to BM inclusion is depicted in Fig. 3. Interestingly, the calculated FCR(s) of our respective diets, when multiplied with the dietary arginine content, seem to 'hit the target' of arginine requirements by crayfish (e.g., FCR of BM 100 × Arginine in BM 100 = Fulfillment of arginine requirement). As per the crayfish growth trajectory (quantified in the previous section), the group fed on 100% BM failed to show reasonably good growth. They were dominated by smaller-sized runts, poorest of the FCR and PER, but no significant mortality. Among the limited studies testing flocculated microbial meals in crustacean diets [reviewed in 8], BM inclusions were mostly up to 10-30% (of the total diet) or 30% (of fishmeal replacement). Good results in terms of growth were usually obtained at the maximum inclusion levels 8 . Like the present study, two previous studies had tested BM (on Litopenaeus vannamei) at a broader inclusion level from 17 to 84% of the total diet 17,26 . Despite different target species, the results seem close to that of the present study. Above 41-53% BM inclusion, the growth advantages were gradually lost 17,26 . Looking deeper into the aspects of our BM 100 -protein compared to control, BM 33 , or BM 66 -protein, the arginine seems to be a bottleneck for reasonably good growth (Tables 2,  4). Other EAAs, which could also be critical (e.g., methionine and lysine), were comparable-to-higher in BM 100 than in other diets (Table 4). Although methionine and lysine levels in diets fell short of our formulated crayfish nutritional standard (Table 1), at least it fulfilled penaeid EAA standards of NRC 4 . It hints that NRC's penaeid EAA standards cover well for most of the EAA requirements in crayfish, except for arginine (and tryptophan could not be judged). Arginine levels in BM (Table 4) neither fulfilled crayfish nor penaeid standards (Table 1).
Biofloc has been previously criticized for being partly deficient in arginine [27][28][29] . The arginine coefficient (proportion of total protein, in %) of biofloc meals, be it commercial ones like Novacq (2.38% 19 ), FeedKind (2.54% 20 ), or in the present study (2.73%) seem to have close resemblance (CV 5.5%). If we consider the mean arginine coefficient of BM from these data (2.55%) and tally it to fulfill the optimum arginine requirement of crayfish (minimum 1.8%), the crude protein level of such BM should be at least 70%. It is beyond the expected range of ordinary bioflocs 22 . BM harvested from high TSS systems (due to infrequent sedimentation or water exchange) can have lower protein content 10 . For example, the crude protein content of a biofloc can drop by − 34.5% if the Table 1. Optimum dietary nutritional requirement of freshwater crayfish and its comparison with NRC (2011) standards for penaeid shrimps (usually adopted as status quo). *In parentheses-proposed reconsideration of calculated standards, based on high TGC obtained in the present trial. **Digestible values converted to crude values assuming 90% apparent digestibility. www.nature.com/scientificreports/ TSS of the system is let to increase from ≤ 200 mg L −1 to 800-1000 mg L −110 . As such, BM harvested from a low TSS system would have higher arginine (0.72%) compared to a high TSS system (0.47% arginine) (recalculated from 10 ; using mean arginine coefficient = 2.55% of total protein). Even with aging biofloc, the content of arginine (also other EAAs) may decline. For example, from the 10th day to the 30th day of a biofloc culture, the arginine levels can decrease by 25-41% (recalculated from 27 ). However, some specially produced commercial flocculated meals can have a high arginine coefficient (e.g., 5.3% of the protein in ProFloc 17 ). Among all the EAAs, arginine content in red swamp crayfish seems maximum 3,21,30,31 , indicating a supposedly higher arginine demand in crayfish. The same is true for marbled crayfish Procambarus virginalis 32 . Arginine is perhaps the most limiting EAA in most crustacean diets and is required between 1.6-2.7% of diet 33 . Due to the poor activity of the urea cycle in crustaceans, arginine is indispensable for growth 33,34 . Arginine functions as a phosphagen in crustaceans, being the only amino acid providing amidino group for the synthesis of creatine-a major reserve of high-energy phosphate for ATP regeneration 33,35 .   www.nature.com/scientificreports/ Table 3. Heavy metals and mineral content (mean ± SE; dry matter basis) in the tail muscle and hepatopancreas of red swamp crayfish Procambarus clarkii fed graded levels of biofloc meal. BDL = below detection limit (Mn and Fe: < 2 mg kg -1 , Cd: < 0.002 mg kg -1 ); different letters in superscript denote groups with significant differences as derived from Tukey's HSD multiple range test (α = 0.05). www.nature.com/scientificreports/ Risk of heavy metals bioaccumulation or mineral stress from biofloc meal. The contents of heavy metals in BM were below the critical pollution limits. No critical limits were breached in the crayfish body that could qualify BM as a feedstuff capable of inducing unsafe heavy metal biomagnification, rendering them unfit for consumption. Content of Cd and Mn were mostly below the detection limits (Table 3). Except for mercury, hepatopancreas contained a higher amount of heavy metals (and minerals) than muscle. Hepatopancreas of crayfish, like most crustaceans, have been reported to be major storage of minerals, including heavy metals 3,37 .
