Utilization of several plant proteins by gibel carp (Carassius auratus gibelio)

Introduction

Fish meal is one of the most expensive ingredients in prepared fish diets. Because of the increasing price and strict environmental legislation, fish nutritionists have tried to use less expensive plant protein sources to partially or totally replace fish meal. In the context of research on substitutes for fish meal in the diets of fish, numerous studies have been made in the past decade (Tacon 1990).

A number of studies have evaluated various protein sources for fish feeds, but the results were not consistent. Unfortunately, attempts by feed manufacturers and nutritionists to replace the fish meal component of practical fish feeds with alternative protein sources have generally led to reduced feed efficiency and growth (Tacon and Jackson 1985; Xie and Cui 1998). There are several factors affecting the utilization of plant proteins, including poor palatability, low digestibility, poor utilization of protein and amino acids, anti-nutritional factors and other unknown factors (Cho et al. 1974; Higgs et al. 1979; Viola et al. 1981; Krogdahl 1989; Pongmaneerat et al. 1993; Masumoto et al. 1996; Xie and Cui 1998; Becker and Makkar 1999). Because of the various methods and materials used in different studies the results are often difficult to compare.

Soybean meal, rapeseed meal/canola meal and cottonseed meal are widely used plant proteins. The potential use of potato protein concentrate has recently been proposed as this protein source has a wide availability, high protein content and good amino acid composition (Moyano et al. 1992; Xie and Jokumsen 1997). Initial trials with salmonid fishes, however, have produced discouraging results, probably because of the toxic effects of solanine (Tacon and Jackson 1985). There has been only one report on the utilization of potato protein concentrate by the warmwater cyprinids, which usually have a higher tolerance to plant protein than salmonids (Parova and Par 1979).

Gibel carp (Carassius auratus gibelio L.) is a subspecies of goldfish. This subspecies is a natural triploid and reproduces gynogenetically. Over the past two decades, because of its higher growth rate, gibel carp has almost totally replaced the diploid goldfish (crucian carp) and become one of the most important aquaculture species in China. Some initial studies have been made on the nutritional requirements of this warmwater omnivore (He et al. 1990). The optimal protein requirement for juveniles has been suggested to be close to 40% (He et al. 1990). Currently, commercial feeds for the culture of gibel carp in China mostly use plant proteins (especially soybean cake) as the major protein source (M. Xue, pers. comm.), and an evaluation of various plant proteins will be useful for the improvement of diet formulation. Poor growth by fish fed plant proteins, when compared with fish fed fish meal, could be caused by lower intake, lower digestibility or poor protein, or energy retention. Identification of such causes will provide insight into development of techniques to improve utilization of plant proteins. The purpose of the present study was to investigate growth performance of gibel carp fed diets containing soybean cake, cottonseed cake, rapeseed cake, peanut cake or potato protein concentrate at high inclusion levels and to identify the causes for the relatively poor growth by gibel carp fed these plant proteins.

Materials and methods

The growth trial was carried out in a recirculation system consisting of 20 fibreglass tanks (diameter: 80 cm, water volume: 300 L) with a conical bottom. Each tank was equipped with a clinoptilolite filter (50 L volume). Clinoptilolite is a natural zeolite with a strong absorption capacity for ammonia. Dechlorinated tap water was introduced into each tank at a rate of around 350 ml min⁻¹. The water was recirculated through the filter at 650 ml min⁻¹. Aeration was provided intermittently for a half hour every 1 h. Dissolved oxygen was maintained above 5 mg L⁻¹, pH was around 6.8, ammonia-nitrogen below 1 mg L⁻¹ and water temperature was 25 ± 2°C. The photoperiod was 12 L:12 D, with the light period from 08:00 to 20:00 h.

Gibel carp were produced after induced spawning at the hatchery of the Institute of Hydrobiology, the Chinese Academy of Sciences. The fish were transferred for acclimatization into the tanks 4 weeks prior to the growth trial. During the acclimatization period, the fish were fed a practical diet (protein content 35%) to satiation twice a day. In the last week before the experiment, the fish were fed an equal mixture of the experimental diets.

The experimental diet (Table 1) was formulated to contain 35% crude protein, 4.5% lipid and 24% carbohydrate; 1% Cr₂O₃ was added as an indicator for digestibility determination. Five plant proteins including soybean cake (SBC), potato protein concentrate (PPC), peanut cake (PNC), rapeseed cake (RSC) and cotton seed cake (CSC) were used in the diets. Fish meal was used as the reference protein source. In each diet, 15% fish meal (27% of total protein) was included, and the remainder of the protein came from the tested plant proteins. Formulation and chemical composition of the diets are shown in Tables 1 and 2. The diet was made into 1-mm floating pellets by extrusion at 110°C, oven-dried at 60°C and stored at 4°C.

