Introduction
Early weaning of young rabbits ( < 26 days) has been proposed as a means of reducing does energy deficit by decreasing milk energy output and providing adequate nutrition of young rabbits as soon as they begin to eat solid feed (Pascual, Reference Pascual2001; Gidenne and Fortun-Lamothe, Reference Gidenne and Fortun-Lamothe2002; Xiccato et al., Reference Xiccato, Trocino, Sartori and Queaque2004). To use this method, however, the nutritional demands of young rabbits must be fully known. To date, several authors have dealt with nutrition of early weaned rabbits and some interesting results regarding enzyme supplementation and heat processing, a suitable source of protein and energy or fibre and energy level in the starter diet have been reported (Gutiérrez et al., Reference Gutiérrez, Espinosa, García, Carabaño and De Blas2002a and Reference Gutiérrez, Espinosa, García, Carabaño and De Blasb and Reference Gutiérrez, Espinosa, García, Carabaño and De Blas2003; Xiccato et al., Reference Xiccato, Trocino, Sartori and Queaque2003; Nicodemus et al., Reference Nicodemus, Pérez-Alba, Carabaño, De Blas, Badiola, Pérez de Rozas and García2004). However, the main problem is still the digestive health of early weaned rabbits. Faced with a specific pathology, early weaning tended to increase the health risk (Gidenne and Fortun-Lamothe, Reference Gidenne and Fortun-Lamothe2004; Volek et al., Reference Volek, Marounek and Skřivanová2005). The most common method of prevention of digestive disorders in rabbits is the use of a feed supplemented with an antibiotic. As antibiotic growth promoters have been banned in the European Union, there is a need for effective alternatives which can be used in rabbits.
It is possible to increase the resistance of early weaned rabbits to digestive troubles through dietary supplementation with mannan-oligosaccharide (MOS) or inulin-type fructans (Spring et al., Reference Spring, Wenk, Dawson and Newman2000). MOS is derived from the outer cell wall of the yeast Saccharomyces cerevisiae. It was shown that MOS had the ability to bind pathogenic bacteria expressing type-1 fimbriae such as Salmonella spp. or Escherichia coli (Spring et al., Reference Spring, Wenk, Dawson and Newman2000). In traditionally weaned rabbits (between 28 and 35 days of age), the addition of MOS to a diet stimulated villi development and caecal volatile fatty acid concentration and reduced caecal pH and mortality (Fonseca et al., Reference Fonseca, Falcão, Kocher and Spring2004; Pinheiro et al., Reference Pinheiro, Alves, Mourão, Guedes, Pinto, Spring and Kocher2004). Also fructo-oligosaccharides have been widely studied, mainly due to their beneficial effect on health of animals (Flickinger et al., Reference Flickinger, Van Loo and Fahey2003). Inulin is a polysaccharide composed of β-2,1-linked fructosyl moieties, mostly with a terminal glucose. Most of the inulin commercially available today is extracted from chicory roots. These types of inulin have a degree of polymerisation of 2 to 60, with an average ranging from 9 to 25 (Meyer et al., Reference Meyer, van Nuenen, Venema, Van der Kamp, NG Asp, Miller-Jones and Schaafsma2004). In rabbits, the addition of inulin to diet had a beneficial effect on caecal fermentative activity (Morisse et al., Reference Morisse, Maurice, Boilletot and Cotte1993; Maertens et al., Reference Maertens, Aerts and De Boever2004; Volek et al., Reference Volek, Marounek and Skřivanová2005), which could improve the protection of weaned rabbits against common pathogens. However, effect of dietary supplementation with MOS or inulin on digestive health and growth performance of early weaned rabbits is still insufficiently documented. Therefore, the aim of this study was to evaluate the effect of dietary supplementation with either mannan-oligosaccharide (MOS, Bio-Mos®, Alltech Inc.) or inulin (Frutafit® IQ, Sensus, Netherlands) on the health status, microbial activity (fermentative activity, fibrolytic activity, fibre digestion), digestibility of nutrients and growth of early weaned rabbits.