With increasing BM fraction in the diet, the concentration of Hg significantly increased in hepatopancreas (control → BM 33 and BM 66 → BM 100 ; p < 0.05), while other metals did not show any significant trend (Table 3). Except for Cd, all metals were significantly higher (p < 0.05) in the hepatopancreas of BM 100 fed crayfish compared to the control group. Such accumulation of heavy metal in hepatopancreas is capable of impairing metabolism in crayfish 37 . The concentration of Fe exhibits a rather 'bell curve' pattern, peaking at BM 33 and receding thereafter, only in the muscle (Table 3). Cd and Zn did not exhibit any pattern as such. The heavy metal contents in crayfish and BM are given in Tables 3 and 4, respectively. Globally, the total ash content in biofloc may range between 13-46% (reviewed by 22 ), also applicable in our case. The problem of high ash content in most biofloc, limiting its inclusion in diets (despite good protein content), has been briefly discussed in Sabry Neto et al. 38 . One previous study, which studied BM at a high enough inclusion level, attributed high ash and probable toxic effects of trace minerals to retarded growth in Litopenaeus vannamei fed > 60% BM in a diet 26 . Owing to high ash content in BM, mineral stress seems plausible in the present study as well (see Tables 1,4). By mineral stress, we imply even if 10% of the ash or minerals from BM are digested by crayfish, it is potentially much higher 'bioavailable minerals' in the body than their optimum physiological limits. Information on this aspect have been limited for shrimps [reviewed in 39,40] and none for crayfish 3,36 . In shrimps (Penaeus monodon, P. japonicus), retarded growth was observed when excessive mineral premixes were supplemented in a practical diet 39 , or more specifically, when trace minerals like Fe and Mn exceeded levels of 0.01% each in the diet 40 . The BM 100 had all these factors (ash, Fe, and Mn) in excess (Table 4). Heavy metal stress could also be plausible. Any significant absorption of Hg in the body (presented above) is capable of impairing crayfish metabolism 37 , provoking hyper-osmoregulation in crustaceans 41 , with repercussions on aggravated energy expenditure 42 . Our metadata derived models show TGC in crayfish deteriorates at dietary ash levels > 14% (also in BM 100 ), during which the retention of ash is merely < 10% of total dietary intake (see supplementary material and Fig S2, S3). Thus ≥ 90% of the ingested ash (exceeding physiological limits) are excreted through digestive and osmoregulatory (metabolic) pathways. It has its own energy cost, which could have been utilized for protein-sparing or growth 42 .

Recycling biofloc waste as a novel feedstuff for crayfish: Strengths and bottlenecks. Compar-
ing the nutritional standards for crayfish with observed performance in growth trials, few strengths and bottlenecks of BM were realized (Tables 1 and 4). In terms of advantages: (a) BM has a high crude protein content (43.9%); (b) crude fiber content in BM (4.9%) was in the optimum range for crayfish, and; (c) BM is a rich supplier of minerals. However, there are more bottlenecks than limited advantages. BM has excessive total ash detrimental to crayfish growth, with probable manifestations on hyper-osmoregulation and energy expenditure (discussed above). A mediocre crude lipid content (4.5%) is another bottleneck for supplying non-protein energy. These, in combination, render the non-protein energy: protein ratio (NPE: P = 3.7 cal non-protein energy per 1 mg protein) in BM insufficient for effective protein sparing (≈growth). At such low NPE:P, the proteins are catabolized for meeting energy demand (even after oxidizing carbohydrates and lipids), rather than building biomass 42 . It is further compounded by arginine deficiency in BM (~ 14-20% less than an optimum requirement)-probably the most critical essential amino acid for crayfish (discussed above).
