2.1 Experimental design

Juvenile P. clarkii (4.82 ± 0.15 g, 60.03 ± 0.52 mm, mean ± SE) were obtained from ponds at the Selection and Reproduction Center of Crayfish, Qianjiang, Hubei Province, China. A 50-day feeding experiment was conducted in eight experimental concrete ponds (90 juveniles per pond of 9 m²), which followed the National Institutes of Health guide for the care and use of Laboratory Animals (NIH Publications No. 8023, revised 1978). There were four replicate ponds for the two treatments (26% and 30% protein levels). The experimental crayfish were acclimated to the culture conditions for one week before the experiment. At the beginning of the experiment, healthy juveniles were collected and randomly allocated to 8 concrete ponds. Throughout the whole experimental period, water flow rates in ponds were approximately 7 L/min with constant aeration. Water depth was maintained at approximately 27 cm, and H. verticillata was planted in 35 polyethylene flowerpots (0.44 m diameter) which served as both shelters and foods with coverage of approximately 60% in each pond. The water temperature, pH and dissolved oxygen (DO) were measured by a YSI probe (Yellow Springs Instruments), which were 27.27 ± 1.06°C, 9.3 ± 0.05 and 4.33 ± 0.70 mg/L throughout the study. The concentrations of ammonia nitrogen, nitrite, chemical oxygen demand, total nitrogen, total phosphorus and chlorophyll-a were determined using standard methods (APHA (American Public Health Association), 1992). Specifically, ammonia nitrogen was 0.1700 ± 0.015 mg/L, with 0.0412 ± 0.002 mg/L for nitrite, 1.2009 ± 0.020 mg/L for total nitrogen, 0.0345 ± 0.017 mg/L for total phosphorus, 7.1347 ± 0.100 mg/L for chemical oxygen demand and 17.6489 ± 0.440 μg/L for chlorophyll-a respectively.

2.2 Feeding management

Throughout the experiment, crayfish were fed twice daily (8:00 and 18:00). For each feeding practice, the experimental diets were put into a plastic pallet (30 × 15 cm), a diagram of which was described in our previous study (Jin et al., 2019). Crayfish were fed with an excess weighted diet. Two diets (containing 26% and 30% protein levels) were chosen based on the existing knowledge of dietary requirements of P. clarkii. The experimental diets followed common commercial diet formulation from Charoen Pokphand Group (WHS001-2016). They were designed specifically for crayfish, differing in protein levels (Table 1). Ingredients were ground and sieved with a 60-mesh screen, and then completely mixed and made into 2-mm pellets. The diets were dried at 45°C by an air flux oven (Yiheng DHG-9030) and stored at −20°C. After their feeding activities stopped within one hour, the remaining diet was removed, dried and reweighted (Van Ham et al., 2003). The amount of diet consumed by crayfish was calculated. The experiment lasted 50 days.

2.3 Sample collection

All crayfish were starved for 24 h and then collected for growth performance parameters measurement at the end of the experiment. Twelve male crayfish and twelve female crayfish of each pond (48 crayfish for each treatment totally) were randomly sampled and stored at −20°C for muscle composition analysis after chill-killed using the ice-water bath. Furthermore, samples of eight individuals from each treatment were also chill-killed and maintained for stable isotope analysis.

2.4 Growth performance

Parameters for crayfish growth performance such as survival rate, final length (L), final weight (W), gonad weight, liver weight and muscle weight were recorded and calculated as follows:

where N₀ and N_t are the initial and final number of P. clarkii per treatment; W₀ and W_t is the initial and final weight of P. clarkii; L₀ and L_f is the initial and final length of P. clarkii; W_g is the gonad weight of P. clarkii, and W_l is the liver weight of P. clarkii; and T is the number of experimental days.

2.5 Muscle composition analysis

Samples of crayfish abdominal muscle and artificial diets were analysed for protein, lipid, moisture and ash contents. Protein content was determined using the Kjeldahl method (N × 6.25) (William, 1980) with a 4800 Kjeltec Auto Analyzer (FOSS Tecator, Haganas). Lipid content was determined by chloroform–methanol extraction (Folch et al., 1957). Moisture content was determined by placing a 1-g sample into a convection oven (105°C) for 2 h and drying it to constant weight (William, 1980). Ash content was determined by placing a 1-g sample combusting at 550°C in a muffle furnace for approximately 10 h (William, 1980).