Material and methods
Experimental diets
Three experimental diets were formulated: diet C (control, without additives), diet M (supplemented with 0.3% mannan-oligosaccharide), and diet I (supplemented with 4% inulin). Common feed components were used (Table 1). The diets were similar in the level of crude protein (CP), fat and fibre fraction but differed in the level of starch and fructans (Table 2). Both diet C and diet M contained more starch and less fructans than diet I (130 and 7, 134 and 7, 120 and 47 g/kg, respectively). Mannan-oligosaccharide (Bio-Mos, Alltech, Inc., Nicholasville, KY) was added to the vitamin pre-mix in substitution of the pre-mix support (wheat flour), and then included in the diet M. The diet I was supplemented with 4% chicory inulin, at the expense of wheat bran, by means of Frutafit® IQ (SENSUS, 4704 RG Roosendaal, The Netherlands). Frutafit® IQ is the instantised version of powdered inulin. According to Meyer et al. (Reference Meyer, van Nuenen, Venema, Van der Kamp, NG Asp, Miller-Jones and Schaafsma2004), about 85% of inulin molecules in the Frutafit® IQ consist of six or more glycosyl residues. In accordance with recent results (Gutiérrez et al., Reference Gutiérrez, Espinosa, García, Carabaño and De Blas2003; Xiccato et al., Reference Xiccato, Trocino, Sartori and Queaque2003; Volek et al., Reference Volek, Marounek and Skřivanová2005) regarding some aspects of the formulation of starter diet for early weaned rabbits, the diets contained sunflower meal instead of soya-bean meal, a higher level of fat, and a lower level of starch and dietary pectin. The anticoccidial Robenidine (66 mg/kg diet) was used. No other antimicrobials were used during trial.
Table 1 Ingredients (%) of control (C), mannan-oligosaccharide (M) and inulin (I) diet
† Per kg supplement: vitamin A-1 200 000 IU; vitamin D3-200 000 IU; vitamin E-5 g; Vitamin K3-0.2 g; vitamin B1-0.3 g; vitamin B2-0.7 g; vitamin B6-0.4 g; niacinamide-5 g; Ca-pantothenate-2 g; folic acid-0.17 g; biotin-20 mg; vitamin B12-2 mg; choline-60 g; lysine-25 g; dl- methionine-100 g.
‡ Supplemented with 0.3% mannan-oligosaccharide (BIO-MOS, Alltech, Inc., Nicholasville, KY).
Table 2 Chemical composition (g/kg) of control (C), mannan-oligosaccharide (M) and inulin (I) diet
† According to the sequential method of Van Soest et al. (Reference Van Soest, Robertson and Lewis1991); NDF = neutral-detergent fibre, ADF = acid-detergent fibre, ADL = acid-detergent lignin.
‡ Calculated according to Gidenne (Reference Gidenne2003).
§ Calculated according to Knudsen (Reference Knudsen1997).
‖ Non-nitrogenous cellular content = organic mater – NDF – CP.
¶ Calculated according to Maertens et al. (Reference Maertens, Perez, Villamide, Cervera, Gidenne and Xiccato2002).
Health status and growth performance trial
A total of 330 Hyplus® rabbits (616 ± 79 g), weaned at 25 days of age, were randomly allocated into three groups (group C, M and I). All rabbits were individually ear tagged. The experiment was carried out under practical conditions on a commercial farm, in spring 2005. At weaning, rabbits were moved from the maternal sector to the fattening sector and put in all-wire cages (30 × 30 × 33 cm), two per cage. A minimum environmental temperature was maintained at 16°C during the experiment.
The control diet (diet C, without additives) was fed to rabbits of the C group from weaning to slaughter at 74 days of age. Diet M (MOS) and diet I (inulin) were fed to rabbits of the respective group (group M and group I) from weaning to 46 days of age, then rabbits received control diet till slaughter.
The diets and water were offered ad libitum to all rabbits during the entire experimental period. As the trial was performed on a commercial farm without the possibility of measuring feed intake individually, consumption of feed was measured weekly per group, and therefore statistical evaluation of feed intake was not possible. Feed intake (g per rabbit per day) was calculated as follows: consumption of feed was divided by both numbers of rabbits within period and days within period (Volek et al., Reference Volek, Marounek and Skřivanová2006). Rabbits which died during period were excluded from calculation. Animals were individually weighed every week.