A retrospective evaluation of BM 100 or BM (as a feedstuff for crayfish) applying our metadata derived 'growthretention models' (supplementary Fig S3, S4, S6) could explain few nutrient utilization scenarios behind low growth in BM 100 . The ash, protein, and lipid retentions from BM should be less than 5%, 10%, and 3% of dietary intakes, respectively (predicted). For control, BM 33, and BM 66 diets, these retentions were well above the identified thresholds qualifying for reasonably good growth in crayfish (refer to supplementary material). Comprehensively, the retarded growth problem with solely feeding on biofloc biomass could be a synergistic effect of-(a) arginine deficiency, (b) mineral and heavy metal stress, and, (c) low non-protein energy to protein ratio.

Methods
Calculation of crayfish nutritional standards, growth trajectory, and its nutritional dependencies. In the absence of centralized nutrition recommendations for freshwater crayfish species, unlike other commercially important crustaceans (e.g., penaeid shrimps, see NRC 4 ), available literature was meta-analyzed. Peer-reviewed and published articles (in English or at least with English abstract) were searched online (search engines: Web of Science, Scopus, and Google Scholar) using keywords like 'growth trials' , 'crayfish' , 'nutrition' , 'proximate composition' , 'body composition' , 'amino acids' , 'heavy metals' , 'optimum requirement' were used in different combinations (depending on target information). Altogether 27 articles were sourced and data extracted for meta-analyses. Detailed methodology on each meta-analysis (i.e., formulation of nutritional standards, calculation of growth trajectory and feed utilization parameters, quantification of nutritional dependencies on growth) are provided in the supplementary material.
Collection of biofloc biomass. Biofloc biomass was obtained from a well-established indoor, freshwater biofloc system, stocked with Nile tilapia Oreochromis niloticus at a stocking density of 35 kg m −3 . Commercial pellets (TILAPICO 3 mm, Coppens, The Netherlands) were used as standard feed for fish. Fish feed was given twice daily based on a feed amount equivalent to 2.5% of the fish body weight. Wheat flour (35.56% C; 2.38% N) Scientific Reports | (2020) 10:19607 | https://doi.org/10.1038/s41598-020-76692-0 www.nature.com/scientificreports/ served as a carbon source which was applied daily with feed (22.05% C; 7.07% N) in a ratio of 1:0.6 (feed: flour). Assuming a 30% retention of nutrients from feed to fish, the projected C: N ratio was ≈6:1. Such a low C: N ratio favored frequent harvest of young and N-rich wet biofloc biomass 6,10 to be converted to dry matter for the ensuing experiment. Biofloc biomass was drained daily through a pump and a vortex separation device so that the suspended solids level stayed between 25 and 50 ml L −1 in the system. After separation, biofloc was filtered through a nylon screen (mesh size 60 μm) to drain the excess water. The filtrate was then dried at 80 °C to obtain a material of solid consistency. After obtaining enough dried biofloc, the samples were grounded by a hammer mill to yield finer particles and hereinafter referred to as the biofloc meal (BM).
Preparation of experimental feed. Commercial pellets (TILAPICO 3 mm, Coppens, The Netherlands) were used as the basal diet due to its similar protein content with our test ingredient (BM). The commercial 'fish feed' was chosen due to a lack of established 'crayfish feeds' in the market. Even the available ones appeared to be random feed mixtures targeted for ornamental crayfish keeping. Inclusion of BM by replacing basal diet was done on a weight by weight basis. All feeds were isonitrogenous. The graded inclusion levels were 0% (basal diet = control diet), 33% (67% basal + 33% BM; diet BM 33 ), 66% (34% basal + 66% BM; diet BM 66 ) and 100% (only BM; diet BM 100 ). Feed pellets (pellet size 2 mm) were cold extruded, dried (12 h; 45 °C), vacuum sealed, and stored at 4 °C till further use. The diet samples were analyzed in an accredited third-party laboratory (AGRO-LA, spol. s.r.o., https ://www.agrol a.cz/zemed elske -a-potra vinar ske-sluzb y/) employing analytical methods (ISO verified and certified protocols in the Czech Republic) for proximate composition, essential amino acids (EAAs; except tryptophan due to analytical error), heavy metals, and essential mineral contents. Detailed composition of basal diet, treatment diets and the biofloc meal are summarized in Table 4.