2.6 Stable isotope analysis

Samples of crayfish, experimental diets and H. verticillata were separately collected and oven-dried at 60°C for 48 h to a constant weight. They were then very finely ground into a homogeneous powder (<200 μm). All samples were then placed in a 3.5 × 5 mm tip cup for δ¹⁵N and δ¹³C isotope analysis (Jin et al., 2019). Specifically, 3-mg samples were combusted and the gases were analysed by gas chromatography and continuous flow-mass spectrometry (MAT-253, Thermo Fisher Scientific). The reference standards of δ¹³C and δ¹⁵N were Vienna Pee Dee Belemnite and atmospheric N₂ respectively. They were measured to a precision of ±1%. The isotope values for δ¹⁵N (‰) and δ¹³C (‰) were calculated using the following equation:

2.7 Statistical analyses

The variations in the survival rates between the two treatments were analysed using pairwise permutation tests. Student's t tests were used to analyse the variations in other growth parameters (W, L, GSI, HSI, SGR_W, SGR_L and muscle weight), muscle composition, and crayfish δ¹³C and δ¹⁵N values between the two treatments. The Kruskal–Wallis test was used to analyse the differences in δ¹³C and δ¹⁵N values between the two experimental diets and H. verticillata. Growth performance parameters were analysed using principal component analysis (PCA). For the stable isotope data, we calculated the contributions of the experimental diets and the H. verticillata to crayfish growth using the ‘SIAR’ package in R. All analyses were performed using R (version 3.3.2, R Core Team, 2017), and the significance level was set to 0.05 (p < 0.05).

3.1 Growth performance

The growth parameter results of the males and females are provided in Table 2 and Data S1. The survival rates were 84.45% (26% dietary protein level) and 70.56% (30% dietary protein level). There were no significant differences between the two treatments (Student's t test, t = 2.06, p = 0.09).

The male P. clarkii fed the 26% dietary protein level had significantly higher muscle weight than those fed the 30% dietary protein level (Student's t test, t = 2.30, p = 0.03). The male P. clarkii had no significant differences in W, L, GSI, HSI, SGR_W and SGR_L between the two treatments (Student's t test, p > 0.05). Similarly, the female P. clarkii had no significant differences in all parameters (W, L, GSI, HSI, SGR_W, SGR_L and muscle weight) between the two treatments (Student's t test, p > 0.05).

PCA revealed no significant differences in all the growth parameters of both sexes between the two treatments (Figure 1). Specifically, PC1 included W, L, SGR_W, SGR_L and muscle weight, and explained 56.19% of the variance among samples. PC2 mainly separated females and males into two groups by GSI and HSI, explaining 30.26% of the variance. The two components accounted for 86.44% of the total variance. Taking both sexes into consideration, Figure 1 illustrated that crayfish fed a 26% dietary protein level had no significant differences in all growth parameters when compared with those a 30% dietary protein level. This result indicated reduced dietary protein levels from 30% to 26% did not impair the growth performance of juvenile P. clarkii.

3.2 Muscle composition

The ash, lipid, moisture and protein contents in P. clarkii muscles from the two different treatments are provided in Figure 2 and Data S2. P. clarkii fed the 26% dietary protein level had significantly higher protein and ash contents than the crayfish fed the 30% dietary protein level (Student's t test, p < 0.001). Conversely, the moisture content within the P. clarkii had the opposite trend, and it was significantly lower in P. clarkii fed the 26% dietary protein level (Student's t test, p < 0.001). The P. clarkii fed the 30% dietary protein level had higher lipid content than those fed the 26% dietary protein level, but the difference was not statistically significant between the two treatments (Student's t test, p = 0.17).

3.3 Stable isotope analysis

The carbon and nitrogen stable isotope ratios (δ¹³C and δ¹⁵N, Data S3) in P. clarkii and the food resources are provided in Figures 3 and 4 respectively. The δ¹³C values in P. clarkii ranged from −22.87 ‰ to −19.12 ‰, with no significant differences observed between the two treatments (Student's t test, p > 0.05). The δ¹³C values in the food resources were −21.18 ‰ to −20.5 ‰ for experimental Diet 1 (26% protein level), −21.24 ‰ to −19.2 3‰ for experimental Diet 2 (30% protein level) and −26.58‰ to −23.07‰ for H. verticillata. The δ¹³C values of the H. verticillata were significantly lower than those of the experimental diets (Kruskal–Wallis test, p < 0.05). The δ¹⁵N values of P. clarkii ranged from 4.34 ‰ to 5.66 ‰, with no significant differences observed between crayfish in the two treatments (Student's t test, p > 0.05). H. verticillata had significantly lower δ¹⁵N values (1.92 ‰–3.46 ‰) than the two experimental diets (Kruskal–Wallis test, p < 0.05), which were 4.25 ‰–4.3 ‰ for diet 1 and 4.91 ‰–5.05 ‰ for diet 2 respectively.

