- Email: [email protected]
Using microbial fatty acids to improve understanding of the contribution of solid associated bacteria to microbial mass in the rumen
Animal Feed Science and Technology 150 (2009) 197–206
Contents lists available at ScienceDirect
Animal Feed Science and Technology journal homepage: www.elsevier.com/locate/anifeedsci
Using microbial fatty acids to improve understanding of the contribution of solid associated bacteria to microbial mass in the rumen R.J.B. Bessa a,∗, M.R.G. Maia a, E. Jerónimo a, A.T. Belo b, A.R.J. Cabrita c, R.J. Dewhurst d, A.J.M. Fonseca e a
REQUIMTE, Unidade de Produc¸ão Animal, Instituto Nacional de Recursos Biológicos, Fonte Boa, 2005-048 Vale de Santarém, Portugal b Unidade de Produc¸ão Animal, Instituto Nacional de Recursos Biológicos, Fonte Boa, 2005-048 Vale de Santarém, Portugal c REQUIMTE, SAECA, Faculdade de Ciências, Universidade do Porto, Campus Agrário de Vairão, Rua Padre Armando Quintas, 4485-661 Vairão VC, Portugal d Teagasc, Animal Bioscience Centre, Dunsany, County Meath, Ireland e REQUIMTE, ICBAS, Instituto de Ciências Biomédicas de Abel Salazar, Universidade do Porto, Campus Agrário de Vairão, Rua Padre Armando Quintas, 4485-661 Vairão VC, Portugal
a r t i c l e
i n f o
Article history: Received 17 October 2007 Received in revised form 19 September 2008 Accepted 24 September 2008 Keywords: Microbial markers Odd and branched chain fatty acids Purine bases Rumen bacteria
a b s t r a c t This study sought to distinguish liquid-(LAB) and detached (SAB1 ) and undetached (SAB2 ) solid-associated bacteria through their fatty acid (FA) and purine base (PB) profiles. Fatty acids and PB were also evaluated as internal microbial markers for estimating microbial biomass associated with rumen particles. Four merino rams fitted with rumen cannulae and fed dehydrated alfalfa pellets provided rumen contents. In 3 consecutive weeks, rumen contents were collected and samples of LAB and SAB1 , total rumen content (TRC), washed rumen particles (WRP) and rumen particles after SAB1 extraction (ERP) were obtained and analysed for PB and FA. The SAB2 biomass composition was estimated from the non-NDF organic matter (OM) remaining in ERP. The concentration of total SAB biomass in particles was estimated using both PB and odd and branched-chain fatty acids (OBCFA). Concentrations of PB and
Abbreviations: CP, crude protein; ERP, extracted rumen particles; FA, fatty acids; LAB, liquid associated bacteria; NAN, nonammonia N; NDFom, neutral detergent fibre not assayed with a heat stable amylase and expressed exclusive of residual ash; OBCFA, odd and branched-chain fatty acids; OM, organic matter; PB, purine bases; SAB1 , detachable solid associated bacteria; SAB2 , undetachable solid associated bacteria; TRC, total rumen content; WRP, washed rumen particles. ∗ Corresponding author. Tel.: +351 243 767308; fax: +351 243 767307. E-mail address: [email protected] (R.J.B. Bessa). 0377-8401/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.anifeedsci.2008.09.005
198
R.J.B. Bessa et al. / Animal Feed Science and Technology 150 (2009) 197–206
OBCFA were highly correlated among the different rumen fractions. Marked differences between LAB and SAB populations occurred with LAB having higher PB content, lower FA content and a higher proportion (g/100 g fatty acids) of OBCFA than did SAB. The chemical composition of SAB1 and SAB2 was similar, except for the 15% higher crude protein content of the latter. The concentration of OBCFA (mg/g microbial OM) did not differ between bacterial fractions. The PB/OBCFA ratio (mg/mg) was higher in LAB (2.08) than in SAB (0.94). The ratio between branched-chain and odd-linear-chain FA was higher in LAB (2.26) than in SAB (1.46). Extraction of PB and OBCFA from WRP with our SAB detachment procedure was 61% and 31%, respectively. Estimated SAB1 and total SAB biomass (mg OM/g WRP) were 158 and 266, and 47 and 164, respectively, using PB and OBCFA as microbial markers. This study suggests that the OBCFA have potential as internal microbial markers in rumen ecosystem studies. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Internal microbial markers have been extensively used to predict the contribution of microbial crude protein (CP) to duodenal non-ammonia N (NAN) flows (Broderick and Merchen, 1992) and to study the distribution of microbes within the rumen ecosystem (Legay-Carmier and Bauchart, 1989). Several substances have been proposed as internal markers, including diaminopimelic acid, d-alanine, 2-aminoethylphosphonic acid, chitin, adenylate energy charge, purines and ribonucleic acid. Purine bases (PB) were regarded as one of the best methods by Broderick and Merchen (1992). However, because no internal marker has proven completely satisfactory (Broderick and Merchen, 1992), the search for alternative internal markers continues. Additionally, no marker has yet addressed the problem of obtaining representative samples of populations from either the rumen or the duodenum. The techniques used to remove attached bacteria (i.e., blending, pummelling, chilling, chemical treatment or a combination of treatments) vary in their efficacy (Martín-Orúe et al., 1998). Therefore, there is still a need to clarify the composition of undetached microbes associated with rumen particles, particularly in calculating microbial flows. Microbial fatty acid (FA) composition has been extensively used in systematic (Moss, 1981; Kaneda, 1991) and ecologic studies (Olsson, 1999). Recently, emphasis has been given to use of odd and branched chain fatty acids (OBCFA) in rumen studies (Vlaeminck et al., 2006a). These FA apparently fulfill most of the requirements needed for an internal marker of rumen microbial biomass as they are stable compounds, easy to measure, and are present at only trace levels in most plants (Diedrich and Henschel, 1990). Additionally, their pattern can vary between both liquid-(LAB) and solid-associated bacteria (SAB) (Vlaeminck et al., 2006b), as well among microbial species (Vlaeminck et al., 2006a). This study aimed to distinguish LAB, detachable (SAB1 ) and undetached (SAB2 ) microbes associated with rumen particles through their FA and PB profiles. Fatty acids and PB were also evaluated as internal microbial markers to estimate microbial biomass associated with rumen particles. 2. Materials and methods 2.1. Collection and fractionation of rumen contents Four merino rams (50 ± 3.2 kg) fitted with rumen cannulae (6 cm diameter) were used to provide rumen contents. Rams were fed 1 kg of dehydrated alfalfa pellets (CP: 177 g/kg DM; NDFom: 458 g/kg DM) and 0.1 kg of wheat straw (CP: 36 g/kg DM; NDFom: 700 g/kg DM) in two equal meals at 10:00 and 17:00 h. Rams had continuous access to water and mineral blocks. After 3-week adaptation periods, 1 l of rumen contents was collected from each ram before the morning meal of the same day of 3 consecutive weeks. The procedures used to fractionate rumen contents and to obtain bacterial pellets
R.J.B. Bessa et al. / Animal Feed Science and Technology 150 (2009) 197–206
199
Fig. 1. Procedures used for fractionation of rumen contents and to obtain rumen bacteria pellets.