Crayfish keeping. A total of 120 juvenile red swamp crayfish (Procambarus clarkii; conservation status: least concern) having a mean weight of 7.8 ± 0.7 mg at the onset of exogenous feeding (developmental stage 3), were used as experimental animals (10 individuals per tank; 4 group x triplicate). The experiment lasting for nine weeks was conducted in a series of indoor glass aquaria (54 × 36 × 30 cm, volume 46 L) with aeration and attached to a recirculating aquaculture system. Two baked clay bricks (28.5 × 13.5 × 6.5 cm), each with 39 cross holes (26 and 13 holes with a profile of 1 × 3 cm and 1 × 1 cm, respectively), were placed in each aquarium to provide shelters/refugia for the stocked crayfish 43 . After three weeks, a block of joined polypropylene tubes containing five tubes (length 10 cm, inner diameter 35 mm) was added to each aquarium as an additional shelter for on-growing animals. The bases were represented by three longitudinally joined tubes with a further two tubes positioned pyramidal in the second layer 44 . Altogether, 12 tanks were used and subjected to stable indoor climatic conditions with natural photoperiod (12L:12D).
Growth trial and feed utilization parameters. Crayfish were fed twice a day to apparent satiation (roughly corresponding 5-6% of the body weight) with the abovementioned diets for nine weeks. Uneaten feed, feces, and other wastes were siphoned out manually every morning. Dissolved oxygen (7.9 ± 0.3 mg L −1 ), pH (7.6 ± 0.2), and temperature (21.8 ± 0.3 °C) were measured daily using Oxi 3205 and pH 720 m (WTW GmbH, Weilheim, Germany), respectively. Every three weeks, the body weight was measured using an electronic balance (lowest sensitivity 1 mg) and the number of survivors counted. The feed rationing was revised accordingly. Body weight measurements were taken before feeding. After the trial, final body weight and total length were recorded, including the number of survivors. The animals were not fed before the day of the final measurement. The food conversion ratio (FCR, units), protein efficiency ratio (PER, units), and survivability (%) were determined for each diet following the formulas in Cortes-Jacinto et al. 45 . Live weight gain (LWG) was calculated applying the formula, LWG = final-initial weight (in mg)/ days reared. Coefficient of variance (CV) of body weight (standard deviation × 100 ∕ mean) was calculated as a measure of size heterogeneity. To eliminate statistical biasedness in the data due to hierarchical size distribution in crayfish groups, other measures of central dispersion like interquartile range (IR) and median were included besides the mean. The abovementioned parameters were calculated from the IR, median, and mean estimates of each treatment. All graphical models were generated using the ggplot2 package in R. Statistically significant differences (α level set at 0.05) in body weight, growth, and survivability of crayfish fed on different dietary treatments were tested. The grouped data were first subjected to a Shapiro-Wilk's normality test; then following the p value, either one-way ANOVA with post-hoc Tukey HSD (parametric test), or, Kruskal-Wallis post-hoc Dunn's test with Bonferroni correction (non-parametric test) was selected. The tests were performed using default commands in RStudio v1.2.5042.

Assessment of heavy metals risk from biofloc biomass.
At the end of the experiment, tail muscle and hepatopancreas samples from representative crayfish of each group were collected and frozen (−20 °C). Selected heavy metals (Hg, Cd, Zn; following high bioaccumulation affinity realized in Kouba et al. 46 ) and some additional minerals (Fe, Mn) were analyzed from these samples in the same accredited third-party laboratory. Body (muscle + hepatopancreas) heavy metal levels were compared with maximum permissible limits (Cd or Hg 0.5 mg kg −1 wet weight basis) given in the European Commission 47 for aquatic meat products (in the context of safety for consumption). In the context of agricultural use safety (as fertilizers), the heavy metal content of biofloc meal was determined and compared with Czech EPA limits (Cd 5 mg kg −1 , Hg 4 mg kg −1 dry matter basis) (Decree of Ministry of Environmental of the Czech Republic No. 437/2016 on the Code, 2016).

Ethics approval.
All procedures performed in studies involving animals (Oreochromis niloticus and Procambarus clarkii) were in accordance with the ethical standards approved by the institutional ethics committee (Jihočeská univerzita v Českých Budějovicích Fakulta rybářství a ochrany vod).