Composition, digestibility and particle-associated enzyme activities in rumen digesta as influenced by particle size and time after feeding

Three bulls were given twice daily a diet of grass silage and barley (70:30) on a D M basis. Manual evacuation of rumen contents was made before feeding , 3, 6 and 9 h after feeding. The interval between two subsequent evacuations was 27 h. Particle size distribution in the digesta was determined by wet-sieving. The greatest diurnal variation in ruminal NDF pool occurred in the 3-mm fraction. Small particles (0.04-0.2 mm) contained less NDF and cellulose and more A D L compared with larger particles. Potential NDF digestibility determined by 288 h nylon bag incubation declined as particle size decreased except for the 0.04-mm fraction. NDF digestibility was positively correlated with the proportion of cellulose in NDF and negatively with that of A D L and hemicellulose. Chemical composition of rumen particulate D M was comparatively unaffected by the time after feeding but NDF digestibility decreased with time when averaged across particle sizes. Particle-associated CM Case and xylanase activities increased gradually as the particle size decreased. The activities were 3 to 9-fold higher in the 0.04-mm fraction than in the other four fractions. There was a close positive relationship between enzyme activities and content of neutral-detergent solubles in particles.


INTRODUCTION
It is generally accepted that feed particles can only leave the rumen after reduction of particulate matter below a critical size (Poppi et al., 1980;Ulyatt et al., 1986).Feed is broken down physically by ingestive mastication and rumination.Particle size reduction by microbial fermentation alone is minor (Murphy and Nicoletti, 1984) but microbial digestion renders the plant material more sensitive to rumination.Rates of particle size reduction, fermentation and passage are the major constraints describing digesta kinetics in the rumen.These processes affect rumen fill, feed intake and fermentation end-products.
Microbial attachment to fibrous substrates is an important prerequisite of the degradation of these substrates in the rumen (Cheng et al., 1984).The adherent microbial population may comprise 70-80% of the microbial organic matter (Craig et al., 1987), and there is a close relationship between the rate of digestion of fibre and quantity of adherent bacteria (Gerson et al., 1988) or particle-associated enzyme activities (Silva et al., 1987;Huhtanen and Khalili, 1992).Particle-associated microbial population is dependent on available attachment sites, and is affected by rumen environment (Forsberg and Lam, 1977;Silva et al., 1987;Huhtanen and Khalili, 1992).However, there is little information on distribution of the adherent population relative to particle size or particle composition.Gerson et al. (1988) reported a greater microbial density in small particles than in large particles of meadow hay.
The objectives of the present study was to determine the quantity and distribution of NDF among different particle size fractions, and also between digestible and indigestible fractions.The chemical composition, digestibility and particle-associated enzyme acitivies as influenced by particle size and time after feeding were determined with a view toward improving the understanding of their possible significance with respect to rumen function dynamics.

Animals and diets
Three Friesian bulls (mean live weight 520 kg), each fitted with a rumen cannula, were used.The experimental diet consisted of grass silage made from timothy (Phleum pratense) and barley in the ratio of 70:30 on a DM basis.A daily feed allowance of 7 kg DM day-1 was provided in two equal meals at 7 a.m. and 7 p.m.The animals were kept in individual standings and had free access to water.A commercial mineral mixture (100 g day-1 ) was given to meet the animals' requirements.