Mortality was recorded every day; morbidity was recorded weekly. Health status was evaluated according to European Group on Rabbit Nutrition (Fernández-Carmona et al., Reference Fernández-Carmona, Blas, Pascual, Maertens, Gidenne, Xiccato and García2005). It means that morbidity corresponds to sick rabbits (but still alive within a period) showing digestive troubles or severe loss of weight during a week. An animal was considered morbid only once (within period), even if diarrhoea lasted several days. The ‘health risk index’ was the sum of morbid and dead rabbits, knowing that each animal was considered only once (classed either dead or morbid). Dead rabbits were subjected to autopsy in the State Veterinary Institute in Prague and examined using ISO methods (http://www.iso.org).
Microbial activity
A total of 18 Hyplus® rabbits (636 ± 45 g), 25 days old at the beginning of the trial, were used for caecal digesta evaluation. Rabbits were assigned at random to the three experimental diets (C, M, and I diet) and put in all-wire cages (0.16 m2), one per cage, in a climate-controlled room. Environmental conditions were as follows: temperature 16 ± 1°C, relative humidity ca. 65%. Both feed and water were offered ad libitum. Rabbits were slaughtered at the age of 42 days. After laparatomy, the caeca were excised and weighed. The caeca were emptied by squeezing, the pH of the caecal digesta was measured immediately, and caecal contents were weighed. For analyses of caecal volatile fatty acid and ammonia concentrations, portion of caecal digesta was diluted 1:2 with distilled water, and 50 mg of mercury II chloride were added to inactivate microbial growth. Other portions of caecal digesta was frozen and kept under CO2 at − 80°C until analysed (bacterial fibrolytic activity) or used for dry matter determination. All animals were healthy, i.e. no diarrhoea or loss of live weight were observed.
Apparent digestibility trials
A total of 30 Hyplus® rabbits (629 ± 40 g), 25 days old at the beginning of the experiment, were used in order to determine the coefficient of the total tract apparent digestibility of the experimental diets. Rabbits were assigned at random to three experimental diets (C, M, and I diet). Animals were individually housed in digestibility cages (34.5 × 39.5 × 40.0 cm). Feed and water were offered ad libitum during the whole experimental period. The environmental conditions were as follows: temperature 16 to 18°C, relative humidity ca. 65%. Following 11 days of an adaptation period, the feed intake and total faecal output (caecotrophy was not prevented) were recorded from 36 to 40 days of age for each rabbit according to the European reference method (Perez et al., Reference Perez, Lebas, Gidenne, Maertens, Xiccato, Parigi-Bini, Dalle Zotte, Cossu, Carazzolo, Villamide, Carabaño, Fraga, Ramos, Cervera, Blas, Fernandez, Falcao e Cunha and Bengala Freire1995). All animals were healthy, i.e. no diarrhoea, low feed intake or loss of live weight were registered.
Analytical methods
CP, fat and starch content were determined by the Association of Official Analytical Chemists (AOAC) procedure (1980): contents of CP (6.25 × N by the AOAC procedure 2.055) and fat (AOAC procedure 7.052) in the feed and faeces were determined employing instruments Kjeltec Auto 1030 Analyser and Soxtec 1043 from Tecator Comp. (Sweden), respectively. Starch was determined by the Ewers method (AOAC procedure 14.032). Dry matter was determined by drying of samples of the feed, digesta and faeces at 105°C to constant weight. Neutral-detergent fibre (NDF), acid-detergent fibre (ADF) and acid-detergent lignin (ADL) were determined according to the procedure of Van Soest et al. (Reference Van Soest, Robertson and Lewis1991), using Fibertec 2010 (Tecator Comp., Sweden). Water-insoluble pectins, fructans, as well as digestible energy were calculated from tables (Knudsen, Reference Knudsen1997; Maertens et al., Reference Maertens, Perez, Villamide, Cervera, Gidenne and Xiccato2002; Gidenne, Reference Gidenne2003). The non-nitrogenous cellular content (NNCC), which includes pectins, starch, oligosaccharides and free sugars, was estimated by difference according to the equation: NNCC = OM - CP - NDF (Gidenne and Fortun-Lamothe, Reference Gidenne and Fortun-Lamothe2004), where OM is dry matter minus ash. Caecal contents were diluted with distilled water 1:2 and total volatile fatty acids (VFA) were determined by titration after steam distillation. The VFA molar percentages were estimated by gas