Journal of Chemical Ecology, I1ol. 19, No. 11, 1993
INFLUENCE OF NONPROTEIN NITROGEN ON ESTIMATION OF PROTEIN FROM TOTAL NITROGEN IN FLESHY FRUITS
IDO IZHAKI Department of Biology University of Haifa at Oranim Tivon, 36910 lsrael
(Received March 8, 1993; accepted July 6, 1993) Abstract--The protein content of pulps of 26 fleshy fruit species from east Mediterranean habitats in Israel were estimated using two different methods: (1) the Kjeldahl procedure in which the total recovered nitrogen is multiplied by 6.25 to estimate total proteins, and (2) amino acid analysis by amino acid analyzer. The average protein content obtained by the Kjeldahl procedure was 5.75 % (dry weight) while it was only 3.90% when amino acids were analyzed. The higher value of protein content by the Kjeldahl procedure is most likely the result of a relatively high proportion of nonprotein nitrogen compounds (31%) in these pulps. Therefore the 6.25 factor is not valid and a 4.05 factor may be more accurate for assessing the true protein content of these fleshy fruits. The data also suggest that the more accurate estimate of true protein (Y) from Kjeldahl total nitrogen (X) should be based on the highly significant linear regression between these two variables: Y = 4.885X - 0.6. Key Words--Fmgivory, seed dispersal, amino acids, Kjeldahl, protein, fleshy fruit, nutrition, secondary compounds, plant-animal interactions.
INTRODUCTION
O n e of the most frequently used methods for determining dietary crude protein is the Kjeldahl procedure, published nearly a century ago, which is based on total nitrogen content o f food items (Robbins, 1983). Although this method determines nitrogen, the results are converted to crude protein by multiplying N x 6.25 ( M a y n a r d and Loosli, 1969; Bondi, 1987). The factor 6.25 is derived from the average analysis of protein, being 16 % nitrogen. However, the nitrogen content of proteins from different food types varies, d e p e n d i n g o n the a m o u n t 2605 0098-0331/93/1100-2605507.00/0 9 1993 Plenum Publishing Corporation
2606
IZHAKI
of nonprotein nitrogen compounds. Therefore, the true protein content of many food items calculated by the Kjeldahl method is probably overestimated because crude protein is based on an analysis of total nitrogen and not on protein nitrogen (e.g., Hansen, 1970). The traditional 6.25 factor is widely used for estimating the crude protein content of fleshy pulps of fruits (e.g., Piper, 1986; Debussche et al., 1987; Herrera, 1987; Sakai and Carpenter, 1990). In several studies this factor was used to estimate the nutritional value of the fruit for frugivorous animals (e.g., Foster, 1978; Poddar and Lederer, 1982; Calvert, 1985; Herrera, 1989; Izhaki and Safriel, 1989; Rogers et al., 1990; Izhaki, 1992; Kool, 1992). Because pulps may contain appreciable amounts of nonprotein nitrogen compounds, the 6.25 factor may overestimate the protein content of these fruits. Fruits may contain different sources of nonprotein nitrogen such as inorganic nitrogen, lowmolecular-weight peptides, nucleic acids, free amino acids, ammonium salts, and secondary compounds (Maynard and Loosli, 1969; Lyttleton, 1973). Alkaloids are the best known of the nitrogen-containing secondary metabolites of plants, but nonprotein amino acids, cyanogenic glycosides, and glucosinolates also occur in plants (Harborne, 1991). Milton and Dintzis (1981) suggested that the appropriate nitrogen-to-protein conversion factor for leaves, flowers, and fruit from eight tropical woody species was 4.4. Their data showed that the proteins in these samples averaged 19% nitrogen, and an average of 20% of the total nitrogen in these samples was nonproteinaceous. However, their study was mainly based on leaves, and included only one species of fleshy fruit (Ficus insipida). Here I report the amino acid content of the largest sample of noncultivated fleshy fruit species analyzed to date. For each species, the total protein content was calculated by summing the amino acid content and was compared to the estimate of crude protein by Kjeldahl method.
METHODS AND MATERIALS
Research Species. Twenty-six plant species that produce fleshy fruits in the northern part of Israel were studied. These species belong to 17 families and may be categorized into four life forms: 11 trees, seven shrubs, seven climbers, and one herbaceous plant. Most of these species are bird-dispersed plants (Izhaki, 1986; Izhaki and Safriel, 1985; Bamea et al., 1991; Izhaki et al., 1991), but at least two of them (Ziziphus spina-christi, Styrax officinalis) are mainly mammaldispersed. There is some evidence that frugivorous birds are unable to subsist on a diet comprised of one or several of these fruits (Izhaki, 1992; Izhaki and Safriel, 1989). The nonexclusive suggested explanations of this phenomenon include the low protein content of the pulp (Berthold, 1976; Levey and Karasov,
2607
ESTIMATION OF PROTEIN IN FLESHY FRUIT
1989; Sedinger, 1990), shortage of specific amino acids (Parrish and Martin, 1977; Mack, 1990; Sedinger, 1990), and secondary compounds that interfere with protein metabolism (Izhaki and Safriel, 1989). Methods. Fresh ripe fruits were collected in the field, the seeds removed, and the pulp oven dried at 40~ to constant mass. The dried pulp was ground to powder, and the sample was divided into two fractions for two different analyses. The total nitrogen content of the pulp was determined by the Kjeldahl technique (AOAC, 1984) using a Kjeltec Auto 1030 Analyzer. In this method, the pulp sample was boiled in a concentrated sulfuric acid-catalyst solution until all organic matter was destroyed and the nitrogen was converted to ammonium sulfate. The ammonia was volatilized by sodium hydroxide (AOAC, 1984). Amino acid composition and released ammonia were determined according to the procedure described by Elkin and Griffith (1985). The pulp powders were oxidized with performic acid prior to hydrolysis, and their amino acid contents were determined by cation exchange chromatography using a Biotronic LC 5000 Amino Acid Analyzer. HC1 was used to destroy excess performic acid (Elkin and Griflith, 1985). Released ammonia and 17 common amino acids were measured (Table 1). No analysis was made for tryptophan, but this should not alter protein values or conversion factors significantly (Milton and Dintzis, 1981).