Table 1. Chemical composition of ingredients used in the experimental diets. Values are means of duplicate measurements

Table 2. Diet formulation and chemical composition of experimental diets (% of dry weight)

The fish were weighed individually at the beginning of the experiment after 1 day of starvation. Eighteen tanks were randomly assigned to each of the six diets in triplicate. Thirty fish weighing about 3 g each were stocked in each tank. Triplicate samples of five fish each were taken from the original batch for measurement of initial body composition; these fish were autoclaved at 120°C, homogenized and dried at 70°C.

During the trial, fish were fed to satiation twice a day (at 09:00 and 14:30 h). The daily feed supplied was recorded, and uneaten feed was collected 20 min after feeding, and dried. Faeces were collected twice a day from a faecal trap (settling column) connected to the conical bottom of the tank and dried at 70°C. Only faeces that remained intact were kept for chemical analysis. Recovery rate of uneaten feed was determined by placing six weighed feed samples in tanks without fish for 20 min and then collected, dried and weighed. The recovery rate was used to correct the weight of the uneaten feed.

The trial lasted for 8 weeks. At the end of the trial, all fish were weighed individually after 1 day of starvation. Five fish from each tank were taken for chemical analysis.

Contents of dry matter, crude protein, lipid, ash and energy were determined for the feeds and the fish according to AOAC (1984) methods. Crude protein and energy contents were also determined for faecal samples from each tank. Crude fibre was determined for feed. Dry matter was determined after the sample was oven-dried to constant weight at 105°C. Crude protein was analysed by the Kjeldahl method, crude fat by ether extraction and fibre by drying and ashing after the extraction with 0.5 M H₂SO₄ and 0.5 M NaOH. Ash content was determined after 12 h at 550°C in a muffle furnace. Gross energy was measured by bomb calorimetry. Amino acid contents of feeds and faeces from each tank were measured using a Waters TICO-PAG amino acid automatic analyser. The inert marker, Cr₂O₃, in the diet and faeces was determined as described by Furukawa and Tsukahara 1966). For each variable, duplicate samples were measured.

Specific growth rate (SGR, % day⁻¹) was calculated as:

100 × (Ln FBW – LnIBW)/t where FBW and IBW were the final and initial body weight (g), and t was the experimental period in days.

Feeding rate (FR, % body weight day⁻¹) was calculated as:

Total feed intake × 100/((Initial body weight + final body weight)*0.5).

Conversion efficiency (CE, %) was calculated as:

100 × (wet weight gain)/(total feed intake).

Apparent digestibility coefficient (ADC, %) was calculated as:

100–(Cr₂O₃ content in feed/Cr₂O₃ content in faeces) × (protein, energy or amino acid content in faeces/protein, energy or amino acid content in feed) × 100.

Protein retention efficiency (PRE, %) was:

recovered protein in fish body × 100/protein intake.

Energy retention efficiency (ERE, %) was:

recovered energy in fish body × 100/energy intake.

Possible differences between groups were examined by a multiple range test (Newman–Keuls) after ANOVA. Differences were regarded as significant when P < 0.05. Least-squares regression was used to analyse the quantitative relationship between SGR and intake of digestible protein, amino acids and energy.

Results

Growth performance

The initial body weights of the PNC and RSC groups were significantly smaller than in the PPC group, but the variations were small (2.87–3.04 g). The dietary composition significantly affected SGR, feeding rate and conversion efficiency (Table 3). SGR in fish fed the reference feed was superior compared to the other groups, while SGR in fish fed the PPC was significantly lower. There was no significant difference in SGR among the other groups. Feeding rates ranked in the order: RSC > CSC > FM > PNC > SBC > PPC. Conversion efficiency was highest in groups fed FM, SBC and PNC, followed by groups fed CSC and RSC, and was lowest in the group fed PPC.

Table 3. Growth performance of gibel carp fed different protein sources (mean ± SE)

Protein utilization

The highest protein intake was with groups fed RSC, CSC and RNC, and the lowest with groups fed SBC and PPC (Table 4). Protein digestibility of the FM diet was significantly higher than that of the SBC diet, while there was no significant difference between other diets. Protein retention was significantly higher in groups fed the FM and SBC diets than in the group fed the PPC diet, while there was no significant difference among other groups. There was no significant relationship between SGR and intake of digestible protein (g g⁻¹ day⁻¹) (r = 0.112, P = 0.103).