chromatography at 140°C, employing a column of the Chromosorb WAW with 15% SP 1220 and 1% H3PO4 (Supelco). Ammonia was determined colorimetrically with Nessler reagent after prior separation from interfering compounds by microdiffusion in Conway units (Conway, Reference Conway1957). Activities of cellulase (EC 3.2.1.4), xylanase (EC 3.2.1.32), pectinase (EC 3.2.1.15) and inulinase (EC 3.2.1.7) were assayed according to Kopečný and Bartoš (Reference Kopečný and Bartoš1990) using carboxymethylcellulose (sodium salt, low viscosity), wood xylan, citrus pectin and inulin as substrates at concentration of 4 mg/ml. Contents of the caecum were diluted with a 50 mmol/l phosphate buffer (pH 7.0). For the incubation, 1 ml of substrate and 1 ml of digesta were diluted and mixed. The incubation was performed at 39°C for 1 h, and 2 h in the case of the determination of cellulolytic activity. It was terminated by adding 1 ml of 0.3 mol/l Ba(OH)2 and 1 ml of 0.3 mol/l ZnSO4. The precipitate was removed by centrifugation (10 000 g, 15 min) and reducing sugars were quantified spectrophotometrically at 410 nm by a p-hydroxybenzoic acid hydrazide method (Lever, Reference Lever1977). The quantity of released sugars was expressed as micromole of reducing sugars per gram of dry matter of caecal content and per hour.
Statistical analyses
For a statistical analysis of growth rate, only data from healthy rabbits were used (initial number of rabbits at weaning minus morbidity and mortality) (Table 3), in order to evaluate the effect of diets on growth rate corrected for the influence of sick rabbits (Gidenne et al., Reference Gidenne, Pinheiro and Falcão e Cunha2000; Volek et al., Reference Volek, Marounek and Skřivanová2006). Data on effect of the diet on growth performance, caecal parameters and digestibility coefficient were analysed using the GLM procedure of Statistical Analysis Systems Institute (2001). Sheffe's test was used for mean comparison where appropriate. Data on mortality and morbidity were analysed using the χ2 test. The statistical significance was considered as P < 0.05.
Table 3 Growth performance of early weaned rabbits fed control (C), mannan-oligosaccharide (M) and inulin (I) diet
† RMSE = root mean square error.
‡ n = initial number of rabbits at weaning (110 per group) minus mortality and morbidity.
§ Average data of groups not analysed statistically.
Results
Diets, growth performance and health status
In accordance with the objectives of the experiment, the three experimental diets were similar in their chemical composition (Table 2). They differed in the level of starch and fructans because of the addition of 4% chicory inulin into the diet I at the expense of wheat bran content (Table 1).
During the starter period (25 to 46 days of age), the weight gain tended to be higher in control rabbits than in other groups (35.7, 32.3 and 33.3 g/day in the group C, M and I, respectively; P = 0.11), while no differences were recorded for the whole period (Table 3). There were no substantial differences in the average feed intake during the entire fattening period.
No differences among groups in mortality rate were significant, although a lower mortality rate in rabbits fed diet with mannan-oligosaccharide was observed (21.8, 26.4 and 30.9% in rabbits fed M, I and control diet, respectively; Table 4). In dead rabbits, Clostridium perfringens, Klebsiella pneumoniae and non-hemorrhagic E. coli were identified potential bacterial pathogens in the intestine and Pasteurella multocida, Klebsiella pneumoniae and non-hemorrhagic E. coli in other organs. Pathogenicity of E. coli isolates was not examined. Also coccidiosis was recorded (Eimeria magna, Eimeria media, Eimeria perforans). On the other hand, some differences were observed in both morbidity and the health risk index of young rabbits. The lowest morbidity (P = 0.05) as well as the health risk index were recorded in rabbits fed diet with inulin (40.0, 46.4 and 55.5% in rabbits fed I, M and control diet, respectively; P = 0.03). After a change of diet (at 46 days of age), however, the health risk index increased in the rabbits previously fed diet with additives (mannan-oligosaccharide or inulin), therefore no significant differences in the health status of rabbits were recorded for the entire fattening period.
Table 4 Health status of early weaned rabbits† fed control (C), mannan-oligosaccharide (M) and inulin (I) diet
† 110 rabbits per group at the beginning of the trial.
‡ Numbers of dead and sick rabbits are given in parentheses.