are summarized in Fig. 1. At each collection, one aliquot of total rumen contents (TRC) from each ram was freeze-dried for subsequent analysis. The remaining TRC was filtered through eight layers of surgical gauze; the solid phase was washed, re-suspended in twice its weight of saline solution (0.85 g/100 ml, w/v NaCl, 39 ◦ C) and filtered again through eight layers of surgical gauze. The liquid phase of TRC and the washing liquid were pooled and subjected to differential centrifugation at 500 × g for 5 min, at 4 ◦ C. The supernatant was then centrifuged at 20,000 × g for 20 min, at 4 ◦ C, to obtain the LAB pellet, which was washed with saline solution and re-centrifuged. An aliquot of the washed solid phase of TRC (i.e., washed rumen particles, WRP) was freeze-dried and the remaining material was
200
R.J.B. Bessa et al. / Animal Feed Science and Technology 150 (2009) 197–206
subjected to SAB extraction following the procedure described by Martín-Orúe et al. (1998). Briefly, WRP were re-suspended in saline solution overnight at 4 ◦ C, homogenized 6 × 30 s with a Waring blender (Waring Products Division, New Hartford, CT, USA) and filtered through surgical gauze. The solid phase obtained was washed and freeze-dried (i.e., extracted rumen particles, ERP). The liquid phase, including the washing liquid, was subjected to differential centrifugation, as described above, to obtain the detachable SAB pellet (SAB1 ). All samples were freeze-dried and kept at −20 ◦ C until analysis. 2.2. Analytical procedures Samples of TRC, WRP and ERP were analysed for DM, ash (AOAC, 1990, ID 942.05), Kjeldahl N (AOAC, 1990, ID 954.01) and neutral detergent fibre (NDFom; Van Soest et al., 1991). Sodium sulfite was not added during NDF extraction and ␣-amylase was also not used, and NDF was expressed exclusive of residual ash. Kjeldahl N was also determined on the NDFom residue from ERP (N-NDFom). The protein content of bacterial samples (LAB and SAB1 ) was determined using the Bradford method (Bradford, 1976), using bovine serum albumin as standard (Merck KGaA, Darmstadt, Germany). The TRC, WRP, ERP, LAB and SAB1 samples were analysed for PB and FA contents. PB were analysed by HPLC after perchloric acid hydrolysis following Balcells et al. (1998). Separation was according to the method of Makkar and Becker (1999) at a flow rate of 1.0 ml/min and PB were detected using a UV detector at a wavelength of 254 nm. Fatty acid analysis was conducted by the one-step extraction–methylation method of Sukhija and Palmquist (1988). Fatty acid methyl esters were analysed on a HP6890 gas chromatograph (Hewlett-Packard, Avondale, PA, USA) operating with a flame ionization detector and equipped with a 60 m fused silica capillary column SP-2380 (Supelco, Bellefonte, PA, USA) with 0.25 mm of internal diameter and 0.20 m of film thickness, and using He at 2.0 ml/min as a carrier gas. Peak identification was based on comparison with known standards of FA methyl esters (Sigma Chemical Co., St. Louis, MO, USA). The octadecenoic isomers were not completely resolved and are presented as trans-18:1 (i.e., sum of the peaks eluting between 18:0 and cis-9 18:1), and cis-18:1 (sum of cis-9, cis-11 and cis-12 isomers). The conjugated linoleic acid (CLA) peak reported is consistent with retention time of 18:2 cis-9, trans-11 methyl ester, but could also include several other conjugated isomers, such as 18:2 trans-7, cis-9 and 18:2 trans-8, cis-10. Quantification was made using heneicosanoic acid (21:0) as the internal standard, assuming for all peaks equal proportionality between peak areas and FA methyl ester weight. 2.3. Calculations Consider ‘D’ as the proportion of non-NDF OM that is removed from WRP OM (WRPOM ) due to SAB extraction, calculated as D=
([ERP]NDF − [WRP]NDF ) [ERP]NDF
where: [ERP]NDF = concentration of NDF (mg/g OM) in ERP, and [WRP]NDF = concentration of NDF (mg/g OM) in WRP. The contribution of SABOM to WRPOM can be estimated as SAB1 (mg/g WRPOM ) =
[WRP]m − ([ERP]m × (1 − D)) × 1000 [SAB1 ]m
SAB2 (mg/g WRPOM ) =
[ERP]m × (1 − D) × 1000 [SAB2 ]m
SABT (mg/g WRPOM ) = SAB1 + SAB2 where: SAB1 = bacterial OM biomass extracted from WRPOM ; SAB2 = bacterial OM biomass remaining in the ERP OM; [WRP]m = concentration (mg/g OM) of marker in WRP; and [ERP]m = concentration (mg/g OM) of marker in ERP; [SAB1 ]m = concentration (mg/g OM) of marker in SAB1 ; [SAB2 ]m = concentration (mg/g OM) of marker in SAB2 .
R.J.B. Bessa et al. / Animal Feed Science and Technology 150 (2009) 197–206
201
The protein content of SAB2 was computed as 6.25 × ([ERP]N − ([ERP]N-NDF + [ERP]N-NA )) where: [ERP]N = concentration of N (Kjeldahl) (mg/g OM) in ERP; [ERP]N-NDF = concentration of N (Kjeldahl) (mg/g) in NDF from ERP; [ERP]N-NA = concentration of N present in nucleic acids computed from concentrations of adenine and guanine (mg/g OM) in ERP and complementary pyrimidine bases. The percentage of marker extraction was calculated as Marker extraction (%) =
[WRP]m − ([ERP]m × (1 − D)) . [WRP]m × 100
2.4. Statistical analysis Data were analysed by repeated measures analysis of variance using the MIXED procedure of SAS (2001) with week of sampling as the repeated measure. The compound symmetry covariance structure was used. The model included the fixed effects of rumen fractions, bacterial type or type of marker and week of sampling, the random effect of animal and the random residual error. Orthogonal contrasts were constructed in order to compare: (i) total rumen contents with rumen particles (TRC vs (WRP and ERP)), (ii) washed with extracted rumen particles (WRP vs ERP), (iii) LAB with SAB (LAB vs (SAB1 and SAB2 )), and (iv) detachable and undetachable SAB (SAB1 vs SAB2 ). Correlations between PB and FA concentrations in TRC, WRP and ERP were tested and the Pearson correlation coefficients reported. 3. Results 3.1. Composition of rumen contents PB, N, and FA contents were highest in TRC and lowest in ERP (Table 1). The NDFom content of these fractions followed the opposite trend, increasing from TRC through to ERP. Total FA concentration (mg/g OM) decreased from TRC through to ERP, although the proportions (g/100 g total fatty acids) of iso-14:0, 14:0, iso-15:0, 15:0, iso-16:0, 17:0, cis-18:1, and 20:0 did not change with rumen fractionation (Table 1). The anteiso-15:0, iso-17:0, anteiso-17:0, and the sum of the residual FA (most of them not identified) were lower in rumen particles (WRP and ERP) than in TRC. The proportions of 16:0, 18:0, trans-18:1, 18:2n-6, 18:3n-3, and 22:0 were higher in rumen particles (WRP and ERP) than in TRC. The proportion of iso-15:0, trans-18:1 and 22:0 were higher in ERP than WRP. Total FA and all individual FA concentrations (mg/g OM) in TRC, WRP and ERP were positively correlated with PB concentration (data not shown). The highest Pearson correlation coefficients with PB concentration were for 16:0 (0.91), total FA (0.86) and anteiso-15:0 (0.83), the correlation for total OBCFA being 0.79. Although 16:0 and total FA had higher correlations with PB than OBCFA, they can be quite unspecific. Therefore, OBCFA were used to estimate microbial biomass associated with rumen particles instead of individual FA. 3.2. Composition of rumen bacteria Table 2 presents PB, protein and total FA contents and proportions of FA for rumen bacteria. There were differences between LAB and SAB populations, LAB having higher PB content and PB/protein ratio and lower FA content. The PB and FA contents of SAB1 and SAB2 were similar, though SAB2 contained 15% more protein, and consequently had a lower PB/protein than SAB1 . Fatty acid proportions were different between LAB and SAB; LAB having a higher proportion (g/100 g total FA) of OBCFA, and SAB having a higher proportion of C18 biohydrogenation derived FA (18:0; trans-18:1 and CLA). The CLA was much higher in SAB than in LAB; it was hardly present in LAB. The concentration of OBCFA (mg/g microbial OM) did not differ between bacterial types. The ratio between branched-chain and odd linear-chain FA was higher in LAB than in SAB.
202
R.J.B. Bessa et al. / Animal Feed Science and Technology 150 (2009) 197–206
Table 1 Chemical composition (mg/g OM) and fatty acid (FA) of total rumen contents (TRC), washed ruminal particles (WRP), and extracted ruminal particles (ERP). TRC
WRP
ERP
SEM
TRC vs RP
WRP vs ERP
Concentrations (mg/g OM) Purine bases (PB) N NDF FA OBCFAa
3.47 32.9 586 30.4 2.36
2.19 23.9 733 20.2 1.40
1.01 18.3 872 15.8 1.13
0.223 1.05 11.9 0.65 0.039
<0.001 <0.001 <0.001 <0.001 <0.001
<0.001 <0.001 <0.001 0.003 0.005
Ratios PB/OBCFA Branched-/odd-chain FA
1.48 1.56
1.58 1.31
0.90 1.41
0.134 0.073
0.017 0.017
<0.001 0.204
0.32 1.25 1.03 1.37 2.26 0.62 19.0 0.85 0.59 0.79 29.3 9.70 5.47 4.26 3.49 0.91 0.79 18.0
0.32 1.18 0.97 1.14 2.25 0.58 19.7 0.56 0.51 0.77 30.4 10.90 5.44 5.23 3.94 0.98 0.90 14.3
0.20 1.20 1.21 1.14 2.20 0.67 19.9 0.71 0.45 0.78 31.2 11.60 5.60 4.88 3.77 0.95 1.04 12.8
0.056 0.042 0.069 0.035 0.046 0.058 0.58 0.056 0.025 0.038 0.91 0.530 0.550 0.300 0.150 0.067 0.037 0.98
0.136 0.123 0.487 0.002 0.403 0.941 0.044 0.020 0.002 0.391 0.047 <0.001 0.839 0.025 0.040 0.379 0.001 0.007
0.094 0.712 0.049 0.993 0.431 0.367 0.744 0.103 0.070 0.579 0.352 0.040 0.523 0.300 0.332 0.702 0.008 0.275
FA proportions, g/100 g total FA iso-14:0 14:0 iso-15:0 anteiso-15:0 15:0 iso-16:0 16:0 iso-17:0 anteiso-17:0 17:0 18:0 trans-18:1 cis-18:1 18:2n-6 18:3n-3 20:0 22:0 Othersb a b
Odd and branched-chain fatty acids. The remaining fatty acids, most of them unidentified.
Differences in FA proportions between SAB1 and SAB2 were observed. Compared to SAB1 , the SAB2 had lower proportions of odd-chain FA and some of the branched-chain FA (iso-14:0 and iso-16 and anteiso-17:0), but the ratio between branched-chain and odd linear-chain FA was equal. Moreover, the SAB2 had higher proportions of trans-18:1, CLA and particularly of 18:2n-6 and 18:3n-3 than SAB1 . 3.3. Marker extraction and SAB biomass estimation The extraction of PB and OBCFA from WRP with the current SAB detachment procedure (Table 3) was 61% and 31%, respectively. Detached SAB biomass (SAB1 ) and total SAB biomass estimates were higher for PB than for OBCFA. The undetached SAB (SAB2 ) biomass estimate did not differ between markers. Estimates of total SAB biomass in particles was 27% and 16% of WRPOM , and comprise 99% and 62% of non-NDFOM in WRP, respectively for PB and OBCFA. 4. Discussion 4.1. Composition of rumen contents In this experiment, the contribution of dietary PB can be considered negligible, as the mean PB concentration in WRP was 2190 mg/kg OM in comparison with 7.6 mg/kg OM in dehydrated alfalfa. Additionally, Djouvinov et al. (1998) found that 61% of dietary PB were degraded after 17 h in the rumen. Although the distribution of markers between liquid and solid phases was not measured,
R.J.B. Bessa et al. / Animal Feed Science and Technology 150 (2009) 197–206
203
Table 2 Chemical composition (mg/g organic matter, OM) and fatty acid (FA) (g/100 g total fatty acids) in liquid-(LAB) and solid associated bacteria (SAB).