Experimental procedure
The experimental measurements were carried out after an adaptation period of 21 d.Rumen pool size of total digesta was measured by manually removing all the rumen contents.Measurements were made at 7, 10, 13 and 16 h on four consecutive days.The time interval between each emptying was therefore 27 h.In later studies (unpublished) this procedure had no effect on rumen microbial activity as indicated by a similar degradation of hay DM in 24 h before and during the 3-d rumen evacuation period.Rumen contents was weighed, thoroughly mixed, sampled and returned to the rumen.Each sample was divided into two subsamples of which one was used for determination of DM content and subsequently for NDF, ADF and ADL analyses.The other subsample was used for particle size analyses and enzyme assays.
Rumen digesta was fractioned into various particle size groups using a wet-sieving technique.The sample (60 g wet weight) was first sieved using screens (diameter 20 cm) with mesh sizes 3.0, 1.0 and 0.5 mm.The slurry passed through the screens was then filtered through two nylon bags (pore sizes 200 and 40 fim).The bag of larger pore size (external dimensions 6x12 cm) was placed within the bag of smaller pore size (external dimensions 8x16 cm).The tops of the bags were tied with a tube of running tap water.During rinsing the bags were occasionally squeezed.All particle size distribution measurements were made in triplicate.
Immediately after rumen evacuation one sample of rumen contents was sieved, using a similar procedure as described above, in order to measure enzyme-activities of microbes firmly attached to rumen particles.After sieving, the material was transferred quantitatively into nylon bags (pore size 40/mi), washed under running tap water and squeezed thoroughly by hand.For determination of particle-associated carboxymethylcellulase (CMCase) and xylanase activities, a sample of 1.0 g of each particle size fraction was used.The details of enzyme extraction and enzyme assays are described by Silva et al. (1987) and Huhtanen and Khalili (1992).The DM content of the particles was determinend in duplicate.Enzyme activities were expressed as /xmol reducing sugars produced per minute per g DM under the contidions used.The activities were corrected for the substrate breakdown during the extraction procedure.

Statistical analyses
The model used to analyze rumen pool size data was: A; + Hj + e^where A and H are the effects of animal and evacuation time.For some parameters, the effect of time was further separated into single degree comparisons using polynomial contrasts.The model used to analyze composition, degradability and enzyme activities of rumen particles was: A i + H j +e ij + S k + (AS) ik + (HS) jk + e ijk , where A, H and S are the effects of animal, time and particle size, and e$j and e ijk are main plot (6 df) and sub-plot (24 df) errors.Correlation coefficients between chemical composition, NDFD and enzyme acitivities were calculated.

RESULTS
The silage used contained 207 g DM kg-' and 24.6 g N, 520 g NDF and 281 g ADF kg-1 DM.The corresponding values for barley were 871,21.1,187and 46, respectively.The silage was of high quality in terms of low pH (3.73) and low concentrations of fermentation acids and ammonia N.
The amount of rumen digesta decreased from 55.2 kg at 3 h to 39.0 kg at 12 h after feeding (linear trend P< 0.001) (Table 1).As diurnal variation in digesta composition was small, differences in the pool size of DM and cell wall constituents were similar to those in the total digesta.The reduction in the total weight of DM occurred in two phases with an initial rapid loss during the first TABLE 1 The effect of time after feeding on the pool size of rumen digesta, dry matter (DM), distribution of DM in various particle groups and pool size of cell wall constituents 3 h after feeding followed by a more gradual decline between 3-12 h after feeding.The rate at which DM left the rumen was 548 g h-1 between 0-3 h, and 208 gh-1 between 3 and 12 h.The amount of DM leaving the rumen over the first 3 h after feeding was calculated as follows: the amount of DM present in the rumen at 3 h was subtracted from the sum of DM present before feeding and that ingested.The greatest decrease in rumen DM pool occurred in the 3-mm fraction accounting for 57% of the total amount of DM disappeared between 3-12 h after feeding.The changes in the particle size distribution with time after feeding is shown in Fig. 1.There was a reduction in the 3-mm fraction and generally an inrease in the 0.2-and < 0.04-mm fractions with time but very little change occurred in other fractions.
Figure 1.Change in particle size distribution of rumen contents with time after feeding.
Postprandial changes in the composition of rumen particulate matter were fairly small (Table 2).The only significant (P < 0.05) differences were in ADF and C content which both increased with time.The NDFD decreased (P < 0.05) with time, and a similar trend was also observed for ADFD.On the other hand, chemical composition, NDFD and ADFD varied significantly (P< 0.001) with particle size.NDF and ADF contents were much lower and ADL content higher in the 0.04-mm fraction than in the other fractions.The composition of NDF was also affected by the particle size; the proportion of C decreased and that of HC and ADL increased gradually as the particle size decreased.The difference was Chemical composition of feed particles affected by time after feeding and particle size especially distinct between the 0.02-and 0.04-mm fractions.The NDFD decreased from 750 mg g-1 in the 3-mm fraction to 567 mg g-1 in the 0.2-mm fraction.In the smallest fraction (0.04 mm) NDFD was similar to that in the 3-mm fraction but ADFD was lower.