Caecal traits, microbial activity, and digestibility of experimental diets
The effect of the dietary treatments on several caecal traits and microbial activity is shown in Table 5. No significant effect of the diets was detected on caecum relative weight and its content. However, strong changes were observed in caecal microbial activity. Total caecal VFA concentration was significantly higher (P < 0.01) and the pH lower in rabbits fed inulin diet than in the caecum of other rabbits (5.85, 5.98, and 5.68 in rabbits fed control, M and I diet, respectively; P < 0.01). Also caecal ammonia concentration strongly decreased when rabbits received diet with inulin (143.5, 142.6 and 102.6 μg/ml in rabbits fed control, M and I diet, respectively; P = 0.01). Acetate molar proportion was higher (P = 0.01) and that of propionate and other VFA as well as the propionate/butyrate ratio significantly lower in rabbits fed diet supplemented with inulin than in other rabbits. Butyrate molar proportion was significantly higher in rabbits fed the diet with mannan-oligosaccharide (P < 0.01). The other differences were detected on the caecal bacterial fibrolytic activity. A significantly higher activity of inulinase was recorded in rabbits fed the diet supplemented with inulin. The activity of cellulase tended to be lower in rabbits fed the diet supplemented with mannan-oligosaccharide (P = 0.10).
Table 5 Caecal parameters and caecal microbial activity in early weaned rabbits fed control (C), mannan-oligosaccharide (M) or inulin (I) diet at 42 days of age
a,b Values in the same row with unlike superscript differ significantly (P < 0.05).
† RMSE = root mean square error.
‡ The assay was based on the increase in reducing power, calibrated by corresponding monomeric sugars. The activity was expressed as micromole of reducing sugars per gram of dry matter of caecal content per hour.
Whole tract digestibility coefficients of experimental diets are shown in Table 6. A lower digestibility of cellulose (P = 0.03) was observed in rabbits fed the diet supplemented with mannan-oligosaccharide than in rabbits fed the other diets. No significant differences were recorded in mean live weight and feed intake during digestibility measurements.
Table 6 Total tract apparent digestibility (TTAD) of control (C), mannan-oligosaccharide (M) and inulin (I) diet
a,b Values in the same row with unlike superscript differ significantly (P < 0.05).
† RMSE = root mean square error.
‡ Mean live weight and feed intake during digestibility measurements (from 36 to 40 days old).
§ Non-nitrogenous cellular content (NNCC) = OM - CP - NDF.
Discussion
In the present study we observed substantial differences in the caecal metabolism in rabbits (Table 5). Compared with the control and MOS diet, the inulin addition strongly influenced the caecal fermentative activity (VFA, pH, ammonia, fermentation pattern) in a beneficial way, which is in agreement with previous studies on rabbits (Morisse et al., Reference Morisse, Maurice, Boilletot and Cotte1993; Maertens et al., Reference Maertens, Aerts and De Boever2004; Volek et al., Reference Volek, Marounek and Skřivanová2005) or rats (Juśkiewicz et al., Reference Juśkiewicz, Zduńczyk, Klewicki and Gomez-Villalva2004 and Reference Juśkiewicz, Zduńczyk and Wróblewska2005). Whatever the diet, the activity of pectinase was higher than the activities of other fibrolytic enzymes (xylanase, inulinase, cellulase), which is in agreement with the hierarchy in fibrolytic activities reported in other studies (Marounek et al., Reference Marounek, Vovk and Skřivanová1995; Gidenne et al., Reference Gidenne, Jehl, Segura and Michalet-Doreau2002). In rabbits fed the diet supplemented with mannan-oligosaccharide, we found the lowest activity of cellulase (P = 0.10) in comparison with other groups. Some negative effects of MOS on the activity of caecal microflora were observed in turkeys. Juśkiewicz et al. (Reference Juśkiewicz, Zduńczyk and Jankowski2003) reported a decrease in the bacterial glycolytic activity (especially in α-glucosidase, α-galactosidase and β-galactosidase activities) in the caecal digesta of young turkeys fed diets with MOS. In our study, a significantly higher activity of inulinase was recorded in rabbits fed the diet with inulin, which apparently corresponds to a high caecal fermentative activity in these animals. The results of our study indicated a very good utilisation of inulin-type fructans by the caecal microflora of rabbits.