Concentrations (mg/g OM) PB Protein Fatty acids (FA) OBCFAa
LAB
SAB1
SAB2
SEM
LAB vs SAB
SAB1 vs SAB2
20.9 393 61.0 10.2
9.7 450 113.9 10.3
8.0 520 126.6 9.3
0.85 17.3 4.30 0.75
<0.001 0.001 <0.001 0.490
0.161 0.031 0.084 0.225
Ratios PB/protein (mg/g) OBCFA/protein (mg/g) PB/OBCFA Branched-/odd linear FA
54.5 26.4 2.08 2.27
20.6 23.2 0.94 1.46
14.9 17.9 0.89 1.45
2.36 1.55 0.122 0.081
<0.001 0.002 <0.001 <0.001
0.083 0.013 0.588 0.943
FA proportions, g/100 g total FA iso-14:0 14:0 iso-15:0 anteiso-15:0 15:0 iso-16:0 16:0 iso-17:0 anteiso-17:0 17:0 18:0 trans-18:1 cis-18:1 18:2n-6 CLA 18:3n-3 20:0 22:0 Othersb OBCFA
0.72 1.03 2.48 4.98 4.19 1.06 19.9 1.93 0.93 1.06 23.3 5.69 4.96 2.84 0.07 1.47 0.65 0.50 22.3 17.3
0.39 0.67 1.02 1.56 2.67 0.88 20.9 0.71 0.78 1.00 30.8 9.85 5.06 3.01 0.47 1.72 0.71 0.60 17.2 9.02
0.16 1.19 1.21 1.14 2.20 0.67 19.9 0.71 0.45 0.78 31.2 11.56 5.60 4.87 0.66 3.76 0.95 1.04 12.0 7.32
0.069 0.068 0.134 0.333 0.165 0.089 0.542 0.130 0.033 0.043 1.03 0.490 0.439 0.147 0.048 0.075 0.090 0.064 0.78 0.792
<0.001 0.152 <0.001 <0.001 <0.001 0.016 0.489 <0.001 <0.001 0.002 0.002 <0.001 0.153 0.002 <0.001 <0.001 0.137 0.006 <0.001 <0.001
0.016 <0.001 0.296 0.362 0.042 0.077 0.186 0.981 <0.001 0.001 0.825 0.008 0.084 <0.001 0.006 <0.001 0.101 0.003 0.005 0.133
SAB1 : detachable bacterial OM biomass; SAB2 : undetachable bacterial OM biomass. a b
Odd and branched-chain fatty acids. The remaining fatty acids, most of them unidentified.
the higher PB and FA contents of TRC in comparison with WRP was likely due to the presence of liquid phase biomass (e.g., LAB and protozoa), biomass associated with very small highly colonized particles (Legay-Carmier and Bauchart, 1989; Yang et al., 2001), and eventually fatty acid soaps in suspension.
Table 3 Percentage extraction of marker from rumen particles and microbial biomass concentration in washed rumen particle organic matter (WRPOM ) estimated using either purine bases (PB) or total odd and branched-chain fatty acids (OBCFA) as microbial markers. PB Marker extraction from WRP (%) Biomass (mg/g WRPOM ) SAB1 SAB2 SABT SABT (% non-NDFOM in WRP)
61.0 158 108 266 99
OBCFA 30.5 47 117 164 62
SEM 3.99 22.5 8.3 20.4 4.8
SAB1 : detachable bacterial OM biomass; SAB2 : undetachable bacterial OM biomass; SABT : SAB1 + SAB2 .
P 0.002 0.010 0.391 0.019 0.021
204
R.J.B. Bessa et al. / Animal Feed Science and Technology 150 (2009) 197–206
4.2. Composition of rumen bacteria The present data on chemical composition of rumen bacteria is in general accordance with published information. The higher PB content and PB/protein ratio and lower FA content of LAB compared to SAB populations are in accordance with previous reports (Merry and McAllan, 1983; Olubobokun and Craig, 1990; Martin et al., 1994; Vicente et al., 2004). These differences might be related to distinct distributions of bacterial species between liquid and particulate phases (Minato et al., 1966) as well to differences in metabolic activities (Merry and McAllan, 1983; Bauchart et al., 1990) and growth rate (Bates et al., 1985). Concentrations of PB in SAB1 agree with some previous reports (Volden et al., 1999; Rodriguez et al., 2000), but are lower than those reported by others using a similar analytical method (Perez et al., 1998; Gonzaléz-Ronquilho et al., 2004; Vicente et al., 2004). The higher C18 biohydrogenation derived FA (18:0, trans 18:1, and CLA) in SAB than in LAB agree with previous studies (Bauchart et al., 1990; Vlaeminck et al., 2006b) and could have been related to higher biohydrogenation activity in rumen particles (Harfoot et al., 1975). The CLA concentration was much higher in SAB than in LAB, and it was hardly present in LAB. Reports on CLA concentration in LAB and SAB are scarce, but Vlaeminck et al. (2006b) found higher CLA proportions in SAB than in LAB, whereas this pattern was only partly reflected in data of Kim et al. (2005). The ratio between branched-chain and odd linear-chain FA was higher in LAB than in SAB, as found by Vlaeminck et al. (2006b) when collating published data. The present data are the first estimate of chemical composition of undetached microbes associated with rumen particles (SAB2 ). The chemical composition of SAB1 and SAB2 was similar, except for the 15% higher protein content of the latter. This difference could be due to real differences in chemical composition, to presence of non-microbial N in non-NDFOM remaining in ERP (SAB2 ) or/and to different methodology used for quantification of the different bacterial fractions. The protein content of SAB2 was estimated using a factorial approach where N associated with NDFom and N associated with nucleic acids (averaging 37% and 4% of ERP total N, respectively) were subtracted from total N in ERP. Moreover, the SAB2 had higher proportions of trans-18:1, CLA and particularly of 18:2n-6 and 18:3n-3 than SAB1 . The higher concentration of 18:2n-6 (6.17 vs 3.42 mg/g non-NDFOM ) and of 18:3n-3 (4.77 vs 1.92 mg/g non-NDF OM ) in SAB2 than in SAB1 suggests that plant structural membranes (e.g., from chloroplasts) might still be present in ERP. As far as we know, there are no simultaneous published values for total PB, N and OBCFA (and hence direct PB/OBCFA ratios) of rumen bacteria, but Vlaeminck et al. (2005) found a slightly higher PB/OBCFA ratio in SAB (1.17) in available literature. Both TRC and WRC had an intermediate PB/OBCFA ratio between LAB and SAB. Both PB and OBCFA are expected to be of microbial origin, and that the PB/OBCFA and branched-chain and odd linear-chain FA ratios did not differ between SAB1 and SAB2 suggests that the biomass composition of detached SAB (SAB1 ) and SAB remaining adherent in ERP (SAB2 ) is quite similar, and that the differences between SAB1 and SAB2 resulted from the presence of an N rich non-NDFOM fraction of plant (i.e., feed) origin. Indeed, we can estimate that microbial N should only be 73% of N present in SAB2 fraction when using both PB/OBCFA and PB/protein ratios. Unlike the PB/protein ratio, the OBCFA/protein ratio did not differ between LAB and SAB1 (P=0.12; not shown in Table 2), which is an advantage for use of OBCFA as an internal marker of microbial mass flowing from the rumen. 4.3. Marker extraction and SAB biomass estimation There were large differences between PB and OBCFA extraction from WRP (61% vs 31%), that are reflected in the large difference in SAB1 estimates (158 vs 47 mg/g WRPOM ) obtained with each marker. Marker extraction is computed as the difference between marker concentration in WRP and ERP (corrected for losses of non-NDFOM ) and might reflect PB leakage and solubilisation associated with bacterial cell lysis and damage during extraction. Martín-Orúe et al. (1998) also reported high PB extraction from WRP with their SAB detachment procedures (60–71%). Conversely, extraction of OBCFA was close to at reported for diaminopimelic acid (Legay-Carmier and Bauchart, 1989) which, like OBCFA is not contained in the cytoplasm, suggesting that the low extractions observed might be due to persistence of the bacterial cell wall and membranes adherent to particles even after cell damage or lysis. As
R.J.B. Bessa et al. / Animal Feed Science and Technology 150 (2009) 197–206
205