Time after feeding (h)
Figure 2. Changes in potential NDF digestibility with particle size and time after feeding.
The distribution of NDF in various particle pools is shown in Table 3.The greatest reduction with time after feeding occurred in the total NDF and DNDF retained by the 3.0 mm sieve (linear trend P< 0.001).The amount of NDF in the 0.04-mm fraction also decreased linearly with time.As a result of decreasing NDFD with time, relative diurnal changes in the amount of rumen NDF retained by 3.0 and 1.0-mm sieves were greater for digestible than for indigestible fraction.The average rate of disappearance of particles was 96.7 and 17.4 g h-1 for NDF retained by the 3.0 and 1.0-mm sieves.
The effects of time and particle size on particle-associated enzyme activities are shown in Table 4.Both CMCase and xylanase activities increased with time (linear effect P < 0.05).The enzyme acitivities increased gradually as the particle size decreased from > 3.0 mm to 0.2-0.5 mm, and increased to 3.0-4.5-fold in the 0.04-mm fraction as compared with 0.2-mm fraction.Time x size interaction both in CMCase and xylanase activity was significant (P< 0.01).Xylanase increased from 20.3 to 44.7 between 3 -12 h post-feed in the 3-mm fraction while the variation in smaller particles was much less; e.g.32.4 to 46.8 in the 0.5-mm fraction and 40.0 to 47.7 in the 0.2-mm fraction.The content of neutral-detergent (ND) solubles of particulate DM was highly correlated with CMCase (r = 0.953) and xylanase (r = 0.899).