The isotopic Bayesian mixing model estimated the relative contributions of the food resources (diet 1, diet 2 and H. verticillata) to tissue growth in P. clarkii. Results identified the artificial diet was the main food consumed by P. clarkii in both treatments (mean contribution: 58.07% for diet 1 and 62.66% for diet 2, Table 3). However, when the dietary protein levels decreased from 30% to 26%, the mean contribution of H. verticillata increased from 37.34% to 41.93%. This result indicates the diet of P. clarkii shifted from the artificial diet to the easily available H. verticillata when the dietary protein levels in the artificial feed decreased.

4.1 Effects of dietary protein levels on growth performance of P. clarkii

Reducing the dietary protein level from 30% to 26% did not significantly affect growth performance in crayfish. In this situation, feeding P. clarkii with a dietary protein level of 26% ensured successful crayfish production with reduced production costs. Previous numerous laboratory experiments could not consider the contributions of natural foods and still confirmed the growth of P. clarkii did not benefit from high dietary protein levels. For instance, juvenile P. clarkii exhibited the highest growth rate when fed a diet with a 27% protein level (Wu et al., 2007). Other researches have suggested that optimal dietary protein levels are 26% – 30% for juvenile P. clarkii (Jover et al., 1999; Ling et al., 2012; Wu et al., 2007; Xu, Liu, et al., 2013; Zhang et al., 2012). Similar results have also been found in other species such as Macrobrachium americanum (Méndez-Martínez et al., 2017), Ctenopharyngodon idella (Xu et al., 2016), C. quadricarinatus (Cortés-Jacinto et al., 2003), M. carcinus (Benítez-Mandujano & Ponce-Palafox, 2014) and L. vannamei (Shahkar et al., 2014). All of these studies showed that an excessive input of dietary protein had negative effects on the growth of the cultured organisms.

For intensive aquaculture operations, artificial diets can exceed 50% of the production costs (Keckeis & Schiemer, 1992; Wong et al., 2016), and the diet cost highly depends on the proportion of protein. To provide crayfish at desirable market sizes in the shortest timeframe, farmers feed crayfish diets high in protein. However, high protein inputs in culture systems cause water pollution, low dissolved oxygen levels, and decreased efficiency in food absorption and impact immune systems, resulting in huge economic losses (Craig et al., 2017; Henry & Fountoulaki, 2014; Martinez-Cordova et al., 2003; Velazco-Vargas et al., 2014). Considering the similar production potential by the two treatments in the current study, we suggest reducing the dietary protein levels from 30% to 26% to maintain aquaculture production at a minimum economic loss. This change not only brings numerous benefits to farmers but is also the key to improving the economic and environmental sustainability of P. clarkii culture.

4.2 Effects of dietary protein levels on muscle composition of P. clarkii

The results of muscle composition analysis showed the P. clarkii fed the 26% protein diet had significantly higher crude protein and ash contents than those fed the 30% protein diet. However, dietary protein levels did not significantly influence the lipid content within the crayfish muscles. This result suggests reducing the dietary protein level to 26% would not negatively impact crayfish muscle composition. Furthermore, crude protein content in the crayfish muscles tended to decrease with the increased dietary protein level. This result supports previous research on the crayfish Astacus leptodactylus (Ghiasvand et al., 2012) and crab Portunus trituberculatus (Jin et al., 2013). However, not all previous researches support the same. Some studies found that muscle crude protein content tended to increase with an increase in dietary protein levels in P. clarkii (Li, 2012; Yu, 2011), C. quadricarinatus (Pavasovic et al., 2007) and M. americanum (Méndez-Martínez et al., 2017). Other studies demonstrated that dietary protein levels had no significant effect on muscle composition in P. clarkii (Ling et al., 2012), C. quadricarinatus (Thompson et al., 2004), M. carcinus (Benítez-Mandujano & Ponce-Palafox, 2014), L. vannamei (Hu et al., 2008) and M. nipponense (Zhang et al., 2017).