far as we know, there is no previous report on the level of extraction of OBCFA with the SAB detachment procedure. The proportion of non-NDFOM that disappeared from WRP with the current SAB detachment procedure was 0.159 (S.E. = 0.014). However, the PB/OBCFA ratio of the non-NDFOM that disappeared was 3.20 (S.E. = 0.434), much higher than the PB/OBCFA ratio of SAB1 (0.94), SAB2 (0.89), WRP (1.58) and ERP (0.90). The high value of PB/OBCFA ratio in the extracted material is consistent with the much higher PB extraction than from OBCFA. The SAB1 biomass estimates expressed as non-NDFOM that disappeared from WRP was 99% and 32% for PB and OBCFA, respectively. In three of 12 observations, the SAB biomass estimates using PB as marker produced an obvious overestimation (176%, 129% and 111% of WRP non-NDFOM ) and if these values were excluded the average is reduced to 78% (S.D. = 12.6). Therefore, the estimates of SAB1 biomass in WRP using PB indicate that SAB1 constitutes most of the non-NDFOM that disappeared from WRP with this SAB detachment procedure. However, the large difference in PB/OBCFA ratio in SAB isolates (0.94) and non-NDFOM that disappeared from WRP (3.20) suggests that SAB1 estimated using OBCFA as markers might be more credible. Martín-Orúe et al. (1998) conducted direct quantitative recoveries of extracted SAB biomass and concluded that only 32% of extracted PB, and 15 N, were recovered in SAB extracted biomass. Applying this recovery rate to SAB1 estimated with PB (158 × 0.32) a corrected estimate of 50.5 mg/g of WRPOM is obtained. This is very close to the SAB1 biomass estimated with OBCFA (47 mg/g of WRPOM ), suggesting that most of the extracted OBCFA would be recovered in SAB extracted biomass. Estimates of total SAB biomass in particles (27% and 16% of WRPOM , respectively for PB and OBCFA) are consistent with the range of values reported by Craig et al. (1987) and Legay-Carmier and Bauchart (1989). 5. Conclusions This study shows that the OBCFA are potentially useful as internal microbial markers in the rumen ecosystem. Differences in the level of extraction of PB and OBCFA from particles with the adopted SAB detachment procedure may reflect the interaction between bacterial lysis and the location of the internal markers in the cell (i.e., cytoplasmic for PB and in membranes for OBCFA). The composition of SAB1 and SAB2 was similar, except for the 15% higher protein content of SAB2 . Unlike the PB/N (mg/mg) ratio, the OBCFA/protein (mg/mg) ratio did not differ between LAB and SAB, which may be an advantage for using OBCFA as internal markers of rumen microbial mass. Acknowledgements M.R.G. Maia and E. Jerónimo gratefully acknowledge the receipt of scholarships from Fundac¸ão para a Ciência e Tecnologia (FCT), Portugal (Grants SFRH/BD/6976/2001 and SFRH/BD/23675/2005, respectively). References AOAC, 1990. Official Methods of Analysis, 15th ed. Association of Official Analytical Chemists, Arlington, VA, USA. Balcells, J., Ganuza, J.M., Pérez, J.F., Martín Orúe, S.M., Ronquillo, M.G., 1998. Urinary excretion of purine derivatives as an index of microbial-nitrogen intake in growing rabbits. Br. J. Nutr. 79, 373–380. Bates, D.B., Gillet, J.A., Barao, S.A., Bergen, W.G., 1985. The effect of specific growth rate and stage of growth on nucleic acid-protein values of pure cultures and mixed ruminal bacteria. J. Anim. Sci. 61, 713–724. Bauchart, D., Legay-Carmier, F., Doreau, M., Gaillard, B., 1990. Lipid metabolism in liquid-associated and solid-adherent bacteria in rumen contents of dairy cows offered lipid supplements diets. Br. J. Nutr. 63, 563–578. Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248–254. Broderick, G.A., Merchen, N.R., 1992. Markers for quantifying microbial protein synthesis in the rumen. J. Dairy Sci. 75, 2618–2632. Craig, W.M., Broderick, G.A., Ricker, D.B., 1987. Quantitation of microorganisms associated with the particulate phase of ruminal ingesta. J. Nutr. 117, 56–62. Diedrich, M., Henschel, K.P., 1990. The natural occurrence of unusual fatty-acids. 1. Odd numbered fatty-acids. Nahrung 34, 935–943. Djouvinov, D.S., Nakashima, Y., Todorov, N., Pavlov, D., 1998. In situ degradation of feed purines. Anim. Feed Sci. Technol. 71, 67–77.
206
R.J.B. Bessa et al. / Animal Feed Science and Technology 150 (2009) 197–206
Gonzaléz-Ronquilho, M., Balcells, J., Belenguer, J., Castrillo, C., Mota, M., 2004. A comparison of purine derivatives excretion with conventional methods as indices of microbial yield in dairy cows. J. Dairy Sci. 87, 2211–2221. Harfoot, C.G., Noble, R.C., Moore, J.H., 1975. The role of plant particles, bacteria and cell-free supernatant fractions of rumen contents in the hydrolysis of trilinolein and the subsequent hydrogenation of linoleic acid. Antonie van Leeuwenhoek 41, 533–542. Kaneda, T., 1991. Iso- and anteiso-fatty acids in bacteria: biosynthesis, function, and taxonomic significance. Microbiol. Rev. 55, 288–302. Kim, E.J., Sanderson, R., Dhanoa, M.S., Dewhurst, R.J., 2005. Fatty acid profiles associated with microbial colonization of freshly ingested grass and rumen biohydrogenation. J. Dairy Sci. 88, 3230. Legay-Carmier, F., Bauchart, D., 1989. Distribution of bacteria in the rumen contents of dairy cows given a diet supplemented with soya-bean oil. Br. J. Nutr. 61, 725–740. Makkar, H.P.S., Becker, K., 1999. Purine quantification in digesta from ruminants by spectrophotometric and HPLC methods. Br. J. Nutr. 81, 107–112. Martin, C., Williams, A.G., Michalet-Doreau, B., 1994. Isolation and characteristics of the protozoal and bacterial fractions from bovine ruminal contents. J. Anim. Sci. 72, 2962–2968. Martín-Orúe, S.M., Balcells, J., Zakraoui, F., Castrillo, C., 1998. Quantification and chemical composition of mixed bacteria harvested from solid fractions of rumen digesta, effect of detachment procedure. Anim. Feed Sci. Technol. 71, 269–282. Merry, R.J., McAllan, A.B., 1983. A comparison of the chemical composition of mixed bacteria harvested from the liquid and solid fractions of rumen digesta. Br. J. Nutr. 50, 701–709. Minato, H., Endo, A., Higuchi, M., Ootomo, Y., Uemura, T., 1966. Ecological treatise on the rumen fermentation I. The fractionation of bacteria attached to the rumen digesta solids. J. Gen. Microbiol. 12, 39–52. Moss, C.W., 1981. Gas–liquid chromatography as an analytical tool in microbiology. J. Chromatogr. 203, 337–347. Olsson, P.A., 1999. Signature fatty acids provide tools for determination of the distribution and interactions of mycorrhizal fungi in soil. FEMS Microbiol. Ecol. 29, 303–310. Olubobokun, J.A., Craig, W.M., 1990. Quantity and characteristics of microorganisms associated with ruminal fluid or particles. J. Anim. Sci. 68, 3360–3370. Perez, J.F., Balcells, J., Fondevila, M., Guada, J.A., 1998. Composition of liquid- and particle-associated bacteria and their contribution to the rumen outflow. Aust. J. Agric. Res. 49, 907–914. Rodriguez, C.A., Gonzalez, J., Alvir, M.R., Repetto, J.L., Centeno, C., Lamrani, F., 2000. Composition of bacteria harvested from the liquid and solid fractions of the rumen of sheep as influenced by feed intake. Br. J. Nutr. 84, 369–376. SAS, 2001. User’s Guide. Statistics, Release 8.2. SAS Inst. Inc., Cary, NC, USA. Sukhija, P.S., Palmquist, D.L., 1988. Rapid method for determination of total fatty acid content and composition of feedstuffs and feces. J. Agric. Food Chem. 36, 1202–1206. Van Soest, P.J., Robertson, J.B., Lewis, B.A., 1991. Methods for dietary fiber, neutral detergent fiber and non-starch polysaccharides in relation to animal nutrition. J. Dairy Sci. 74, 3583–3597. Vicente, F., Guada, J.A., Surra, J., Balcells, J., Castrillo, C., 2004. Microbial contribution to duodenal purine flow in fattening cattle given concentrate diets, estimated by purine N labelling (N-15) of different microbial fractions. Anim. Sci. 78, 159–167. Vlaeminck, B., Dufour, C., Van Vuuren, A.M., Cabrita, A.R.J., Dewhurst, R.J., Demeyer, D., Fievez, V., 2005. Use of odd and branchedchain fatty acids in rumen contents in rumen contents and milk as a potential microbial marker. J. Dairy Sci. 88, 1031–1042. Vlaeminck, B., Fievez, V., Cabrita, A.R.J., Fonseca, A.J.M., Dewhurst, R.J., 2006a. Factors affecting odd- and branched-chain fatty acids in milk: a review. Anim. Feed Sci. Technol. 131, 389–417. Vlaeminck, B., Fievez, V., Demeyer, D., Dewhurst, R.J., 2006b. Effect of forage:concentrate ratio on fatty acid composition of rumen bacteria isolated from ruminal and duodenal digesta. J. Dairy Sci. 89, 2668–2678. Volden, H., Mydland, L.T., Harstad, O.M., 1999. Chemical composition of protozoal and bacterial fractions isolated from ruminal contents of dairy cows fed diets differing in nitrogen supplementation. Acta Agric. Scand. A 49, 235–244. Yang, W.Z., Beauchemin, K.A., Rode, L.M., 2001. Effect of dietary factors on distribution and chemical composition of liquid- or solid-associated bacterial populations in the rumen of dairy cows. J. Anim. Sci. 79, 2736–2746.