Table 4. Protein utilization of gibel carp fed different protein sources (mean ± SE)

Energy utilization

Energy intake was highest in groups fed RSC and CSC diets, followed by groups fed PNC and FM, and lowest in groups fed SBC and PPC diets (Table 5). Energy digestibility was highest for the FM, CSC and PNC diets, followed by the RSC and PPC diets, and lowest for the SBC diet. Energy digestibility for fishmeal, cotton seed cake and peanut cake was significantly higher, while that for soybean cake was lower (P < 0.05). Energy retention was highest in the group fed FM diet, and lowest in the group fed the PPC diet. There was no significant difference between other groups. There was no significant relationship between SGR and intake of digestible energy (kJ g⁻¹ day⁻¹) (r = −0.048, P = 0.639).

Table 5. Energy utilization of gibel carp fed different protein sources (mean ± SE)

Amino acid utilization

Table 6 shows the apparent digestibility of the amino acids in the diets. Amino acid digestibility was usually highest for the fish meal diet and lowest for the SBC diet.

Table 6. Apparent digestibility of amino acids for gibel carp fed different protein sources (%) (mean ± SE)

There was no significant relationship between SGR and intake of digestible lysine (r  =  −0.048, P  =  0.610), methionine (r  =  0.104, P  =  0.112) or total essential amino acids (r  =  0.343, P  =  0.006) (g g⁻¹ day⁻¹).

Body composition

The different diets significantly affected body lipid and ash content, but did not affect the contents of dry matter, crude protein or energy content of the different groups (Table 7). The fish fed RSC had the highest, while the fish fed PPC showed the lowest, lipid content (P < 0.05). The fish fed PPC had the highest, while those fed PNC had the lowest, ash content (P < 0.05).

Table 7. Body composition of gibel carp fed different protein sources (mean ± SE) (of wet weight)

Discussion

In most studies, decreased growth was observed in fish when a large proportion of the fish meal was replaced by soybean cake (Xie et al. 1998). Factors suggested to be the causes for decreased growth in fish fed soybean cake include low palatability (Kim 1974; Mohsen and Lovell 1990), low digestibility (Lei et al. 1996; Masumoto et al. 1996), deficiency in some essential amino acids, especially lysine (Viola et al. 1982) and methionine (Jackson et al. 1982), poor utilization of protein (Boonyaratpalin et al. 1998), and low energy (Viola et al. 1981, 1982). In the present study, total feed intake was lower in gibel carp fed the SBC diet than in the control, but protein and energy intakes were similar. The SBC group had lower digestibilities of dry matter, protein, energy and most amino acids than the control. Conversion efficiency and protein retention efficiency were similar between the SBC group and control, but energy retention efficiency was significantly lower. Thus the lower growth rate in the SBC group was mainly caused by lower digestibilities and energy retention efficiency. Several recent studies showed that pretreated soybean protein products, in which some of the toxins were removed, could replace most of the fishmeal in the diets for salmonids (Davies et al. 1989; Kaushik et al. 1995; Refstie et al. 1998). The lower digestibility and energy retention efficiency in SBC-fed gibel carp were probably mainly caused by the existence of anti-nutritional factors.

Decreased growth of fish fed the diet with the inclusion of rapeseed cake was reported to be due to low feed intake (Higgs et al. 1982), low digestibility of dry matter and/or protein (Hilton and Slinger 1986), unbalanced amino acids (Tacon and Jackson 1985) and/or high fibre contents (Hilton et al. 1983). The present study showed that the fish fed the RSC diet had a higher feeding rate than the control (FM diet). The digestibility of dry matter and energy was lower than in the control while the digestibility of protein was similar to that in the control. Except for methionine and glutamine, digestibilities of amino acids were lower in the RSC diet than in the FM diet. Lysine and methionine contents in the RSC diet were lower than the requirements. As the fibre content in the RSC diet was lowest in the present study, a high fibre content was not the cause for the lower growth rate. Fish fed the RSC diet also had lower conversion efficiency and energy retention than the control. The decreased growth of gibel carp fed RSC diets was largely caused by low digestibility of dry matter, energy and most of the amino acids, as well as low energy retention efficiency. Compared to soybean meal, Hardy and Sullivan (1983) reported no significant difference in the growth of the rainbow trout fed soybean meal and canola meal. Although the present study showed similar growth rates for the gibel carp fed SBC and RSC diets, the conversion efficiency was lower for the RSC diet, while the feeding rate, apparent digestibility of dry matter and energy were higher. The protein retention and energy retention were similar for the SBC diet and the RSC diet. Toxins, including phytic acid, tannins and other phenols, are reported to be the cause for low growth for fish fed rapeseed cake (Jones 1979; Tacon and Jackson 1985). Myrosinase hydrolyses glucosinolates, and one of the hydrolytic products is goitrin, a compound which can alter thyroid function in trout and salmon (Higgs et al. 1979).