In the present study we found some differences in the digestibility of diets (Table 6). A significantly lower digestibility coefficient of cellulose was observed in rabbits fed the diet supplemented with MOS. Similarly, a significantly lower digestibility coefficient of crude fibre was observed in the growing pigs fed a diet with MOS in comparison with control pigs (Lipiński et al., Reference Lipiński, Purwin, Tywończuk, Zduńczyk, Wróblewska, Łaniewska-Trokenheim and Warmińska-Radyko2005). In our study, lower digestibility of cellulose can be explained by lower cellulolytic activity observed in the caecum of these rabbits (Table 5).
Supplementation of diets with mannan-oligosaccharide or inulin had no effect on the rabbit growth, except a slightly lower weight gain during starter period in comparison with rabbits fed the control diet (P = 0.11; Table 3). Similarly, no significant effect of dietary supplementation with MOS or fructo-oligosaccharides on the weight gain of rabbits was observed in other studies (Luick et al., Reference Luick, El-Sayaad and Cheeke1992; Scapinello et al., Reference Scapinello, de Faria, Furlan and Michelan2001; Fonseca et al., Reference Fonseca, Falcão, Kocher and Spring2004). There were no substantial differences among groups in average feed intake values for the entire fattening period.
No mortality was registered before 30 days of age. On the contrary, mortality rate increased sharply between 32 and 46 days of age. There were no significant differences among groups in mortality during starter period (25 to 46 days of age), although it is noteworthy that a lower mortality rate in rabbits fed the diet with mannan-oligosaccharides compared with the control diet occurred (Table 4). Fonseca et al. (Reference Fonseca, Falcão, Kocher and Spring2004) reported a significant reduction of mortality rate when traditionally weaned rabbits were fed a diet supplemented with mannan-oligosaccharide during the entire fattening period, but they used a larger number of animals in groups. In our study rabbit mortality was associated with specific enteropathy. A higher susceptibility of early weaned rabbits to a specific pathology, and thereby a high health risk index observed in this study corresponds with other studies (Gidenne and Fortun-Lamothe, Reference Gidenne and Fortun-Lamothe2004; Volek et al., Reference Volek, Marounek and Skřivanová2005), where no antibiotics were used. As mentioned above, coccidiosis was also observed. In the present study we used the anticoccidial Robenidine (66 mg/kg diet), which is not apparently sufficient for the protection of rabbits against these pathogens. In fact, the resistance of some coccidia (E. magna, E. media, E. perforans) to this drug has been mentioned (Licois, Reference Licois2004). In spite of non-significant differences in mortality rate among groups during the starter period, morbidity of rabbits as well as the health risk index were affected by dietary treatments used in this study. Both the lowest morbidity (P = 0.05) and health risk index (P = 0.03) were recorded in rabbits fed the diet supplemented with inulin. Similarly, a protective effect of fructo-oligosaccharides was recorded in animals experimentally infected with enteropathogenic E. coli (Morisse et al., Reference Morisse, Maurice, Boilletot and Cotte1993), and also when inulin was added to a high-pectin diet (Volek et al., Reference Volek, Marounek and Skřivanová2005). The better health status of rabbits fed the diet with inulin observed in our study was apparently associated with the beneficial effect of inulin on the caecal microbial activity (Table 5). In fact, Gidenne and Licois (Reference Gidenne and Licois2005) reported that a high fibre intake improved the resistance of the growing rabbit to a specific enteropathy experimentally induced by an enteropathogenic E. coli, in association with a higher caecal fermentative activity and lower caecal pH. When the diets were changed (at 46 days of age), the health risk index increased in the rabbits previously fed with the mannan-oligosaccharide or inulin diet, thus no significant differences in the health status of rabbits were recorded for the whole fattening period (Table 4).
Conclusion
Supplementation of the starter diet with mannan-oligosaccharide or inulin had no effect on rabbit growth during the whole fattening period. Compared with control and MOS diet, the inulin addition remarkably affected the caecal microbial activity in a beneficial way, which lowered both the morbidity and health risk index of rabbits in the first weeks after weaning. No effect was detected on the mortality of rabbits. The results of our study suggest the importance of using inulin-type fructans in the nutrition of young rabbits. The higher health risk index of rabbits observed after change of diet indicates that prebiotics should be given for a longer time during the fattening period.
Acknowledgements
This study was supported by the Czech Science Foundation, Project no. 523/03/D011