Contents lists available at ScienceDirect
Animal Feed Science and Technology journal homepage: www.elsevier.com/locate/anifeedsci
Using microbial fatty acids to improve understanding of the contribution of solid associated bacteria to microbial mass in the rumen R.J.B. Bessa a,∗, M.R.G. Maia a, E. Jerónimo a, A.T. Belo b, A.R.J. Cabrita c, R.J. Dewhurst d, A.J.M. Fonseca e a
REQUIMTE, Unidade de Produc¸ão Animal, Instituto Nacional de Recursos Biológicos, Fonte Boa, 2005-048 Vale de Santarém, Portugal b Unidade de Produc¸ão Animal, Instituto Nacional de Recursos Biológicos, Fonte Boa, 2005-048 Vale de Santarém, Portugal c REQUIMTE, SAECA, Faculdade de Ciências, Universidade do Porto, Campus Agrário de Vairão, Rua Padre Armando Quintas, 4485-661 Vairão VC, Portugal d Teagasc, Animal Bioscience Centre, Dunsany, County Meath, Ireland e REQUIMTE, ICBAS, Instituto de Ciências Biomédicas de Abel Salazar, Universidade do Porto, Campus Agrário de Vairão, Rua Padre Armando Quintas, 4485-661 Vairão VC, Portugal
a r t i c l e
i n f o
Article history: Received 17 October 2007 Received in revised form 19 September 2008 Accepted 24 September 2008 Keywords: Microbial markers Odd and branched chain fatty acids Purine bases Rumen bacteria
a b s t r a c t This study sought to distinguish liquid-(LAB) and detached (SAB1 ) and undetached (SAB2 ) solid-associated bacteria through their fatty acid (FA) and purine base (PB) profiles. Fatty acids and PB were also evaluated as internal microbial markers for estimating microbial biomass associated with rumen particles. Four merino rams fitted with rumen cannulae and fed dehydrated alfalfa pellets provided rumen contents. In 3 consecutive weeks, rumen contents were collected and samples of LAB and SAB1 , total rumen content (TRC), washed rumen particles (WRP) and rumen particles after SAB1 extraction (ERP) were obtained and analysed for PB and FA. The SAB2 biomass composition was estimated from the non-NDF organic matter (OM) remaining in ERP. The concentration of total SAB biomass in particles was estimated using both PB and odd and branched-chain fatty acids (OBCFA). Concentrations of PB and
Abbreviations: CP, crude protein; ERP, extracted rumen particles; FA, fatty acids; LAB, liquid associated bacteria; NAN, nonammonia N; NDFom, neutral detergent fibre not assayed with a heat stable amylase and expressed exclusive of residual ash; OBCFA, odd and branched-chain fatty acids; OM, organic matter; PB, purine bases; SAB1 , detachable solid associated bacteria; SAB2 , undetachable solid associated bacteria; TRC, total rumen content; WRP, washed rumen particles. ∗ Corresponding author. Tel.: +351 243 767308; fax: +351 243 767307. E-mail address: [email protected] (R.J.B. Bessa). 0377-8401/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.anifeedsci.2008.09.005
198
R.J.B. Bessa et al. / Animal Feed Science and Technology 150 (2009) 197–206
OBCFA were highly correlated among the different rumen fractions. Marked differences between LAB and SAB populations occurred with LAB having higher PB content, lower FA content and a higher proportion (g/100 g fatty acids) of OBCFA than did SAB. The chemical composition of SAB1 and SAB2 was similar, except for the 15% higher crude protein content of the latter. The concentration of OBCFA (mg/g microbial OM) did not differ between bacterial fractions. The PB/OBCFA ratio (mg/mg) was higher in LAB (2.08) than in SAB (0.94). The ratio between branched-chain and odd-linear-chain FA was higher in LAB (2.26) than in SAB (1.46). Extraction of PB and OBCFA from WRP with our SAB detachment procedure was 61% and 31%, respectively. Estimated SAB1 and total SAB biomass (mg OM/g WRP) were 158 and 266, and 47 and 164, respectively, using PB and OBCFA as microbial markers. This study suggests that the OBCFA have potential as internal microbial markers in rumen ecosystem studies. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Internal microbial markers have been extensively used to predict the contribution of microbial crude protein (CP) to duodenal non-ammonia N (NAN) flows (Broderick and Merchen, 1992) and to study the distribution of microbes within the rumen ecosystem (Legay-Carmier and Bauchart, 1989). Several substances have been proposed as internal markers, including diaminopimelic acid, d-alanine, 2-aminoethylphosphonic acid, chitin, adenylate energy charge, purines and ribonucleic acid. Purine bases (PB) were regarded as one of the best methods by Broderick and Merchen (1992). However, because no internal marker has proven completely satisfactory (Broderick and Merchen, 1992), the search for alternative internal markers continues. Additionally, no marker has yet addressed the problem of obtaining representative samples of populations from either the rumen or the duodenum. The techniques used to remove attached bacteria (i.e., blending, pummelling, chilling, chemical treatment or a combination of treatments) vary in their efficacy (Martín-Orúe et al., 1998). Therefore, there is still a need to clarify the composition of undetached microbes associated with rumen particles, particularly in calculating microbial flows. Microbial fatty acid (FA) composition has been extensively used in systematic (Moss, 1981; Kaneda, 1991) and ecologic studies (Olsson, 1999). Recently, emphasis has been given to use of odd and branched chain fatty acids (OBCFA) in rumen studies (Vlaeminck et al., 2006a). These FA apparently fulfill most of the requirements needed for an internal marker of rumen microbial biomass as they are stable compounds, easy to measure, and are present at only trace levels in most plants (Diedrich and Henschel, 1990). Additionally, their pattern can vary between both liquid-(LAB) and solid-associated bacteria (SAB) (Vlaeminck et al., 2006b), as well among microbial species (Vlaeminck et al., 2006a). This study aimed to distinguish LAB, detachable (SAB1 ) and undetached (SAB2 ) microbes associated with rumen particles through their FA and PB profiles. Fatty acids and PB were also evaluated as internal microbial markers to estimate microbial biomass associated with rumen particles. 2. Materials and methods 2.1. Collection and fractionation of rumen contents Four merino rams (50 ± 3.2 kg) fitted with rumen cannulae (6 cm diameter) were used to provide rumen contents. Rams were fed 1 kg of dehydrated alfalfa pellets (CP: 177 g/kg DM; NDFom: 458 g/kg DM) and 0.1 kg of wheat straw (CP: 36 g/kg DM; NDFom: 700 g/kg DM) in two equal meals at 10:00 and 17:00 h. Rams had continuous access to water and mineral blocks. After 3-week adaptation periods, 1 l of rumen contents was collected from each ram before the morning meal of the same day of 3 consecutive weeks. The procedures used to fractionate rumen contents and to obtain bacterial pellets
R.J.B. Bessa et al. / Animal Feed Science and Technology 150 (2009) 197–206
199
Fig. 1. Procedures used for fractionation of rumen contents and to obtain rumen bacteria pellets.