Cottonseed cake inclusion was reported to decrease the growth of fish in several studies (El-Sayed 1990). Lysine and methionine were mostly reported to be the limiting amino acids (Tacon and Jackson 1985; El-Sayed 1990). Low lysine, methionine and cystine availability in cottonseed meal and their possible limitation on fish growth were highlighted by Jauncey and Ross (1982). Gossypol was reported to be the main toxin in cottonseed meal and could be a major reason causing decreased fish growth (Jauncey and Ross 1982; El-Sayed 1990). Decreased growth of fish fed cottonseed flour may also be due to the high level of cyclopropeniod fatty acids, reported to cause adverse effects in fish (Hendricks et al. 1980). The present study showed that the fish fed CSC diets had slower growth than the control. In the present study, the CSC diet had a low lysine content and a higher arginine content than that in the fish meal diet (Table 8). Apparent digestibilities of dry matter, protein and energy were similar to those in the fish meal diet, but digestibilities of lysine, leucine, threonine and isoleucine were lower. The lower digestibilities could be caused by gossypol (Jauncey and Ross 1982). Fish fed the CSC diet had similar protein retention to that in the control, but lower energy retention.

Table 8. Amino acid composition of experimental diets (% of dry matter)

Most of the work on potato protein has been applied to salmonids fish (Moyano et al. 1989; Xie and Jokumsen 1997). Low feeding rate and poor growth and feed utilization was reported in rainbow trout fed diets with potato protein concentrate (Xie and Jokumsen 1997). Moyano et al. (1989) suggested that the poor acceptance and poor growth in trout fed potato protein was due to the existence of factors affecting palatability and/or nutritional value (Moyano et al. 1987) and the high content of non-protein nitrogen (Chick and Slack 1949). The present study also showed that fish fed the PPC diet had the lowest feeding rate. Fish fed the PPC diet had similar digestibilities of most amino acids to those in the control, but the energy digestibility and protein and energy retention were much lower. Thus low palatability and poor postabsorption utilization of protein and energy were the major causes of reduced growth in PPC-fed fish. Solanine is reported to be the main anti-nutritional factor in potato protein (Moyano et al. 1989, 1992) and its effect on nutrient digestion and utilization should be studied quantitatively.

Wu and Jan (1977) reported that the growth of Tilapia aurea fed on an all groundnut protein diet was 58% worse than the fish meal control. They suggested that the poor performance of groundnut-fed fish was due to the poor amino acid profile. Jackson et al. (1982) regarded low methionine content as the most likely explanation of the poor growth of tilapia that were fed groundnut protein. In the present study, the methionine content was higher than that required for carp, but the lysine content was much lower (Table 8). PNC-fed fish had similar levels of feed intake, digestibility of protein, most amino acids and energy, and protein retention to those in the control, but energy retention was lower. Thus the lower growth rate in the PNC-fed fish was mainly caused by poor postabsorption retention of energy. Flavacin is a toxin that frequently develops in peanut cake due to fungal growth; whether this toxin has adverse effects on energy utilization is not clear.

In conclusion, the present study showed that inferior growth of fish fed fish meal was caused by different factors in gibel carp fed diets with different plant proteins. Except for the diet containing potato protein concentrate, gibel carp fed plant proteins had a similar protein and energy intake to those of fish fed fish meal, suggesting that gibel carp accepted these plant proteins well. Energy retention was lower in fish fed any of the plant proteins than in fish fed fish meal. Reduced protein and amino acid digestibility and lower energy retention efficiency were likely to be the most important causes for the reduced growth rate in gibel carp fed plant proteins. The lack of difference in growth rate by gibel carp fed various plant proteins, with the exception of the poorly accepted and utilized potato protein concentrate, suggests that the relatively expensive soybean cake can potentially be replaced by other cheaper plant proteins in practical diets for this species.

In the present study, only fingerlings were tested, and the plant proteins were incorporated at a relatively high level. Future experiments should be carried out on gibel carp of different sizes, and on the suitable inclusion level of various plant proteins that yields the best economic gains.

Acknowledgements

This project was supported by the National Natural Science Foundation of China (39670573 and 39625006) and Chengguang Project (985003078) of Wuhan City, and partly by the Ministry of Science and Technology of China (96–008–02–03) and the Chinese Academy of Sciences (KZ 951-A1–102–01).