RESULTS
AND
DISCUSSION
The amino acid composition varied greatly among the species (Table 1). Ephedra aphylla and Withania somnifera both have similar, relatively high, total protein estimates (14.5% and 14.2%, respectively), yet E. aphylla is rich in phenylalanine and valine, whereas W. somnifera is rich in aspartic acid, glutamic acid, and arginine (Table 1). Differences in amino acid composition were also detected in congeneric species. Although the general patterns of amino acid composition and total protein were relatively similar for the two studied species of Rhamnus, R. lycioides pulp was twice as rich in aspartic acid as R. alaternus (Table 1). Protein quality, expressed by amino acid composition, may have more relevance to patterns of frugivory than total protein per se (Parrish and Martin, 1977; Mack, 1990). The total nitrogen recovered by the Kjeldahl determination was significantly higher than the total nitrogen recovered by the amino acid analysis for all studied species (paired t test on arcsin square-root transformed values, T = 11.6, df = 25, P < 0.0001 Table 2). The average difference between these two results is 31% and may be considered as nonproteinaceous nitrogen content of these fleshy fruits (Table 2). This average is larger than the value (20%) reported by Milton and Dintzis (1981) for tropical plants. In three plant species, Osyris alba, Rhamnus lycioides, and Tamus orientalis, 50 % or more of their pulp nitrogen may
2608
IZHAga
TABLE 1. TOTAL PULP AMINO ACID COMPOSITION OF 26 FRUIT SPECIES FROM EAST MEDITERRANEAN HABITATS (% Dry Weight)
Cysteine
Aspartic acid
Methionine
Threonine
Serine
Glutamic acid
Ephedra aphylla (Ephedraceae)
0.232
Osyris alba (Santalaceae) Laurus nobilis (Lauraceae) Rosa canina (Rosaceae) Crataegus monogyna (Rosaceae)
0.045
0.853
0.178
0.461
0.393
2.068
1.116
0.293
0.050
0.111
0.124
0.252
0.174
0.040 0.055
0.390 0.471
0.123 0.045
0.163 0.126
0.162 0.148
0.352 0.271
0.264 0.396
Pistacia lentiscus (Anacardiaceae)
0.021
0.186
0.027
0.068
0.073
0.106
0.090
0.074
0.362
0.131
0.174
0.206
0.364
Pistacia atlantica (Anacardiaceae)
0.464
0.067
0.469
0.143
0.199
0.250
0.391
0.482
0.059
0.363
0.100
0.151
0.188
0.348
0.300
0.035
0.204
0.042
0.095
0.106
0.197
0.117
0.031
0.186
0.024
0.062
0.074
0.152
0.070
0.023
0.333
0.025
0.068
0.063
0.130
0.096
Ziziphus spina-christi (Rhamnaceae) Myrtus communis (Myrtaceae)
0.032
0.737
0.026
0.078
0.102
0.239
0.080
0.031
0.150
0.032
0.066
0.073
0.167
0.087
Arbutus andrachne (Ericaceae)
0.014
0.074
0.013
0.036
0.040
0.092
0.041
Styrax officinalis (Styracaceae) Phillyrea latifolia (Oleaceae) Withania somnifera (Solanaceae) Rubia tenuifolia (Rubiaceae)
0.023
0.157
0.039
0.051
0.073
0.248
0.043
0.056
0.319
0.048
0.140
0.169
0.330
0.163
0.415
1.698
0.253
0.612
0.707
1.792
1.178
0.030
0.256
0.032
0.075
0.091
0.175
0.221
Viburnum tinus (Caprifoliaceae) Lonicera etrusca (Caprifoliaeeae)
0.044
0.238
0.062
0.114
0.125
0.238
0.136
0.028
0.228
0.038
0.095
0.104
0.236
0.091
Bryonia spp. (Cucurbitaceae) Asparagus aphyllus (Liliaceae)
0.195 0.167
1.521 0.523
0.189 0.081
0.573 0.250
0.818 0.269
0.956 0.521
0.730 1.090
Ruscus aculeatus (Liliaceae)
0.056
0.335
0.039
0.113
0.143
0.378
0.223
Smilax aspera (Liliaceae) Tamus orientalis (Dioscoreaceae) Arum dioscoridis (Araceae)
0.076
0.356
0.049
0.131
0,149
0.290
0.171
0.038
0.259
0.042
0.111
0.127
0.332
0.126
0.083
0.406
0.073
0.174
0.179
0.874
0.209
PLant species
Pistacia palaestina (Anacardiaceae) Rhus coriaria (Anacardiaceae) Rhamnus alaternus (Rhamnaceae) Rhamnus lycioides (Rhamnaceae)
Proline
be attributable to nonprotein compounds. Using the traditional 6.25 factor yielded an average of 5.75 % crude protein in the analyzed fruits while the average protein content based on the total amino acid content was only 3.90% (paired t test on arcsin square-root transformed values, T = 11.0, df = 25, P < 0.0001, Table 2). These data suggest that levels of protein in fruit pulp are even lower than previously estimated. This would lend support to the suggestion that protein
ESTIMATION OF PROTEIN IN FLESHY FRUIT
2609
TABLE 1. CONTINUED
PhenylGlycine
Alanine
Valine
Isoleucine
Leucine
Tyrosine
alanine
Histidine
Lysine
Arginine
Ammonia
0,652 0.140
0.768
1.250
0.478
0,850
0.549
2.9t0
0.226
0,430
0.302
0.793
0.135
0.147
0.110
0.194
0.095
0.127
0.064
0.116
0.193
0.045
0.174
0.204
0.216
0.163
0.279
0.144
0,161
0.083
0.238
0.181
0.073
0.127
0.128
0.145
0.107
0,166
0.117
0.116
0.081
0.128
0.209
0.061
0.067
0.072
0.088
0.057
0.103
0.057
0,063
0.033
0.085
0.070
0.048
0.202
0.220
0.222
0.176
0.301
0.188
0.205
0.116
0.275
0.246
0.104
0.255
0.271
0.298
0.220
0.356
0.289
0.245
0.147
0.378
0.374
0.080
0.193
0.201
0,234
0.151
0.247
0.186
0.170
0.094
0.238
0.195
0.082
0.109
0.114
0.121
0.101
0.173
0.097
0.117
0.045
0.136
0.104
0.076
0.085
0.076
0.083
0,069
0.119
0.053
0,070
0.034
0.056
0.064
0.122
0.077
0.064
0.093
0.050
0.117
0.057
0,065
0.032
0.068
0.066
0.051
0.100
0.128
0.097
0,072
0.143
0.077
0.075
0.046
0.109
0.541
0,096
0.094
0.084
0.083
0.063
0.117
0.069
0.071
0.039
0.086
0.079
0,041
0.042
0.044
0.050
0.037
0.057
0.024
0.034
0.016
0.037
0.041
0.043
0.061
0.102
0.069
0.034
0.080
0.039
0,048
0.032
0.051
0.056
0.041
0,172
0.161
0.178
0.155
0,247
0.135
0.156
0.052
0,032
0.166
0.080
0.709
0,755
0.663
0,584
0.866
0.573
0.559
0.389
0.911
1.213
0.326
0,141
0.107
0.108
0.084
0.122
0.077
0.084
0.037
0.059
0.089
0.047
0.151
0,129
0.140
0.110
0.197
0.094
0.127
0.050
0.141
0.131
0.093
0.112
0.110
0.110
0.107
0,159
0.087
0.103
0.033
0.052
0.108
0.064
0.654
0.723
0.841
0,663
0.868
0.596
0.719
0.208
0.576
0.892
0.264
0.422
0.321
0.327
0.268
0.395
0.293
0.277
0.115
0.265
0.299
0.103
0.168
0.182
0.157
0.128
0.150
0.109
0.111
0.069
0,155
0,263
0.111