We identified the crude lipid content was not significantly affected by dietary protein levels, which was in agreement with previous research on P. trituberculatus (Huo et al., 2014) and C. quadricarinatus (Thompson et al., 2004). However, our results disagreed with previous studies in P. clarkii (Li, 2012; Su et al., 2009; Xu, Liu, et al., 2013), A. leptodactylus (Ghiasvand et al., 2012), P. trituberculatus (Jin et al., 2013), M. carcinus (Benítez-Mandujano & Ponce-Palafox, 2014), M. nipponense (Zhang et al., 2017) and M. americanum (Méndez-Martínez et al., 2017). In our study, the ash and moisture contents had an opposite tendency with an increase in the dietary protein level. In contrast, most studies found no significant differences in the ash and moisture contents with increasing dietary protein levels (Catacutan, 2002; Ghiasvand et al., 2012; Hu et al., 2008; Huo et al., 2014; Jin et al., 2013; Méndez-Martínez et al., 2017; Wu et al., 2007; Zhang et al., 2017). Many factors could be responsible for the different muscle compositions within cultured organisms. For instance, the protein, lipid and ash contents are size-dependent and can also be affected by the life stage and energy intake of the organism (Shearer, 1994). The different diet formulations also had significant effects on muscle composition. For instance, Shearer (1994) proved that diets containing energy from carbohydrates produced higher body protein levels than diets containing the same level of energy produced from lipids. In addition, other natural foods such as detritus, phytoplankton, benthos and zooplankton would also affect body composition in practical culture systems. For instance, Lei et al. (2013) demonstrated that fish feeding on phytoplankton and zooplankton had higher levels of polyunsaturated fatty acids. Therefore, our results indicate that feeding crayfish with high dietary protein levels is unnecessary because crayfish derive a substantial part of their dietary requirements from natural foods in the culture system. We also conclude that a high dietary protein input during farming is not the best feeding strategy for P. clarkii.

4.3 Natural foods contributions and implications for sustainable aquaculture

In this study, the stable isotope analysis showed that H. verticillata was an important component of P. clarkii diets, and contributed 37.34% and 41.93% when crayfish were fed artificial dietary protein levels of 30% and 26% respectively. The results confirm our hypothesis that a reduction in the dietary protein levels does not negatively influence the growth and muscle composition in P. clarkii because of the natural supplementary nutrition available from H. verticillata. In addition to H. verticillata, other natural foods also contribute to the growth of P. clarkii, such as benthic detritus, sediment, planktonic, zooplankton and invertebrates (Alcorlo et al., 2004; Grey & Jackson, 2012; Gutierrez-Yurrita et al., 1998; Huner, 1981; Kreider & Watts, 1998). Similar results have been identified in other crayfish. For instance, in semi-intensive pond cultures of C. quadricarinatus, the contribution of natural plants to crayfish growth can be as high as 44% (Joyce & Pirozzi, 2016). For Paranephrops zealandicus, terrestrial detritus constitutes up to 58.3% of its stomach contents (Hollows et al., 2002). These results also highlight the benefits of natural foods that contribute to crayfish growth, potentially reducing production costs.

4.4 Implications of reduced dietary protein levels for sustainable aquaculture

At the management level, reducing the dependence on high dietary protein inputs in farming systems and maximizing the utilization of natural foods as alternative food sources is of high significance to further reduce production costs (especially feed costs) to aquaculture profitability (Tacon, 1997). The nutritional and economic importance of natural foods has been well recognized in the culture of many crustacean species, increasing both pond productivity and yield. For example, previous studies on C. destructor and L. stylirostris have demonstrated the growth-enhancing effects of natural foods in ponds, which contribute 28%–79% and 37%–40% of their growth, reducing artificial diet costs (Cardona et al., 2015; Duffy et al., 2011). Similar results have also been reported in L. vannamei (Gamboa-Delgado et al., 2011; Porchas-Cornejo et al., 2012; Roy et al., 2012; Xu, Liu, et al., 2013), Farfantepenaeus brasiliensis (Emerenciano et al., 2012), C. quadricarinatus (Viau et al., 2012), M. rosenbergii (Correia et al., 2003), L. stylirostris (Cardona et al., 2015) and Marsupenaeus japonicus (Arapi et al., 2012). All these studies support that maximizing the use of natural foods in the overall nutritional budget of pond-cultured crayfish not only improve crustacean species growth but also greatly reduce production costs. It is important for farmers to focus on managing the naturally available foods in ponds to maximize production profits. As the artificial food involves large production costs, further studies should be encouraged to assist in the utilization of natural foods and explore sustainable feeding strategies.