are summarized in Fig. 1. At each collection, one aliquot of total rumen contents (TRC) from each ram was freeze-dried for subsequent analysis. The remaining TRC was filtered through eight layers of surgical gauze; the solid phase was washed, re-suspended in twice its weight of saline solution (0.85 g/100 ml, w/v NaCl, 39 ◦ C) and filtered again through eight layers of surgical gauze. The liquid phase of TRC and the washing liquid were pooled and subjected to differential centrifugation at 500 × g for 5 min, at 4 ◦ C. The supernatant was then centrifuged at 20,000 × g for 20 min, at 4 ◦ C, to obtain the LAB pellet, which was washed with saline solution and re-centrifuged. An aliquot of the washed solid phase of TRC (i.e., washed rumen particles, WRP) was freeze-dried and the remaining material was
200
R.J.B. Bessa et al. / Animal Feed Science and Technology 150 (2009) 197–206
subjected to SAB extraction following the procedure described by Martín-Orúe et al. (1998). Briefly, WRP were re-suspended in saline solution overnight at 4 ◦ C, homogenized 6 × 30 s with a Waring blender (Waring Products Division, New Hartford, CT, USA) and filtered through surgical gauze. The solid phase obtained was washed and freeze-dried (i.e., extracted rumen particles, ERP). The liquid phase, including the washing liquid, was subjected to differential centrifugation, as described above, to obtain the detachable SAB pellet (SAB1 ). All samples were freeze-dried and kept at −20 ◦ C until analysis. 2.2. Analytical procedures Samples of TRC, WRP and ERP were analysed for DM, ash (AOAC, 1990, ID 942.05), Kjeldahl N (AOAC, 1990, ID 954.01) and neutral detergent fibre (NDFom; Van Soest et al., 1991). Sodium sulfite was not added during NDF extraction and ␣-amylase was also not used, and NDF was expressed exclusive of residual ash. Kjeldahl N was also determined on the NDFom residue from ERP (N-NDFom). The protein content of bacterial samples (LAB and SAB1 ) was determined using the Bradford method (Bradford, 1976), using bovine serum albumin as standard (Merck KGaA, Darmstadt, Germany). The TRC, WRP, ERP, LAB and SAB1 samples were analysed for PB and FA contents. PB were analysed by HPLC after perchloric acid hydrolysis following Balcells et al. (1998). Separation was according to the method of Makkar and Becker (1999) at a flow rate of 1.0 ml/min and PB were detected using a UV detector at a wavelength of 254 nm. Fatty acid analysis was conducted by the one-step extraction–methylation method of Sukhija and Palmquist (1988). Fatty acid methyl esters were analysed on a HP6890 gas chromatograph (Hewlett-Packard, Avondale, PA, USA) operating with a flame ionization detector and equipped with a 60 m fused silica capillary column SP-2380 (Supelco, Bellefonte, PA, USA) with 0.25 mm of internal diameter and 0.20 m of film thickness, and using He at 2.0 ml/min as a carrier gas. Peak identification was based on comparison with known standards of FA methyl esters (Sigma Chemical Co., St. Louis, MO, USA). The octadecenoic isomers were not completely resolved and are presented as trans-18:1 (i.e., sum of the peaks eluting between 18:0 and cis-9 18:1), and cis-18:1 (sum of cis-9, cis-11 and cis-12 isomers). The conjugated linoleic acid (CLA) peak reported is consistent with retention time of 18:2 cis-9, trans-11 methyl ester, but could also include several other conjugated isomers, such as 18:2 trans-7, cis-9 and 18:2 trans-8, cis-10. Quantification was made using heneicosanoic acid (21:0) as the internal standard, assuming for all peaks equal proportionality between peak areas and FA methyl ester weight. 2.3. Calculations Consider ‘D’ as the proportion of non-NDF OM that is removed from WRP OM (WRPOM ) due to SAB extraction, calculated as D=
([ERP]NDF − [WRP]NDF ) [ERP]NDF
where: [ERP]NDF = concentration of NDF (mg/g OM) in ERP, and [WRP]NDF = concentration of NDF (mg/g OM) in WRP. The contribution of SABOM to WRPOM can be estimated as SAB1 (mg/g WRPOM ) =
[WRP]m − ([ERP]m × (1 − D)) × 1000 [SAB1 ]m
SAB2 (mg/g WRPOM ) =
[ERP]m × (1 − D) × 1000 [SAB2 ]m
SABT (mg/g WRPOM ) = SAB1 + SAB2 where: SAB1 = bacterial OM biomass extracted from WRPOM ; SAB2 = bacterial OM biomass remaining in the ERP OM; [WRP]m = concentration (mg/g OM) of marker in WRP; and [ERP]m = concentration (mg/g OM) of marker in ERP; [SAB1 ]m = concentration (mg/g OM) of marker in SAB1 ; [SAB2 ]m = concentration (mg/g OM) of marker in SAB2 .
R.J.B. Bessa et al. / Animal Feed Science and Technology 150 (2009) 197–206
201
The protein content of SAB2 was computed as 6.25 × ([ERP]N − ([ERP]N-NDF + [ERP]N-NA )) where: [ERP]N = concentration of N (Kjeldahl) (mg/g OM) in ERP; [ERP]N-NDF = concentration of N (Kjeldahl) (mg/g) in NDF from ERP; [ERP]N-NA = concentration of N present in nucleic acids computed from concentrations of adenine and guanine (mg/g OM) in ERP and complementary pyrimidine bases. The percentage of marker extraction was calculated as Marker extraction (%) =
[WRP]m − ([ERP]m × (1 − D)) . [WRP]m × 100
2.4. Statistical analysis Data were analysed by repeated measures analysis of variance using the MIXED procedure of SAS (2001) with week of sampling as the repeated measure. The compound symmetry covariance structure was used. The model included the fixed effects of rumen fractions, bacterial type or type of marker and week of sampling, the random effect of animal and the random residual error. Orthogonal contrasts were constructed in order to compare: (i) total rumen contents with rumen particles (TRC vs (WRP and ERP)), (ii) washed with extracted rumen particles (WRP vs ERP), (iii) LAB with SAB (LAB vs (SAB1 and SAB2 )), and (iv) detachable and undetachable SAB (SAB1 vs SAB2 ). Correlations between PB and FA concentrations in TRC, WRP and ERP were tested and the Pearson correlation coefficients reported. 3. Results 3.1. Composition of rumen contents PB, N, and FA contents were highest in TRC and lowest in ERP (Table 1). The NDFom content of these fractions followed the opposite trend, increasing from TRC through to ERP. Total FA concentration (mg/g OM) decreased from TRC through to ERP, although the proportions (g/100 g total fatty acids) of iso-14:0, 14:0, iso-15:0, 15:0, iso-16:0, 17:0, cis-18:1, and 20:0 did not change with rumen fractionation (Table 1). The anteiso-15:0, iso-17:0, anteiso-17:0, and the sum of the residual FA (most of them not identified) were lower in rumen particles (WRP and ERP) than in TRC. The proportions of 16:0, 18:0, trans-18:1, 18:2n-6, 18:3n-3, and 22:0 were higher in rumen particles (WRP and ERP) than in TRC. The proportion of iso-15:0, trans-18:1 and 22:0 were higher in ERP than WRP. Total FA and all individual FA concentrations (mg/g OM) in TRC, WRP and ERP were positively correlated with PB concentration (data not shown). The highest Pearson correlation coefficients with PB concentration were for 16:0 (0.91), total FA (0.86) and anteiso-15:0 (0.83), the correlation for total OBCFA being 0.79. Although 16:0 and total FA had higher correlations with PB than OBCFA, they can be quite unspecific. Therefore, OBCFA were used to estimate microbial biomass associated with rumen particles instead of individual FA. 3.2. Composition of rumen bacteria Table 2 presents PB, protein and total FA contents and proportions of FA for rumen bacteria. There were differences between LAB and SAB populations, LAB having higher PB content and PB/protein ratio and lower FA content. The PB and FA contents of SAB1 and SAB2 were similar, though SAB2 contained 15% more protein, and consequently had a lower PB/protein than SAB1 . Fatty acid proportions were different between LAB and SAB; LAB having a higher proportion (g/100 g total FA) of OBCFA, and SAB having a higher proportion of C18 biohydrogenation derived FA (18:0; trans-18:1 and CLA). The CLA was much higher in SAB than in LAB; it was hardly present in LAB. The concentration of OBCFA (mg/g microbial OM) did not differ between bacterial types. The ratio between branched-chain and odd linear-chain FA was higher in LAB than in SAB.
202
R.J.B. Bessa et al. / Animal Feed Science and Technology 150 (2009) 197–206
Table 1 Chemical composition (mg/g OM) and fatty acid (FA) of total rumen contents (TRC), washed ruminal particles (WRP), and extracted ruminal particles (ERP). TRC
WRP
ERP
SEM
TRC vs RP
WRP vs ERP
Concentrations (mg/g OM) Purine bases (PB) N NDF FA OBCFAa
3.47 32.9 586 30.4 2.36
2.19 23.9 733 20.2 1.40
1.01 18.3 872 15.8 1.13
0.223 1.05 11.9 0.65 0.039
<0.001 <0.001 <0.001 <0.001 <0.001
<0.001 <0.001 <0.001 0.003 0.005
Ratios PB/OBCFA Branched-/odd-chain FA
1.48 1.56
1.58 1.31
0.90 1.41
0.134 0.073
0.017 0.017
<0.001 0.204
0.32 1.25 1.03 1.37 2.26 0.62 19.0 0.85 0.59 0.79 29.3 9.70 5.47 4.26 3.49 0.91 0.79 18.0
0.32 1.18 0.97 1.14 2.25 0.58 19.7 0.56 0.51 0.77 30.4 10.90 5.44 5.23 3.94 0.98 0.90 14.3
0.20 1.20 1.21 1.14 2.20 0.67 19.9 0.71 0.45 0.78 31.2 11.60 5.60 4.88 3.77 0.95 1.04 12.8
0.056 0.042 0.069 0.035 0.046 0.058 0.58 0.056 0.025 0.038 0.91 0.530 0.550 0.300 0.150 0.067 0.037 0.98
0.136 0.123 0.487 0.002 0.403 0.941 0.044 0.020 0.002 0.391 0.047 <0.001 0.839 0.025 0.040 0.379 0.001 0.007
0.094 0.712 0.049 0.993 0.431 0.367 0.744 0.103 0.070 0.579 0.352 0.040 0.523 0.300 0.332 0.702 0.008 0.275
FA proportions, g/100 g total FA iso-14:0 14:0 iso-15:0 anteiso-15:0 15:0 iso-16:0 16:0 iso-17:0 anteiso-17:0 17:0 18:0 trans-18:1 cis-18:1 18:2n-6 18:3n-3 20:0 22:0 Othersb a b
Odd and branched-chain fatty acids. The remaining fatty acids, most of them unidentified.
Differences in FA proportions between SAB1 and SAB2 were observed. Compared to SAB1 , the SAB2 had lower proportions of odd-chain FA and some of the branched-chain FA (iso-14:0 and iso-16 and anteiso-17:0), but the ratio between branched-chain and odd linear-chain FA was equal. Moreover, the SAB2 had higher proportions of trans-18:1, CLA and particularly of 18:2n-6 and 18:3n-3 than SAB1 . 3.3. Marker extraction and SAB biomass estimation The extraction of PB and OBCFA from WRP with the current SAB detachment procedure (Table 3) was 61% and 31%, respectively. Detached SAB biomass (SAB1 ) and total SAB biomass estimates were higher for PB than for OBCFA. The undetached SAB (SAB2 ) biomass estimate did not differ between markers. Estimates of total SAB biomass in particles was 27% and 16% of WRPOM , and comprise 99% and 62% of non-NDFOM in WRP, respectively for PB and OBCFA. 4. Discussion 4.1. Composition of rumen contents In this experiment, the contribution of dietary PB can be considered negligible, as the mean PB concentration in WRP was 2190 mg/kg OM in comparison with 7.6 mg/kg OM in dehydrated alfalfa. Additionally, Djouvinov et al. (1998) found that 61% of dietary PB were degraded after 17 h in the rumen. Although the distribution of markers between liquid and solid phases was not measured,
R.J.B. Bessa et al. / Animal Feed Science and Technology 150 (2009) 197–206
203
Table 2 Chemical composition (mg/g organic matter, OM) and fatty acid (FA) (g/100 g total fatty acids) in liquid-(LAB) and solid associated bacteria (SAB).