0.240
0.184
0.168
0,129
0.214
0.216
0.130
0.071
0.161
0.191
0.094
0.146
0.189
0.143
0,110
0.178
0.114
0.116
0.044
0.084
0.179
0.070
0.241
0.338
0.276
0.169
0.285
0.149
0.217
0,073
0.114
0.279
0.078
limitation is one reason for the inadequacy of a strict fruit pulp diet (Levey and Karasov, 1989; Sedinger, 1990). The average ratio between total amino acids and total amino acid nitrogen was 6.32 (Table 2). Therefore, the amino acids present in these fruits averaged 15.8% nitrogen (100/6.32) which is close to the expected value of 16%, but since these fruits contained nonprotein nitrogen as discussed before, one should reevaluate the valid conversion factor for the Kjeldahl procedure. The 4.05
Ephedra aphylla (Ephedraceae) Osyris alba (Santalaceae) Laurus nobilis (Lauraceae) Rosa canina (Rosaceae) Crataegus monogyna (Rosaceae) Pistacia lentiscus (Anacardiaceae) Pistacia atlantiea (Anacardiaceae) Pistacia palaestina (Anacardiaceae) Rhus coriaria (Anacardiaceae) Rhamnus alaternus (Rhamnaceae)
Plant species
Estimated protein (N x 6.25) (%) 20.3 6.2 4.3 4.8 2.1 5.1 8.0 6.1 3.0 2.8
Total
nitrogen (%)
3.25 0.99 0.69 0.77 0.34 0.82 1.28
0.98 0.48 0.45
Kjeldahl determination
0.53 0.32 0.27
2.29 0.37 0.51 0.44 0.21 0.63 0.75
N + ammonia) (%)
Nitrogen (amino acid
3.5 2.0 1.4
14.5 2.4 3.4 2.9 1.3 4.0 4.9
(total amino acids) (%)
Protein
Amino acid analysis
4.2 1.7 1.6
15.3 4.2 2.8 3.2 1.1 3.4 5.7
predicted (%)"
Protein
0.7 -0.3 0.2
0.8 1.8 -0.6 0.3 -0.2 -0.6 0.8
Residuals (%)/,
Evaluation of the regression equation
6.54 6.19 5.21
6.34 6.54 6.66 6.56 6.21 6.41 6.54
acids/total amino acid N
3.59 4.14 3.19
4.47 2.43 4.96 3.77 3.57 4.94 3.84
acids/total Kjeldahl N
Total amino
Conversion factors
0.45 0.33 0.39
0.30 0.63 0.26 0.42 0.42 0.23 0.41
Nonproteinous N'
FRUIT SPECIES FROM EAST
Total amino
26 FLESHY MEOITERRANEANHABITATS(Values are Percentage Dry-Weight)
TABLE 2. ANALYSIS OF NITROGEN, PROTEIN, AND CONVERSION FACTORS OF PULPS OF
~v
4.7 6.6 5.75 _+ 4.74
1.06
0,92 + 0.76
3.3 t4.8
0.53 2.37
4.5
2.9
0.46
0.75
2.5
0.40
0.72
17.6
2.82
6.7
3.5
0.56
5.8
2.3
0.37
1.07
I. I
0. t 8
0.93
2.1
5.4
0.86
0.34
3.0
0.48
0.61 _+ 0.59
0,62
0,38
0.48
0.49
0.88
0.29 1.83
0.38
0.28
2.25
0.42
0.19
0.13
0.22
0.51
0.23
"Estimate based on Y = (4.885X) - 0.6, ~'% p r o t e i n (total a m i n o a c i d s ) - % p r o t e i n p r e d i c t e d . " ( % total K j e l d a h l N - % total a m i n o a c i d N ) / % total kjeldahl N .
Myrtus communis (Myrtaceae) Arbutus andrachne (Ericaceae) Styrax officinalis (Styracaceae) Phillyrea latifi~lia (Oleaceae) Withania somnifera (Solanaceae) Rubia tenuifolia (Rubiaceae) Viburnum tinus (Caprifoliaceae) Lonicera etrusca (Capritbliaceae) Bryonia spp. (Cucurbitaceae) Asparagus aphyllus (Liliaceae) Ruscus aculeatus (Liliaceae) Smilax a,wera (Liliaceae) Tamus orientalis (Dioscoreaceae) Arum dioscoridis (Araceae)
(Rhamnaceae)
Rhamnus lycioides (Rhamnaceae) Ziziphus spina-christi 1.5
3.90 + 3.78
4.2
2.4
3.0
2.9
6.0
1.9 12.0
2.3
1.8
14,2
2.8
1.2
0.7
1.4
2.8
1.7
4.6
3. I
2.9
3.9
4,6
2.0 I 1.0
1.6
1.4
13.2
2.1
1.2
0.3
1,1
3.6
0.4
0.7
-0. I
1.0
1,4
0.1 -0. I
-0.7
-0.4
- 1.0
- 0.7
0,0
-0.4
-0.3
0.8
0.2
6.32 _+ 0.38
6.83
6.35
6,27
5.88
6.82
6.41 6,54
6.10
6.61
6,33
6,58
6.48
5.68
6,40
5.44
6.46
4.05 -I_- 0,79
3.99
3.20
4.19
3.1 I
5.58
3.53 5.06
5.00
4,59
5,04
4.93
3.39
4.18
4,26
3.22
3.07
0.37 ! 0.11
0.42
0.50
0.33
0.47
0.18
0.45 0.23
0.18
0.31
0.20
0,25
0.48
0.26
0.33
0.4 I
0.52
t-~
,.~
~
Z
"n
Z 9
> ,--1
~,
2612
IZHAKI
average factor for all species is one approach to converting Kjeldahl total nitrogen to protein content (Table 2). This factor is slightly lower than the 4.4 factor suggested by Milton and Dintzis (1981) for tropical plants but higher than the average factor of 3.55 detected by Herbst (1986) for four tropical fruit species. A significant linear equation Y = 4.885X - 0.6 was detected between protein content and total Kjeldahl N (Figure 1). This equation may be more accurate for estimating protein content than the 4.05 factor because of the positive intercept on the x axis (Figure 1). Despite the relatively good estimation of true protein on total Kjeldahl N based on this regression, nonprotein N estimates range from less than 20% to more than 60% for this set of species. It is possible that nonprotein N may vary even more for other species. Further, analysis of residuals of the regression (Table 2) reveals that estimated total protein from total N can be relatively inaccurate for particular species. For instance, protein is underestimated by 133% for Arbutus andrachne and overestimated by 43 % for Osyris alba upon applying the regression equation (Table 2). Typical errors for other species tend to be in the 10-30% range. In addition, the proportional error tends to be greater with decreasing total N; nevertheless, most fruit pulps tend to be low in N. Thus, the regression of total protein on total N would seem to be adequate only for broad-scale estimates where positive and negative errors would tend to cancel. Such a case is the estimation of total protein intake for animals on a mixed-fruit diet. In contrast, any estimate of protein content derived from total N would be inadequate when accurate interspecific comparisons are 21 18
Z LLI
I5 +
I0
12
o_
9
,( I0 I-
6
+
+
+
3 0 0.0
J
i
i
i
p
i
i
i
0.5
1.0
1.5
2.0
2.5
3.0
3.5
KJELDAHL +
AMINO ACIDS
TOTAL ,~
4.0
N (%)
KJELDAHL
FIG. 1. The crude total protein content (calculated by multiplying total N by the traditional 6.25 factor) and the true protein content (using the total amino acid content) of 26 fruit species plotted against Kjeldahl total N. The linear equation for calculating true protein as a function of Kjeldahl total N is Y = 4.885X - 0.6, (R2 = 0.96, P < 0.0001).