DISCUSSION
The feeding level used (65 g DM (kg LW)-1 ) was restricted below to appetite to ensure that the meals were eaten in a short period.Rapid loss of DM from the rumen during the 3 h after feeding agrees with the observations of Moseley and Jones (1984) and Aitchison et al. (1986).Since the frequency of reticulum contractions increases during eating (Balch, 1971), greater DM loss during this  period may be due to increased passage rate (Moseley and Jones, 1984).Rapid fermentation of forage celi solubles (Van Soest, 1982) and barley starch may also explain the higher removal rate of DM during the initial period after feeding in the present study.The high content of NDF in rumen particulate matter at 3h after feeding also reflects a rapid loss of non-cell wali materiał either due to solubilization or fermentation.
The rate of removal of NDF and DNDF from the rumen appeared to be a more exponential process.The following relations between rumen pool size of NDF and DNDF (> 0.04 mm) and time after feeding were calculated: NDF = 8.191 xe-°.°45ot?r = 0.993;DNDF = 7.898xe-°-654t ,r = 0.996.Usingthe value of 0.010-0.015for the fractional passage rate of DNDF from the rumen (Huhtanen and Khalili, 1991 ;Huhtanen and Jaakkola, 1992) the fractional rate of digestion of DNDF would be 0.050-0.055.
The large proportion (above 60%) of DM in the rumen was below the size (< 1.0 mm) which is considered to be the threshold size to particie passage (Poppi et al., 1980).Therefore, the reduction in particie size by comminution during eating or rumination does not appear to be the main rate limiting factor in clearing digesta from the reticulo-rumen (Ulyatt et al., 1986;Kennedy and Murphy, 1988).
The changes in chemical composition of particulate DM with size were broadly similar to those reported by Waghorn et al. (1986) and Jung et al. (1990) for lucernę diets.The difference in NDF content between the smallest particie size group and the other groups was, however, greater in our study.This may be due the smaller mesh size used in our study (0.04 mm) than in the studies of Jung et al. (1990) (0.15 mm) andWaghorn et al. (1986) (0.25 mm).Lower NDF content in the smali particles may partly be related to a greater attachment of bacteria to smali as compared with large particles (Legay-Carmier and Bauchart, 1989).An increasing N content in rumen particles with decreasing size (Waghorn et al., 1986) supports this.It is also possible that some rumen protozoa was included in the 0.04-mm fraction thereby decreasing NDF content.Rumen protozoa contains fairly large amounts of storage polysaccharides (Czerkawski, 1986) which may not be completely soluble in NDF detergent but soluble in ADF.The different composition of the 0.04-mm fraction may also be associated with the smaller particie size and lower C/HC-ratio in barley than in silage.The effect of time after feeding on the composition of particulate DM and NDF was smali when averaged across particie sizes, in agreement with Waghorn et al. (1986).However, the differences were greater when calculated from rumen NDF pool, e.g.C/NDF decreased from 429 to 385 mg g-i and HC/NDF increased from 523 to 556 mg g-1 at 3 and 12 h post-feeding.
The NDFD of digesta declined as the particie size decreased from > 3.0 mm to 0.2-0.5 mm.Similarly, Jung et al. (1990) reported a lower in vitro DM digestibility for smali than for large particles.This relationship may reflect the fermentation history of the particles.Higher NDFD of the 0.04-mm fraction in the present study may be related to a greater particie loss from the bags during the 12-d ruminal incubation.Although chewing during ingestion and rumination are of major importance in particie size reduction, particie size has also decreased during in situ incubation (Nocek and Kohn, 1988).Among the particie size groups ^ 0.2 mm, NDFD was influenced by the composition of NDF with ADL being the major determinant of NDFD.The C/NDF-ratio was positively and that of HC/NDF negatively correlated with NDFD.This is consistent with the greater rate of removal of C than that of HC from the rumen (0.058 vs. 0.038h-!).These observations agree with higher ruminal and total digestibility of C than that of HC in cattle given similar diets to that used in the present study (Jaakkola et al., 1991;Khalili and Huhtanen, 1991).
Significant time x particie size interaction in NDFD may be related to the changes in the age and composition of particles with time.The pool size of > 3.0 mm particles increases during eating relatively more than that of other fractions, and conseąuently NDFD increases more due to higher NDFD of feed than ruminal digesta (Jung et al,, 1990).On the other hand, later particie size reduction by rumination provides recently ingested feed particles from the 3-mm fraction to the other fractions thereby increasing their NDFD.
Although ingested plant materiał is rapidly colonized in the rumen (Cheng et al., 1984), particle-associated enzyme activities increased between 3-12 h after feeding, in agreement with Williams et al. (1989).They observed that the number of microorganisms attached to digesta particles were similar at 2 and 20 h after feeding, and the increases in enzyme activity did not occur as a result of increased population but were due to increased activities in an attached population.Greater diurnal variation in enzyme activities, NDFD and also NDF pool size in the 3-mm fraction than in the 0.5-and 0.2-mm fractions indicates that most of ingested forage enters this fraction, and that particie breakdown proceeds rapidly.On the other hand, the amounts of NDF entering the smali particie pools from large particles by comminution during rumination, and the amounts leaving by digestion and passage were almost eąual even although the animals were fed twice daily.
The higher enzyme activities in the 0.04-mm fraction than in the other fractions may relate to the increased surface area.Bacterial.attachment is proportional to the surface area exposed to attack (King, 1966).It is also possible that rumen protozoa included in this franction had celluloytic activity (Coleman, 1985).The positive correlation between the enzyme activities and the content of ND solubles support the conclusion of Kennedy and Milligan (1988) that the latter in smali, water-extracted particles might approximate DM of adherent microbes.Except for the 0.04-mm fraction, the content of ND solubles in error of means Significance: t(P < 0.10); *(P < 0.05); **(P < 0.01); ***(P < 0.001) error of means Significance: t(P < 0.10); *(P < 0.05); **(P < 0.01); ***(P < 0.001)

TABLE 3
The effect of time after feeding on distribution of NDF in various fractions

TABLE 4
Particle-associated enzyme activities (/rniol/g DM/min) in various particle size groups