Concentrations (mg/g OM) PB Protein Fatty acids (FA) OBCFAa
LAB
SAB1
SAB2
SEM
LAB vs SAB
SAB1 vs SAB2
20.9 393 61.0 10.2
9.7 450 113.9 10.3
8.0 520 126.6 9.3
0.85 17.3 4.30 0.75
<0.001 0.001 <0.001 0.490
0.161 0.031 0.084 0.225
Ratios PB/protein (mg/g) OBCFA/protein (mg/g) PB/OBCFA Branched-/odd linear FA
54.5 26.4 2.08 2.27
20.6 23.2 0.94 1.46
14.9 17.9 0.89 1.45
2.36 1.55 0.122 0.081
<0.001 0.002 <0.001 <0.001
0.083 0.013 0.588 0.943
FA proportions, g/100 g total FA iso-14:0 14:0 iso-15:0 anteiso-15:0 15:0 iso-16:0 16:0 iso-17:0 anteiso-17:0 17:0 18:0 trans-18:1 cis-18:1 18:2n-6 CLA 18:3n-3 20:0 22:0 Othersb OBCFA
0.72 1.03 2.48 4.98 4.19 1.06 19.9 1.93 0.93 1.06 23.3 5.69 4.96 2.84 0.07 1.47 0.65 0.50 22.3 17.3
0.39 0.67 1.02 1.56 2.67 0.88 20.9 0.71 0.78 1.00 30.8 9.85 5.06 3.01 0.47 1.72 0.71 0.60 17.2 9.02
0.16 1.19 1.21 1.14 2.20 0.67 19.9 0.71 0.45 0.78 31.2 11.56 5.60 4.87 0.66 3.76 0.95 1.04 12.0 7.32
0.069 0.068 0.134 0.333 0.165 0.089 0.542 0.130 0.033 0.043 1.03 0.490 0.439 0.147 0.048 0.075 0.090 0.064 0.78 0.792
<0.001 0.152 <0.001 <0.001 <0.001 0.016 0.489 <0.001 <0.001 0.002 0.002 <0.001 0.153 0.002 <0.001 <0.001 0.137 0.006 <0.001 <0.001
0.016 <0.001 0.296 0.362 0.042 0.077 0.186 0.981 <0.001 0.001 0.825 0.008 0.084 <0.001 0.006 <0.001 0.101 0.003 0.005 0.133
SAB1 : detachable bacterial OM biomass; SAB2 : undetachable bacterial OM biomass. a b
Odd and branched-chain fatty acids. The remaining fatty acids, most of them unidentified.
the higher PB and FA contents of TRC in comparison with WRP was likely due to the presence of liquid phase biomass (e.g., LAB and protozoa), biomass associated with very small highly colonized particles (Legay-Carmier and Bauchart, 1989; Yang et al., 2001), and eventually fatty acid soaps in suspension.
Table 3 Percentage extraction of marker from rumen particles and microbial biomass concentration in washed rumen particle organic matter (WRPOM ) estimated using either purine bases (PB) or total odd and branched-chain fatty acids (OBCFA) as microbial markers. PB Marker extraction from WRP (%) Biomass (mg/g WRPOM ) SAB1 SAB2 SABT SABT (% non-NDFOM in WRP)
61.0 158 108 266 99
OBCFA 30.5 47 117 164 62
SEM 3.99 22.5 8.3 20.4 4.8
SAB1 : detachable bacterial OM biomass; SAB2 : undetachable bacterial OM biomass; SABT : SAB1 + SAB2 .
P 0.002 0.010 0.391 0.019 0.021
204
R.J.B. Bessa et al. / Animal Feed Science and Technology 150 (2009) 197–206
4.2. Composition of rumen bacteria The present data on chemical composition of rumen bacteria is in general accordance with published information. The higher PB content and PB/protein ratio and lower FA content of LAB compared to SAB populations are in accordance with previous reports (Merry and McAllan, 1983; Olubobokun and Craig, 1990; Martin et al., 1994; Vicente et al., 2004). These differences might be related to distinct distributions of bacterial species between liquid and particulate phases (Minato et al., 1966) as well to differences in metabolic activities (Merry and McAllan, 1983; Bauchart et al., 1990) and growth rate (Bates et al., 1985). Concentrations of PB in SAB1 agree with some previous reports (Volden et al., 1999; Rodriguez et al., 2000), but are lower than those reported by others using a similar analytical method (Perez et al., 1998; Gonzaléz-Ronquilho et al., 2004; Vicente et al., 2004). The higher C18 biohydrogenation derived FA (18:0, trans 18:1, and CLA) in SAB than in LAB agree with previous studies (Bauchart et al., 1990; Vlaeminck et al., 2006b) and could have been related to higher biohydrogenation activity in rumen particles (Harfoot et al., 1975). The CLA concentration was much higher in SAB than in LAB, and it was hardly present in LAB. Reports on CLA concentration in LAB and SAB are scarce, but Vlaeminck et al. (2006b) found higher CLA proportions in SAB than in LAB, whereas this pattern was only partly reflected in data of Kim et al. (2005). The ratio between branched-chain and odd linear-chain FA was higher in LAB than in SAB, as found by Vlaeminck et al. (2006b) when collating published data. The present data are the first estimate of chemical composition of undetached microbes associated with rumen particles (SAB2 ). The chemical composition of SAB1 and SAB2 was similar, except for the 15% higher protein content of the latter. This difference could be due to real differences in chemical composition, to presence of non-microbial N in non-NDFOM remaining in ERP (SAB2 ) or/and to different methodology used for quantification of the different bacterial fractions. The protein content of SAB2 was estimated using a factorial approach where N associated with NDFom and N associated with nucleic acids (averaging 37% and 4% of ERP total N, respectively) were subtracted from total N in ERP. Moreover, the SAB2 had higher proportions of trans-18:1, CLA and particularly of 18:2n-6 and 18:3n-3 than SAB1 . The higher concentration of 18:2n-6 (6.17 vs 3.42 mg/g non-NDFOM ) and of 18:3n-3 (4.77 vs 1.92 mg/g non-NDF OM ) in SAB2 than in SAB1 suggests that plant structural membranes (e.g., from chloroplasts) might still be present in ERP. As far as we know, there are no simultaneous published values for total PB, N and OBCFA (and hence direct PB/OBCFA ratios) of rumen bacteria, but Vlaeminck et al. (2005) found a slightly higher PB/OBCFA ratio in SAB (1.17) in available literature. Both TRC and WRC had an intermediate PB/OBCFA ratio between LAB and SAB. Both PB and OBCFA are expected to be of microbial origin, and that the PB/OBCFA and branched-chain and odd linear-chain FA ratios did not differ between SAB1 and SAB2 suggests that the biomass composition of detached SAB (SAB1 ) and SAB remaining adherent in ERP (SAB2 ) is quite similar, and that the differences between SAB1 and SAB2 resulted from the presence of an N rich non-NDFOM fraction of plant (i.e., feed) origin. Indeed, we can estimate that microbial N should only be 73% of N present in SAB2 fraction when using both PB/OBCFA and PB/protein ratios. Unlike the PB/protein ratio, the OBCFA/protein ratio did not differ between LAB and SAB1 (P=0.12; not shown in Table 2), which is an advantage for use of OBCFA as an internal marker of microbial mass flowing from the rumen. 4.3. Marker extraction and SAB biomass estimation There were large differences between PB and OBCFA extraction from WRP (61% vs 31%), that are reflected in the large difference in SAB1 estimates (158 vs 47 mg/g WRPOM ) obtained with each marker. Marker extraction is computed as the difference between marker concentration in WRP and ERP (corrected for losses of non-NDFOM ) and might reflect PB leakage and solubilisation associated with bacterial cell lysis and damage during extraction. Martín-Orúe et al. (1998) also reported high PB extraction from WRP with their SAB detachment procedures (60–71%). Conversely, extraction of OBCFA was close to at reported for diaminopimelic acid (Legay-Carmier and Bauchart, 1989) which, like OBCFA is not contained in the cytoplasm, suggesting that the low extractions observed might be due to persistence of the bacterial cell wall and membranes adherent to particles even after cell damage or lysis. As
R.J.B. Bessa et al. / Animal Feed Science and Technology 150 (2009) 197–206
205