E S T I M A T I O N OF PROTEIN IN F L E S H Y F R U I T
2613
needed. Thus, while the use of the regression is indeed better than the use of the 6.25 conversion factor, the data suggest that it may be risky to use total N to estimate protein in fruit pulps, especially of those species with unusually high or low levels of nonprotein N. The relatively large differences in nitrogen content estimated by the Kjeldahl procedure and the one calculated by the amino acid analysis (Table 2) were probably not an artifact effect of these methods. As discussed earlier by Milton and Dintzis (1981), it is possible that the HC1 hydrolysis in the amino acid analysis does not release all the nitrogen from compounds that are tightly bonded as plant-cell material, while the sulfuric acid in the Kjeldahl method releases this nitrogen. However, most plant protein is found within the cytoplasm of the cell, with only small amounts associated with the cell wall (Lyttleton, 1973; Albersheim, 1975). Further, if such nitrogen exists, it is probably not digestible by animals. Therefore, it seems that most of the differences in nitrogen content between the two methods is due to the high levels of nonprotein nitrogen in wild fruits. There is little information on the nonprotein nitrogen compounds in fruits. The presence of alkaloids and glycosides is frequent among European and Mediterranean species even in ripe fruits (Jordano, 1988; Ehrl6n and Eriksson, 1993, and references therein) and is also reported in other ecosystems (e.g., Calvert, 1985; Potter and Kimmerer, 1986; Gargiullo and Stiles, 1991). Therefore, these compounds may constitute a major fraction of the nonprotein nitrogen compounds in fleshy fruits. In order to confirm this assumption, a direct analysis of secondary compounds in these fruits should be carried out. The significance of such compounds to the evolution of mutual relations between plants that produce fleshy fruits and their frugivorous seed dispersers is still not clear (see Jordano, 1991, and references therein). Acknowledgments--I wish to thank Y. Pancheshnikovafor assisting in the lab, Dr. E. Bamberger for helping during different stages of this study, and an anonymousreviewer for very helpful suggestions and comments that improved the manuscript. Kjeldahl analyses were performed in Miloda Laboratories under D. Izhar's supervision. Amino acid analyses were performedat Aminolab Ltd. under Z. Harduf's supervision. This study was partly supported by a joint research grant of the Technion and University of Haifa, by the Carmel Foundation-Ministryof Environmental Affairs, and by a grant from the Joint German Israeli Research Program. The idea for this study was an outcome of a fruitful personal communicationwith Carlos M. Herrera.
REFERENCES AOAC. 1984. Official Methods of Analysis of the Association of Official Analytical Chemists. AOAC, Washington, D.C. ALBERSHEIM,P. 1975. The walls of growing plant cells. Sci. Am. 232:80-95. BARNEA,A., YOMTOV,Y., and FRIEDMAN,J. 1991. Does ingestion by birds affect seed germination? Funct. Ecol. 5:394-402.
2614
IZHAKI
BERTI-IOLD, P.B. 1976. The control and significance of animal and vegetable nutrition in omnivorous songbirds. Ardea 64:140-154. BONDI, A.A. 1987. Animal Nutrition. John Wiley & Sons, New York. CALVERT, J.J. 1985. Food selection by western gorillas ( G.g. gorilla) in relation to food chemistry. Oecologia 65:236-246. DEBUSSCHE, M., CORTEZ, J., and RIMBAULT,I. 1987. Variation in fleshy fruit composition in the Mediterranean region: The importance of ripening season, life-form, fruit type and geographical distribution. Oikos 49:244-252. EHRLt~N, J., and ERIKSSON, O. 1993. Toxicity in fleshy fruits--a non-adaptive trait? Oikos 66:107113. ELKIN, R.G., and GRJFFITH, J.E. 1985. Hydrolysate preparation for analysis of amino acids in sorghum grains: Effect of oxidative pretreatment. J. Assoc. Anal. Chem. 68:1117-1121. FOSTER, M.F. 1978. Total frugivory in tropical passerines: A reappraisal. Trop. Ecol. 19:131-154. GARGIULLO,M.B., and ST~LES, E.W. 1991. Chemical and nutritional differences between two birddispersed fruits: flex opaca and Rex verticillata. J. Chem. Ecol. 17:1091-1106. HANSEN, E. 1970. Proteins, in A.C. Hulme (ed.). The Biochemistry of Fruits and Their Products, Vol. l, Series on Food, Science & Technology. Academic Press, New York. HARBORNE, J.B. 1991. The chemical basis of plant defense, pp. 45-59, in R.T. Palo and C.T. Robbins (eds.). Plant Defenses Against Mammalian Herbivory. CRC Press, Boca Raton, Florida. HERBST, L.H. 1986. The role of nitrogen from fruit pulp in the nutrition of the frugivorous bat Carollia perspicillata. Biotropica 18:39-44. HERRERA, C.M. 1987. Vertebrate-dispersed plants of the Iberian Peninsula: A study of fruit characteristics. Ecol. Monogr. 57:305-331. HERRERA,C.M. 1989. Frugivory and seed dispersal by carnivorous mammals, and associated fruit characteristics, in undisturbed Mediterranean habitats. Oikos 55:250-262. IZHAKI, I. 1986. Seed dispersal by birds in an eastern Mediterranean scrublands. PhD dissertation. The Hebrew University of Jerusalem, Israel (in Hebrew). IZHAKI, I. 1992. A comparative analysis of nutritional quality of mixed and exclusive fruit diets for yellow-vented bulbuls. Condor 94:912-923. IZHAKI, I., and SAFRIEL,U.N. 1985. Why do fleshy-fruit plants of the Mediterranean scrub intercept fall--but not spring--passage of seed-dispersing migratory birds? Oecologia 67:40-43. IZHAKI, I., and SAFRIEL~U.N. 1989. Why are there so few exclusive frugivorous birds? Experiments on fruit digestibility. Oikos 54:23-32. IZHAKI, I., WALTON,P., and SAFRIEL,U.N. 1991. Seed shadows generated by frugivorous birds in an eastern Mediterranean scrub. J. Ecol. 79:579-590. JORDANO, P. 1988. Diet, fruit choice and variation in body condition of frugivorous warblers in Mediterranean scrubland. Ardea 76:193-209. JORDANO, P. 1991. Fruits and frugivory, pp. 105-156, in M. Fenner (ed.). Seeds: The Ecology of Regeneration in Plant Communities. C.A.B. International, Willingford. KOOL, K.M. 1992. Food selection by the silver leaf monkey, Trachypithecus auratus sondaicus, in relation to plant chemistry. Oecologia 90:527-533. LEVEY, D.J., and KARASOV,W.H. 1989. Digestive responses of temperate birds switched to fruit or insect diets. Auk 106:675-686. LYTTLETON, J.W. 1973. Proteins and nucleic acids, pp. 63-103, in G.W. Butler and R.W. Bailey (eds.). Chemistry and Biochemistry of Herbage, Vol. 1. Academic Press, New York. MACK, A.L. 1990. Is frugivory limited by secondary compounds in fruits? Oikos 57:135-138. MAYNARD, A.B., and LOOSLI, J.K. 1969. Animal Nutrition. McGraw-Hill, New York. MILTON, K., and DINTZlS, F.R. 1981. Nitrogen-to-protein conversion factors for Tropical plant samples. Biotropica 13:177-181.
ESTIMATION OF PROTEIN IN FLESHY FRUIT
2615
PARRISH, J.W., and MARTIN, E.W. 1977. The effect of dietary lysine on the energy and nitrogen balance of the dark-eyed junco. Condor 79:24-30. P~PER, J.K. 1986. Seasonality of fruit characters and seed removal by birds. Oikos 46:303-310. PODDAR, S., and LEDERER, R.J. 1982. Juniper berries as an exclusive winter forage for Townsend's solitaires. Am. Midl. Nat. 108:34-40. POTTER, D.A., and KIMMERER, T.W. 1986. Seasonal allocation of defense investment in llex opaca Ait. and constraints on a specialist leaf miner. Oecologia 69:217-224. ROBBINS, C.T. 1983. Wildlife Feeding and Nutrition. Academic Press, New York. ROGERS, M.F., MA~SELS, F., W~LUAMSON, E.A., FERNANDEZ, M., and TUTIN, C.E.G. 1990. Gorilla diet in the Lope Reserve, Gabon. Oecologia 84:326-339. SAKAI, H.F., and CARPENTER, J.R. 1990. The variety and nutritional value of foods consumed by Hawaiian crow nestlings, and endangered species. Condor 92:220-228. SEDINGER, J.S. 1990. Are plant secondary compounds responsible for negative apparent metabolizability of fruits by passerine birds? A comment on Izhaki and Safriel. Oikos 57:138-140.
INFLUENCE OF NONPROTEIN NITROGEN ON ESTIMATION OF PROTEIN FROM TOTAL NITROGEN IN FLESHY FRUITS
IDO IZHAKI Department of Biology University of Haifa at Oranim Tivon, 36910 lsrael
(Received March 8, 1993; accepted July 6, 1993) Abstract--The protein content of pulps of 26 fleshy fruit species from east Mediterranean habitats in Israel were estimated using two different methods: (1) the Kjeldahl procedure in which the total recovered nitrogen is multiplied by 6.25 to estimate total proteins, and (2) amino acid analysis by amino acid analyzer. The average protein content obtained by the Kjeldahl procedure was 5.75 % (dry weight) while it was only 3.90% when amino acids were analyzed. The higher value of protein content by the Kjeldahl procedure is most likely the result of a relatively high proportion of nonprotein nitrogen compounds (31%) in these pulps. Therefore the 6.25 factor is not valid and a 4.05 factor may be more accurate for assessing the true protein content of these fleshy fruits. The data also suggest that the more accurate estimate of true protein (Y) from Kjeldahl total nitrogen (X) should be based on the highly significant linear regression between these two variables: Y = 4.885X - 0.6. Key Words--Fmgivory, seed dispersal, amino acids, Kjeldahl, protein, fleshy fruit, nutrition, secondary compounds, plant-animal interactions.
INTRODUCTION
O n e of the most frequently used methods for determining dietary crude protein is the Kjeldahl procedure, published nearly a century ago, which is based on total nitrogen content o f food items (Robbins, 1983). Although this method determines nitrogen, the results are converted to crude protein by multiplying N x 6.25 ( M a y n a r d and Loosli, 1969; Bondi, 1987). The factor 6.25 is derived from the average analysis of protein, being 16 % nitrogen. However, the nitrogen content of proteins from different food types varies, d e p e n d i n g o n the a m o u n t 2605 0098-0331/93/1100-2605507.00/0 9 1993 Plenum Publishing Corporation
2606
IZHAKI
of nonprotein nitrogen compounds. Therefore, the true protein content of many food items calculated by the Kjeldahl method is probably overestimated because crude protein is based on an analysis of total nitrogen and not on protein nitrogen (e.g., Hansen, 1970). The traditional 6.25 factor is widely used for estimating the crude protein content of fleshy pulps of fruits (e.g., Piper, 1986; Debussche et al., 1987; Herrera, 1987; Sakai and Carpenter, 1990). In several studies this factor was used to estimate the nutritional value of the fruit for frugivorous animals (e.g., Foster, 1978; Poddar and Lederer, 1982; Calvert, 1985; Herrera, 1989; Izhaki and Safriel, 1989; Rogers et al., 1990; Izhaki, 1992; Kool, 1992). Because pulps may contain appreciable amounts of nonprotein nitrogen compounds, the 6.25 factor may overestimate the protein content of these fruits. Fruits may contain different sources of nonprotein nitrogen such as inorganic nitrogen, lowmolecular-weight peptides, nucleic acids, free amino acids, ammonium salts, and secondary compounds (Maynard and Loosli, 1969; Lyttleton, 1973). Alkaloids are the best known of the nitrogen-containing secondary metabolites of plants, but nonprotein amino acids, cyanogenic glycosides, and glucosinolates also occur in plants (Harborne, 1991). Milton and Dintzis (1981) suggested that the appropriate nitrogen-to-protein conversion factor for leaves, flowers, and fruit from eight tropical woody species was 4.4. Their data showed that the proteins in these samples averaged 19% nitrogen, and an average of 20% of the total nitrogen in these samples was nonproteinaceous. However, their study was mainly based on leaves, and included only one species of fleshy fruit (Ficus insipida). Here I report the amino acid content of the largest sample of noncultivated fleshy fruit species analyzed to date. For each species, the total protein content was calculated by summing the amino acid content and was compared to the estimate of crude protein by Kjeldahl method.
METHODS AND MATERIALS
Research Species. Twenty-six plant species that produce fleshy fruits in the northern part of Israel were studied. These species belong to 17 families and may be categorized into four life forms: 11 trees, seven shrubs, seven climbers, and one herbaceous plant. Most of these species are bird-dispersed plants (Izhaki, 1986; Izhaki and Safriel, 1985; Bamea et al., 1991; Izhaki et al., 1991), but at least two of them (Ziziphus spina-christi, Styrax officinalis) are mainly mammaldispersed. There is some evidence that frugivorous birds are unable to subsist on a diet comprised of one or several of these fruits (Izhaki, 1992; Izhaki and Safriel, 1989). The nonexclusive suggested explanations of this phenomenon include the low protein content of the pulp (Berthold, 1976; Levey and Karasov,
2607
ESTIMATION OF PROTEIN IN FLESHY FRUIT
1989; Sedinger, 1990), shortage of specific amino acids (Parrish and Martin, 1977; Mack, 1990; Sedinger, 1990), and secondary compounds that interfere with protein metabolism (Izhaki and Safriel, 1989). Methods. Fresh ripe fruits were collected in the field, the seeds removed, and the pulp oven dried at 40~ to constant mass. The dried pulp was ground to powder, and the sample was divided into two fractions for two different analyses. The total nitrogen content of the pulp was determined by the Kjeldahl technique (AOAC, 1984) using a Kjeltec Auto 1030 Analyzer. In this method, the pulp sample was boiled in a concentrated sulfuric acid-catalyst solution until all organic matter was destroyed and the nitrogen was converted to ammonium sulfate. The ammonia was volatilized by sodium hydroxide (AOAC, 1984). Amino acid composition and released ammonia were determined according to the procedure described by Elkin and Griffith (1985). The pulp powders were oxidized with performic acid prior to hydrolysis, and their amino acid contents were determined by cation exchange chromatography using a Biotronic LC 5000 Amino Acid Analyzer. HC1 was used to destroy excess performic acid (Elkin and Griflith, 1985). Released ammonia and 17 common amino acids were measured (Table 1). No analysis was made for tryptophan, but this should not alter protein values or conversion factors significantly (Milton and Dintzis, 1981).