far as we know, there is no previous report on the level of extraction of OBCFA with the SAB detachment procedure. The proportion of non-NDFOM that disappeared from WRP with the current SAB detachment procedure was 0.159 (S.E. = 0.014). However, the PB/OBCFA ratio of the non-NDFOM that disappeared was 3.20 (S.E. = 0.434), much higher than the PB/OBCFA ratio of SAB1 (0.94), SAB2 (0.89), WRP (1.58) and ERP (0.90). The high value of PB/OBCFA ratio in the extracted material is consistent with the much higher PB extraction than from OBCFA. The SAB1 biomass estimates expressed as non-NDFOM that disappeared from WRP was 99% and 32% for PB and OBCFA, respectively. In three of 12 observations, the SAB biomass estimates using PB as marker produced an obvious overestimation (176%, 129% and 111% of WRP non-NDFOM ) and if these values were excluded the average is reduced to 78% (S.D. = 12.6). Therefore, the estimates of SAB1 biomass in WRP using PB indicate that SAB1 constitutes most of the non-NDFOM that disappeared from WRP with this SAB detachment procedure. However, the large difference in PB/OBCFA ratio in SAB isolates (0.94) and non-NDFOM that disappeared from WRP (3.20) suggests that SAB1 estimated using OBCFA as markers might be more credible. Martín-Orúe et al. (1998) conducted direct quantitative recoveries of extracted SAB biomass and concluded that only 32% of extracted PB, and 15 N, were recovered in SAB extracted biomass. Applying this recovery rate to SAB1 estimated with PB (158 × 0.32) a corrected estimate of 50.5 mg/g of WRPOM is obtained. This is very close to the SAB1 biomass estimated with OBCFA (47 mg/g of WRPOM ), suggesting that most of the extracted OBCFA would be recovered in SAB extracted biomass. Estimates of total SAB biomass in particles (27% and 16% of WRPOM , respectively for PB and OBCFA) are consistent with the range of values reported by Craig et al. (1987) and Legay-Carmier and Bauchart (1989). 5. Conclusions This study shows that the OBCFA are potentially useful as internal microbial markers in the rumen ecosystem. Differences in the level of extraction of PB and OBCFA from particles with the adopted SAB detachment procedure may reflect the interaction between bacterial lysis and the location of the internal markers in the cell (i.e., cytoplasmic for PB and in membranes for OBCFA). The composition of SAB1 and SAB2 was similar, except for the 15% higher protein content of SAB2 . Unlike the PB/N (mg/mg) ratio, the OBCFA/protein (mg/mg) ratio did not differ between LAB and SAB, which may be an advantage for using OBCFA as internal markers of rumen microbial mass. Acknowledgements M.R.G. Maia and E. Jerónimo gratefully acknowledge the receipt of scholarships from Fundac¸ão para a Ciência e Tecnologia (FCT), Portugal (Grants SFRH/BD/6976/2001 and SFRH/BD/23675/2005, respectively). References AOAC, 1990. Official Methods of Analysis, 15th ed. Association of Official Analytical Chemists, Arlington, VA, USA. Balcells, J., Ganuza, J.M., Pérez, J.F., Martín Orúe, S.M., Ronquillo, M.G., 1998. Urinary excretion of purine derivatives as an index of microbial-nitrogen intake in growing rabbits. Br. J. Nutr. 79, 373–380. Bates, D.B., Gillet, J.A., Barao, S.A., Bergen, W.G., 1985. The effect of specific growth rate and stage of growth on nucleic acid-protein values of pure cultures and mixed ruminal bacteria. J. Anim. Sci. 61, 713–724. Bauchart, D., Legay-Carmier, F., Doreau, M., Gaillard, B., 1990. Lipid metabolism in liquid-associated and solid-adherent bacteria in rumen contents of dairy cows offered lipid supplements diets. Br. J. Nutr. 63, 563–578. Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248–254. Broderick, G.A., Merchen, N.R., 1992. Markers for quantifying microbial protein synthesis in the rumen. J. Dairy Sci. 75, 2618–2632. Craig, W.M., Broderick, G.A., Ricker, D.B., 1987. Quantitation of microorganisms associated with the particulate phase of ruminal ingesta. J. Nutr. 117, 56–62. Diedrich, M., Henschel, K.P., 1990. The natural occurrence of unusual fatty-acids. 1. Odd numbered fatty-acids. Nahrung 34, 935–943. Djouvinov, D.S., Nakashima, Y., Todorov, N., Pavlov, D., 1998. In situ degradation of feed purines. Anim. Feed Sci. Technol. 71, 67–77.
206
R.J.B. Bessa et al. / Animal Feed Science and Technology 150 (2009) 197–206
Gonzaléz-Ronquilho, M., Balcells, J., Belenguer, J., Castrillo, C., Mota, M., 2004. A comparison of purine derivatives excretion with conventional methods as indices of microbial yield in dairy cows. J. Dairy Sci. 87, 2211–2221. Harfoot, C.G., Noble, R.C., Moore, J.H., 1975. The role of plant particles, bacteria and cell-free supernatant fractions of rumen contents in the hydrolysis of trilinolein and the subsequent hydrogenation of linoleic acid. Antonie van Leeuwenhoek 41, 533–542. Kaneda, T., 1991. Iso- and anteiso-fatty acids in bacteria: biosynthesis, function, and taxonomic significance. Microbiol. Rev. 55, 288–302. Kim, E.J., Sanderson, R., Dhanoa, M.S., Dewhurst, R.J., 2005. Fatty acid profiles associated with microbial colonization of freshly ingested grass and rumen biohydrogenation. J. Dairy Sci. 88, 3230. Legay-Carmier, F., Bauchart, D., 1989. Distribution of bacteria in the rumen contents of dairy cows given a diet supplemented with soya-bean oil. Br. J. Nutr. 61, 725–740. Makkar, H.P.S., Becker, K., 1999. Purine quantification in digesta from ruminants by spectrophotometric and HPLC methods. Br. J. Nutr. 81, 107–112. Martin, C., Williams, A.G., Michalet-Doreau, B., 1994. Isolation and characteristics of the protozoal and bacterial fractions from bovine ruminal contents. J. Anim. Sci. 72, 2962–2968. Martín-Orúe, S.M., Balcells, J., Zakraoui, F., Castrillo, C., 1998. Quantification and chemical composition of mixed bacteria harvested from solid fractions of rumen digesta, effect of detachment procedure. Anim. Feed Sci. Technol. 71, 269–282. Merry, R.J., McAllan, A.B., 1983. A comparison of the chemical composition of mixed bacteria harvested from the liquid and solid fractions of rumen digesta. Br. J. Nutr. 50, 701–709. Minato, H., Endo, A., Higuchi, M., Ootomo, Y., Uemura, T., 1966. Ecological treatise on the rumen fermentation I. The fractionation of bacteria attached to the rumen digesta solids. J. Gen. Microbiol. 12, 39–52. Moss, C.W., 1981. Gas–liquid chromatography as an analytical tool in microbiology. J. Chromatogr. 203, 337–347. Olsson, P.A., 1999. Signature fatty acids provide tools for determination of the distribution and interactions of mycorrhizal fungi in soil. FEMS Microbiol. Ecol. 29, 303–310. Olubobokun, J.A., Craig, W.M., 1990. Quantity and characteristics of microorganisms associated with ruminal fluid or particles. J. Anim. Sci. 68, 3360–3370. Perez, J.F., Balcells, J., Fondevila, M., Guada, J.A., 1998. Composition of liquid- and particle-associated bacteria and their contribution to the rumen outflow. Aust. J. Agric. Res. 49, 907–914. Rodriguez, C.A., Gonzalez, J., Alvir, M.R., Repetto, J.L., Centeno, C., Lamrani, F., 2000. Composition of bacteria harvested from the liquid and solid fractions of the rumen of sheep as influenced by feed intake. Br. J. Nutr. 84, 369–376. SAS, 2001. User’s Guide. Statistics, Release 8.2. SAS Inst. Inc., Cary, NC, USA. Sukhija, P.S., Palmquist, D.L., 1988. Rapid method for determination of total fatty acid content and composition of feedstuffs and feces. J. Agric. Food Chem. 36, 1202–1206. Van Soest, P.J., Robertson, J.B., Lewis, B.A., 1991. Methods for dietary fiber, neutral detergent fiber and non-starch polysaccharides in relation to animal nutrition. J. Dairy Sci. 74, 3583–3597. Vicente, F., Guada, J.A., Surra, J., Balcells, J., Castrillo, C., 2004. Microbial contribution to duodenal purine flow in fattening cattle given concentrate diets, estimated by purine N labelling (N-15) of different microbial fractions. Anim. Sci. 78, 159–167. Vlaeminck, B., Dufour, C., Van Vuuren, A.M., Cabrita, A.R.J., Dewhurst, R.J., Demeyer, D., Fievez, V., 2005. Use of odd and branchedchain fatty acids in rumen contents in rumen contents and milk as a potential microbial marker. J. Dairy Sci. 88, 1031–1042. Vlaeminck, B., Fievez, V., Cabrita, A.R.J., Fonseca, A.J.M., Dewhurst, R.J., 2006a. Factors affecting odd- and branched-chain fatty acids in milk: a review. Anim. Feed Sci. Technol. 131, 389–417. Vlaeminck, B., Fievez, V., Demeyer, D., Dewhurst, R.J., 2006b. Effect of forage:concentrate ratio on fatty acid composition of rumen bacteria isolated from ruminal and duodenal digesta. J. Dairy Sci. 89, 2668–2678. Volden, H., Mydland, L.T., Harstad, O.M., 1999. Chemical composition of protozoal and bacterial fractions isolated from ruminal contents of dairy cows fed diets differing in nitrogen supplementation. Acta Agric. Scand. A 49, 235–244. Yang, W.Z., Beauchemin, K.A., Rode, L.M., 2001. Effect of dietary factors on distribution and chemical composition of liquid- or solid-associated bacterial populations in the rumen of dairy cows. J. Anim. Sci. 79, 2736–2746.