RESULTS
AND
DISCUSSION
The amino acid composition varied greatly among the species (Table 1). Ephedra aphylla and Withania somnifera both have similar, relatively high, total protein estimates (14.5% and 14.2%, respectively), yet E. aphylla is rich in phenylalanine and valine, whereas W. somnifera is rich in aspartic acid, glutamic acid, and arginine (Table 1). Differences in amino acid composition were also detected in congeneric species. Although the general patterns of amino acid composition and total protein were relatively similar for the two studied species of Rhamnus, R. lycioides pulp was twice as rich in aspartic acid as R. alaternus (Table 1). Protein quality, expressed by amino acid composition, may have more relevance to patterns of frugivory than total protein per se (Parrish and Martin, 1977; Mack, 1990). The total nitrogen recovered by the Kjeldahl determination was significantly higher than the total nitrogen recovered by the amino acid analysis for all studied species (paired t test on arcsin square-root transformed values, T = 11.6, df = 25, P < 0.0001 Table 2). The average difference between these two results is 31% and may be considered as nonproteinaceous nitrogen content of these fleshy fruits (Table 2). This average is larger than the value (20%) reported by Milton and Dintzis (1981) for tropical plants. In three plant species, Osyris alba, Rhamnus lycioides, and Tamus orientalis, 50 % or more of their pulp nitrogen may
2608
IZHAga
TABLE 1. TOTAL PULP AMINO ACID COMPOSITION OF 26 FRUIT SPECIES FROM EAST MEDITERRANEAN HABITATS (% Dry Weight)
Cysteine
Aspartic acid
Methionine
Threonine
Serine
Glutamic acid
Ephedra aphylla (Ephedraceae)
0.232
Osyris alba (Santalaceae) Laurus nobilis (Lauraceae) Rosa canina (Rosaceae) Crataegus monogyna (Rosaceae)
0.045
0.853
0.178
0.461
0.393
2.068
1.116
0.293
0.050
0.111
0.124
0.252
0.174
0.040 0.055
0.390 0.471
0.123 0.045
0.163 0.126
0.162 0.148
0.352 0.271
0.264 0.396
Pistacia lentiscus (Anacardiaceae)
0.021
0.186
0.027
0.068
0.073
0.106
0.090
0.074
0.362
0.131
0.174
0.206
0.364
Pistacia atlantica (Anacardiaceae)
0.464
0.067
0.469
0.143
0.199
0.250
0.391
0.482
0.059
0.363
0.100
0.151
0.188
0.348
0.300
0.035
0.204
0.042
0.095
0.106
0.197
0.117
0.031
0.186
0.024
0.062
0.074
0.152
0.070
0.023
0.333
0.025
0.068
0.063
0.130
0.096
Ziziphus spina-christi (Rhamnaceae) Myrtus communis (Myrtaceae)
0.032
0.737
0.026
0.078
0.102
0.239
0.080
0.031
0.150
0.032
0.066
0.073
0.167
0.087
Arbutus andrachne (Ericaceae)
0.014
0.074
0.013
0.036
0.040
0.092
0.041
Styrax officinalis (Styracaceae) Phillyrea latifolia (Oleaceae) Withania somnifera (Solanaceae) Rubia tenuifolia (Rubiaceae)
0.023
0.157
0.039
0.051
0.073
0.248
0.043
0.056
0.319
0.048
0.140
0.169
0.330
0.163
0.415
1.698
0.253
0.612
0.707
1.792
1.178
0.030
0.256
0.032
0.075
0.091
0.175
0.221
Viburnum tinus (Caprifoliaceae) Lonicera etrusca (Caprifoliaeeae)
0.044
0.238
0.062
0.114
0.125
0.238
0.136
0.028
0.228
0.038
0.095
0.104
0.236
0.091
Bryonia spp. (Cucurbitaceae) Asparagus aphyllus (Liliaceae)
0.195 0.167
1.521 0.523
0.189 0.081
0.573 0.250
0.818 0.269
0.956 0.521
0.730 1.090
Ruscus aculeatus (Liliaceae)
0.056
0.335
0.039
0.113
0.143
0.378
0.223
Smilax aspera (Liliaceae) Tamus orientalis (Dioscoreaceae) Arum dioscoridis (Araceae)
0.076
0.356
0.049
0.131
0,149
0.290
0.171
0.038
0.259
0.042
0.111
0.127
0.332
0.126
0.083
0.406
0.073
0.174
0.179
0.874
0.209
PLant species
Pistacia palaestina (Anacardiaceae) Rhus coriaria (Anacardiaceae) Rhamnus alaternus (Rhamnaceae) Rhamnus lycioides (Rhamnaceae)
Proline
be attributable to nonprotein compounds. Using the traditional 6.25 factor yielded an average of 5.75 % crude protein in the analyzed fruits while the average protein content based on the total amino acid content was only 3.90% (paired t test on arcsin square-root transformed values, T = 11.0, df = 25, P < 0.0001, Table 2). These data suggest that levels of protein in fruit pulp are even lower than previously estimated. This would lend support to the suggestion that protein
ESTIMATION OF PROTEIN IN FLESHY FRUIT
2609
TABLE 1. CONTINUED
PhenylGlycine
Alanine
Valine
Isoleucine
Leucine
Tyrosine
alanine
Histidine
Lysine
Arginine
Ammonia
0,652 0.140
0.768
1.250
0.478
0,850
0.549
2.9t0
0.226
0,430
0.302
0.793
0.135
0.147
0.110
0.194
0.095
0.127
0.064
0.116
0.193
0.045
0.174
0.204
0.216
0.163
0.279
0.144
0,161
0.083
0.238
0.181
0.073
0.127
0.128
0.145
0.107
0,166
0.117
0.116
0.081
0.128
0.209
0.061
0.067
0.072
0.088
0.057
0.103
0.057
0,063
0.033
0.085
0.070
0.048
0.202
0.220
0.222
0.176
0.301
0.188
0.205
0.116
0.275
0.246
0.104
0.255
0.271
0.298
0.220
0.356
0.289
0.245
0.147
0.378
0.374
0.080
0.193
0.201
0,234
0.151
0.247
0.186
0.170
0.094
0.238
0.195
0.082
0.109
0.114
0.121
0.101
0.173
0.097
0.117
0.045
0.136
0.104
0.076
0.085
0.076
0.083
0,069
0.119
0.053
0,070
0.034
0.056
0.064
0.122
0.077
0.064
0.093
0.050
0.117
0.057
0,065
0.032
0.068
0.066
0.051
0.100
0.128
0.097
0,072
0.143
0.077
0.075
0.046
0.109
0.541
0,096
0.094
0.084
0.083
0.063
0.117
0.069
0.071
0.039
0.086
0.079
0,041
0.042
0.044
0.050
0.037
0.057
0.024
0.034
0.016
0.037
0.041
0.043
0.061
0.102
0.069
0.034
0.080
0.039
0,048
0.032
0.051
0.056
0.041
0,172
0.161
0.178
0.155
0,247
0.135
0.156
0.052
0,032
0.166
0.080
0.709
0,755
0.663
0,584
0.866
0.573
0.559
0.389
0.911
1.213
0.326
0,141
0.107
0.108
0.084
0.122
0.077
0.084
0.037
0.059
0.089
0.047
0.151
0,129
0.140
0.110
0.197
0.094
0.127
0.050
0.141
0.131
0.093
0.112
0.110
0.110
0.107
0,159
0.087
0.103
0.033
0.052
0.108
0.064
0.654
0.723
0.841
0,663
0.868
0.596
0.719
0.208
0.576
0.892
0.264
0.422
0.321
0.327
0.268
0.395
0.293
0.277
0.115
0.265
0.299
0.103
0.168
0.182
0.157
0.128
0.150
0.109
0.111
0.069
0,155
0,263
0.111
0.240
0.184
0.168
0,129
0.214
0.216
0.130
0.071
0.161
0.191
0.094
0.146
0.189
0.143
0,110
0.178
0.114
0.116
0.044
0.084
0.179
0.070
0.241
0.338
0.276
0.169
0.285
0.149
0.217
0,073
0.114
0.279
0.078
limitation is one reason for the inadequacy of a strict fruit pulp diet (Levey and Karasov, 1989; Sedinger, 1990). The average ratio between total amino acids and total amino acid nitrogen was 6.32 (Table 2). Therefore, the amino acids present in these fruits averaged 15.8% nitrogen (100/6.32) which is close to the expected value of 16%, but since these fruits contained nonprotein nitrogen as discussed before, one should reevaluate the valid conversion factor for the Kjeldahl procedure. The 4.05
Ephedra aphylla (Ephedraceae) Osyris alba (Santalaceae) Laurus nobilis (Lauraceae) Rosa canina (Rosaceae) Crataegus monogyna (Rosaceae) Pistacia lentiscus (Anacardiaceae) Pistacia atlantiea (Anacardiaceae) Pistacia palaestina (Anacardiaceae) Rhus coriaria (Anacardiaceae) Rhamnus alaternus (Rhamnaceae)
Plant species
Estimated protein (N x 6.25) (%) 20.3 6.2 4.3 4.8 2.1 5.1 8.0 6.1 3.0 2.8
Total
nitrogen (%)
3.25 0.99 0.69 0.77 0.34 0.82 1.28
0.98 0.48 0.45
Kjeldahl determination
0.53 0.32 0.27
2.29 0.37 0.51 0.44 0.21 0.63 0.75
N + ammonia) (%)
Nitrogen (amino acid
3.5 2.0 1.4
14.5 2.4 3.4 2.9 1.3 4.0 4.9
(total amino acids) (%)
Protein
Amino acid analysis
4.2 1.7 1.6
15.3 4.2 2.8 3.2 1.1 3.4 5.7
predicted (%)"
Protein
0.7 -0.3 0.2
0.8 1.8 -0.6 0.3 -0.2 -0.6 0.8
Residuals (%)/,
Evaluation of the regression equation
6.54 6.19 5.21
6.34 6.54 6.66 6.56 6.21 6.41 6.54
acids/total amino acid N
3.59 4.14 3.19
4.47 2.43 4.96 3.77 3.57 4.94 3.84
acids/total Kjeldahl N
Total amino
Conversion factors
0.45 0.33 0.39
0.30 0.63 0.26 0.42 0.42 0.23 0.41
Nonproteinous N'
FRUIT SPECIES FROM EAST
Total amino
26 FLESHY MEOITERRANEANHABITATS(Values are Percentage Dry-Weight)
TABLE 2. ANALYSIS OF NITROGEN, PROTEIN, AND CONVERSION FACTORS OF PULPS OF
~v
4.7 6.6 5.75 _+ 4.74
1.06
0,92 + 0.76
3.3 t4.8
0.53 2.37
4.5
2.9
0.46
0.75
2.5
0.40
0.72
17.6
2.82
6.7
3.5
0.56
5.8
2.3
0.37
1.07
I. I
0. t 8
0.93
2.1
5.4
0.86
0.34
3.0
0.48
0.61 _+ 0.59
0,62
0,38
0.48
0.49
0.88
0.29 1.83
0.38
0.28
2.25
0.42
0.19
0.13
0.22
0.51
0.23
"Estimate based on Y = (4.885X) - 0.6, ~'% p r o t e i n (total a m i n o a c i d s ) - % p r o t e i n p r e d i c t e d . " ( % total K j e l d a h l N - % total a m i n o a c i d N ) / % total kjeldahl N .
Myrtus communis (Myrtaceae) Arbutus andrachne (Ericaceae) Styrax officinalis (Styracaceae) Phillyrea latifi~lia (Oleaceae) Withania somnifera (Solanaceae) Rubia tenuifolia (Rubiaceae) Viburnum tinus (Caprifoliaceae) Lonicera etrusca (Capritbliaceae) Bryonia spp. (Cucurbitaceae) Asparagus aphyllus (Liliaceae) Ruscus aculeatus (Liliaceae) Smilax a,wera (Liliaceae) Tamus orientalis (Dioscoreaceae) Arum dioscoridis (Araceae)
(Rhamnaceae)
Rhamnus lycioides (Rhamnaceae) Ziziphus spina-christi 1.5
3.90 + 3.78
4.2
2.4
3.0
2.9
6.0
1.9 12.0
2.3
1.8
14,2
2.8
1.2
0.7
1.4
2.8
1.7
4.6
3. I
2.9
3.9
4,6
2.0 I 1.0
1.6
1.4
13.2
2.1
1.2
0.3
1,1
3.6
0.4
0.7
-0. I
1.0
1,4
0.1 -0. I
-0.7
-0.4
- 1.0
- 0.7
0,0
-0.4
-0.3
0.8
0.2
6.32 _+ 0.38
6.83
6.35
6,27
5.88
6.82
6.41 6,54
6.10
6.61
6,33
6,58
6.48
5.68
6,40
5.44
6.46
4.05 -I_- 0,79
3.99
3.20
4.19
3.1 I
5.58
3.53 5.06
5.00
4,59
5,04
4.93
3.39
4.18
4,26
3.22
3.07
0.37 ! 0.11
0.42
0.50
0.33
0.47
0.18
0.45 0.23
0.18
0.31
0.20
0,25
0.48
0.26
0.33
0.4 I
0.52
t-~
,.~
~
Z
"n
Z 9
> ,--1
~,
2612
IZHAKI
average factor for all species is one approach to converting Kjeldahl total nitrogen to protein content (Table 2). This factor is slightly lower than the 4.4 factor suggested by Milton and Dintzis (1981) for tropical plants but higher than the average factor of 3.55 detected by Herbst (1986) for four tropical fruit species. A significant linear equation Y = 4.885X - 0.6 was detected between protein content and total Kjeldahl N (Figure 1). This equation may be more accurate for estimating protein content than the 4.05 factor because of the positive intercept on the x axis (Figure 1). Despite the relatively good estimation of true protein on total Kjeldahl N based on this regression, nonprotein N estimates range from less than 20% to more than 60% for this set of species. It is possible that nonprotein N may vary even more for other species. Further, analysis of residuals of the regression (Table 2) reveals that estimated total protein from total N can be relatively inaccurate for particular species. For instance, protein is underestimated by 133% for Arbutus andrachne and overestimated by 43 % for Osyris alba upon applying the regression equation (Table 2). Typical errors for other species tend to be in the 10-30% range. In addition, the proportional error tends to be greater with decreasing total N; nevertheless, most fruit pulps tend to be low in N. Thus, the regression of total protein on total N would seem to be adequate only for broad-scale estimates where positive and negative errors would tend to cancel. Such a case is the estimation of total protein intake for animals on a mixed-fruit diet. In contrast, any estimate of protein content derived from total N would be inadequate when accurate interspecific comparisons are 21 18
Z LLI
I5 +
I0
12
o_
9
,( I0 I-
6
+
+
+
3 0 0.0
J
i
i
i
p
i
i
i
0.5
1.0
1.5
2.0
2.5
3.0
3.5
KJELDAHL +
AMINO ACIDS
TOTAL ,~
4.0
N (%)
KJELDAHL
FIG. 1. The crude total protein content (calculated by multiplying total N by the traditional 6.25 factor) and the true protein content (using the total amino acid content) of 26 fruit species plotted against Kjeldahl total N. The linear equation for calculating true protein as a function of Kjeldahl total N is Y = 4.885X - 0.6, (R2 = 0.96, P < 0.0001).
E S T I M A T I O N OF PROTEIN IN F L E S H Y F R U I T
2613
needed. Thus, while the use of the regression is indeed better than the use of the 6.25 conversion factor, the data suggest that it may be risky to use total N to estimate protein in fruit pulps, especially of those species with unusually high or low levels of nonprotein N. The relatively large differences in nitrogen content estimated by the Kjeldahl procedure and the one calculated by the amino acid analysis (Table 2) were probably not an artifact effect of these methods. As discussed earlier by Milton and Dintzis (1981), it is possible that the HC1 hydrolysis in the amino acid analysis does not release all the nitrogen from compounds that are tightly bonded as plant-cell material, while the sulfuric acid in the Kjeldahl method releases this nitrogen. However, most plant protein is found within the cytoplasm of the cell, with only small amounts associated with the cell wall (Lyttleton, 1973; Albersheim, 1975). Further, if such nitrogen exists, it is probably not digestible by animals. Therefore, it seems that most of the differences in nitrogen content between the two methods is due to the high levels of nonprotein nitrogen in wild fruits. There is little information on the nonprotein nitrogen compounds in fruits. The presence of alkaloids and glycosides is frequent among European and Mediterranean species even in ripe fruits (Jordano, 1988; Ehrl6n and Eriksson, 1993, and references therein) and is also reported in other ecosystems (e.g., Calvert, 1985; Potter and Kimmerer, 1986; Gargiullo and Stiles, 1991). Therefore, these compounds may constitute a major fraction of the nonprotein nitrogen compounds in fleshy fruits. In order to confirm this assumption, a direct analysis of secondary compounds in these fruits should be carried out. The significance of such compounds to the evolution of mutual relations between plants that produce fleshy fruits and their frugivorous seed dispersers is still not clear (see Jordano, 1991, and references therein). Acknowledgments--I wish to thank Y. Pancheshnikovafor assisting in the lab, Dr. E. Bamberger for helping during different stages of this study, and an anonymousreviewer for very helpful suggestions and comments that improved the manuscript. Kjeldahl analyses were performed in Miloda Laboratories under D. Izhar's supervision. Amino acid analyses were performedat Aminolab Ltd. under Z. Harduf's supervision. This study was partly supported by a joint research grant of the Technion and University of Haifa, by the Carmel Foundation-Ministryof Environmental Affairs, and by a grant from the Joint German Israeli Research Program. The idea for this study was an outcome of a fruitful personal communicationwith Carlos M. Herrera.
REFERENCES AOAC. 1984. Official Methods of Analysis of the Association of Official Analytical Chemists. AOAC, Washington, D.C. ALBERSHEIM,P. 1975. The walls of growing plant cells. Sci. Am. 232:80-95. BARNEA,A., YOMTOV,Y., and FRIEDMAN,J. 1991. Does ingestion by birds affect seed germination? Funct. Ecol. 5:394-402.
2614
IZHAKI
BERTI-IOLD, P.B. 1976. The control and significance of animal and vegetable nutrition in omnivorous songbirds. Ardea 64:140-154. BONDI, A.A. 1987. Animal Nutrition. John Wiley & Sons, New York. CALVERT, J.J. 1985. Food selection by western gorillas ( G.g. gorilla) in relation to food chemistry. Oecologia 65:236-246. DEBUSSCHE, M., CORTEZ, J., and RIMBAULT,I. 1987. Variation in fleshy fruit composition in the Mediterranean region: The importance of ripening season, life-form, fruit type and geographical distribution. Oikos 49:244-252. EHRLt~N, J., and ERIKSSON, O. 1993. Toxicity in fleshy fruits--a non-adaptive trait? Oikos 66:107113. ELKIN, R.G., and GRJFFITH, J.E. 1985. Hydrolysate preparation for analysis of amino acids in sorghum grains: Effect of oxidative pretreatment. J. Assoc. Anal. Chem. 68:1117-1121. FOSTER, M.F. 1978. Total frugivory in tropical passerines: A reappraisal. Trop. Ecol. 19:131-154. GARGIULLO,M.B., and ST~LES, E.W. 1991. Chemical and nutritional differences between two birddispersed fruits: flex opaca and Rex verticillata. J. Chem. Ecol. 17:1091-1106. HANSEN, E. 1970. Proteins, in A.C. Hulme (ed.). The Biochemistry of Fruits and Their Products, Vol. l, Series on Food, Science & Technology. Academic Press, New York. HARBORNE, J.B. 1991. The chemical basis of plant defense, pp. 45-59, in R.T. Palo and C.T. Robbins (eds.). Plant Defenses Against Mammalian Herbivory. CRC Press, Boca Raton, Florida. HERBST, L.H. 1986. The role of nitrogen from fruit pulp in the nutrition of the frugivorous bat Carollia perspicillata. Biotropica 18:39-44. HERRERA, C.M. 1987. Vertebrate-dispersed plants of the Iberian Peninsula: A study of fruit characteristics. Ecol. Monogr. 57:305-331. HERRERA,C.M. 1989. Frugivory and seed dispersal by carnivorous mammals, and associated fruit characteristics, in undisturbed Mediterranean habitats. Oikos 55:250-262. IZHAKI, I. 1986. Seed dispersal by birds in an eastern Mediterranean scrublands. PhD dissertation. The Hebrew University of Jerusalem, Israel (in Hebrew). IZHAKI, I. 1992. A comparative analysis of nutritional quality of mixed and exclusive fruit diets for yellow-vented bulbuls. Condor 94:912-923. IZHAKI, I., and SAFRIEL,U.N. 1985. Why do fleshy-fruit plants of the Mediterranean scrub intercept fall--but not spring--passage of seed-dispersing migratory birds? Oecologia 67:40-43. IZHAKI, I., and SAFRIEL~U.N. 1989. Why are there so few exclusive frugivorous birds? Experiments on fruit digestibility. Oikos 54:23-32. IZHAKI, I., WALTON,P., and SAFRIEL,U.N. 1991. Seed shadows generated by frugivorous birds in an eastern Mediterranean scrub. J. Ecol. 79:579-590. JORDANO, P. 1988. Diet, fruit choice and variation in body condition of frugivorous warblers in Mediterranean scrubland. Ardea 76:193-209. JORDANO, P. 1991. Fruits and frugivory, pp. 105-156, in M. Fenner (ed.). Seeds: The Ecology of Regeneration in Plant Communities. C.A.B. International, Willingford. KOOL, K.M. 1992. Food selection by the silver leaf monkey, Trachypithecus auratus sondaicus, in relation to plant chemistry. Oecologia 90:527-533. LEVEY, D.J., and KARASOV,W.H. 1989. Digestive responses of temperate birds switched to fruit or insect diets. Auk 106:675-686. LYTTLETON, J.W. 1973. Proteins and nucleic acids, pp. 63-103, in G.W. Butler and R.W. Bailey (eds.). Chemistry and Biochemistry of Herbage, Vol. 1. Academic Press, New York. MACK, A.L. 1990. Is frugivory limited by secondary compounds in fruits? Oikos 57:135-138. MAYNARD, A.B., and LOOSLI, J.K. 1969. Animal Nutrition. McGraw-Hill, New York. MILTON, K., and DINTZlS, F.R. 1981. Nitrogen-to-protein conversion factors for Tropical plant samples. Biotropica 13:177-181.
ESTIMATION OF PROTEIN IN FLESHY FRUIT
2615
PARRISH, J.W., and MARTIN, E.W. 1977. The effect of dietary lysine on the energy and nitrogen balance of the dark-eyed junco. Condor 79:24-30. P~PER, J.K. 1986. Seasonality of fruit characters and seed removal by birds. Oikos 46:303-310. PODDAR, S., and LEDERER, R.J. 1982. Juniper berries as an exclusive winter forage for Townsend's solitaires. Am. Midl. Nat. 108:34-40. POTTER, D.A., and KIMMERER, T.W. 1986. Seasonal allocation of defense investment in llex opaca Ait. and constraints on a specialist leaf miner. Oecologia 69:217-224. ROBBINS, C.T. 1983. Wildlife Feeding and Nutrition. Academic Press, New York. ROGERS, M.F., MA~SELS, F., W~LUAMSON, E.A., FERNANDEZ, M., and TUTIN, C.E.G. 1990. Gorilla diet in the Lope Reserve, Gabon. Oecologia 84:326-339. SAKAI, H.F., and CARPENTER, J.R. 1990. The variety and nutritional value of foods consumed by Hawaiian crow nestlings, and endangered species. Condor 92:220-228. SEDINGER, J.S. 1990. Are plant secondary compounds responsible for negative apparent metabolizability of fruits by passerine birds? A comment on Izhaki and Safriel. Oikos 57:138-140.
