Scanning Electron Microscope Study of Bacteria Associated with the Rumen Epithelium of Sheep

Examination of the rumen epithelium of sheep by scanning electron microscopy revealed bacteria associated with the epithelial surface. Comparison of epithelial surfaces from 10 sheep revealed areas that were consistently densely covered with bacteria and other areas where the cover was consistently light. The bacterial populations were frequently of mixed morphological types, but areas populated with a single type were also observed. This finding, together with the discovery of bacterial forms not previously described in rumen contents, suggests that a specific flora may exist on the rumen epithelial surface. The functional significance of such a population is discussed.     

Adhesion of bacteria to epithelial cell surfaces within the reticulo-rumen of cattle

Blocks of tissue were removed from various locations in the bovine digestive tract and fixed and processed for transmission and scanning electron microscopy by techniques that retained adherent bacteria. The distribution of bacteria on the surface of epithelial cells was examined by scanning electron microscopy. This showed intermittent colonization of the epithelia with the formation of occasional microcolonies of morphologically similar bacterial cells. Transmission electron microscopy of ruthenium red-stained material showed the presence of both the glycocalyx of the bovine epithelial cells and fibrous carbohydrate coats surrounding adherent bacteria. The carbohydrate coats appeared to mediate the attachment of bacteria to the epithelium, to food particles, and to each other so that microcolonies were formed. Careful examination of the bacterial colonization of keratinized cells in the process of being sloughed from the surface of the stratified squamous epithelium of the rumen showed that these dead cells were digested by adherent bacteria of a limited number of morphological types. The spatial relationship of this mixed, adherent, microbial population to living and dead epithelial cells and to food particles indicates that digestive processes of some importance may be accomplished by this stationary component of the microbial flora of the digestive tract.     

Adhesion of bacteria to epithelial cell surfaces within the reticulo-rumen of cattle.

Blocks of tissue were removed from various locations in the bovine digestive tract and fixed and processed for transmission and scanning electron microscopy by techniques that retained adherent bacteria. The distribution of bacteria on the surface of epithelial cells was examined by scanning electron microscopy. This showed intermittent colonization of the epithelia with the formation of occasional microcolonies of morphologically similar bacterial cells. Transmission electron microscopy of ruthenium red-stained material showed the presence of both the glycocalyx of the bovine epithelial cells and fibrous carbohydrate coats surrounding adherent bacteria. The carbohydrate coats appeared to mediate the attachment of bacteria to the epithelium, to food particles, and to each other so that microcolonies were formed. Careful examination of the bacterial colonization of keratinized cells in the process of being sloughed from the surface of the stratified squamous epithelium of the rumen showed that these dead cells were digested by adherent bacteria of a limited number of morphological types. The spatial relationship of this mixed, adherent, microbial population to living and dead epithelial cells and to food particles indicates that digestive processes of some importance may be accomplished by this stationary component of the microbial flora of the digestive tract.     

Factors influencing bacterial attachment to the ruminal epithelium in vitro

Some factors influencing bacterial attachment to the rumen epithelium were studied in vitro using mixed rumen bacteria (upper- or lower-layer bacteria formed at bacterial sediment by centrifugation), isolated from steers fed a roughage diet, and rumen epithelial cells collected from beef cattle given low-concentrate (50%; LC) and high-concentrate (90%; HC) diets. Optimal incubation conditions for bacterial attachment to rumen epithelial cells were 39 degrees C for 30 min. The bacteria isolated from the upper layer had a higher attaching activity to the LC epithelial cells than those of the lower layer. A higher degree of bacterial attachment was observed using the rumen epithelium from steers fed the LC diet rather than the HC diet (p<0.01). Ethylenediamine dihydroiodide (EDDI) added at 10 through 40 &mgr;g/ml increased bacterial attachment to the HC epithelial cells. Ammonia at 50 through 100 &mgr;g/ml positively affected bacterial attachment to both LC and HC epithelial cells. Bacterial attachment to the HC epithelial cells was enhanced (p<0.01) by the addition of a reducing agent (L-cysteine.HCl) but no increase was noted with LC cells. L- or D-lactate, volatile and unsaturated fatty acids markedly decreased bacterial attachment to rumen epithelial cells.     

Urease activity of adherent bacteria in the sheep rumen

In experiments on six sheep fed on a low protein diet (6.2 g N/day), it was found that the urease activity of the rumen fluid did not change significantly in the first 6 hours after feeding and that it ranged from 45 to 75 nkat.ml-1. The major portion was bound to the bacterial fraction and formed about 70% of total rumen fluid activity. Urease activity determined in food particles with adherent bacteria removed from the rumen before and 3 and 6 hours after feeding ranged from 20 to 26 nkat.g-1 food (wet weight), and on rumen wall samples with adherent bacteria from 30 to 800 nkat per 2.5 cm2 tissue. Again, no significant changes correlated to the time after feeding were found. The results show that urease activity in the sheep rumen is localized on food particles and on rumen wall epithelium with adherent bacteria, as well as in the rumen fluid.     

Production of alkaline phosphatase by epithelial cells and adherent bacteria of the bovine rumen and abomasum

Three distinct bacterial populations have been defined in the bovine rumen: the rumen fluid population; the population associated with food particles; and the population adherent to the rumen epithelium. Alkaline phosphatase activity has been reported in cells of the first two populations and here we report that assays of rumen epithelial samples containing both tissue and bacteria also contain the enzyme. The reaction product technique has localized the enzyme both in adherent bacteria and in cell of the stratified squamous rumen epithelium. The epithelium of the abomasum shows much lower levels of alkaline phosphatase activity.     

Medium Without Rumen Fluid for Nonselective Enumeration and Isolation of Rumen Bacteria

Colony counts which approximated those in a habitat-simulating, rumen fluid-agar medium (RFM) were obtained in medium 10, a medium identical to the RFM except for the replacement of rumen fluid with 1.5 x 10(-6)m hemin, 0.2% Trypticase, 0.05% yeast extract, and a 6.6 x 10(-2)m volatile fatty acid mixture qualitatively and quantitatively similar to that in rumen fluid. Single deletion of Trypticase, yeast extract, or the volatile fatty acid mixture from medium 10 significantly reduced colony counts. Colony counts were also reduced when medium 10 was modified to contain higher concentrations of Trypticase or volatile fatty acids. Significant differences were found between colony counts obtained from diluted rumen contents of animals fed a cracked corn-urea diet, and the colony counts obtained from animals fed either a cracked corn-soyean oil meal or an alfalfa hay-grain diet. Qualitative differences were found between the predominant bacterial strains isolated from rumen contents of animals fed cracked corn diets and strains isolated from animals fed alfalfa hay-grain. Regardless of differences in the predominant flora associated with diet, medium 10 and the RFM supported growth of similar bacterial populations. The results show that medium 10 is suitable for enumeration and isolation of many predominant rumen bacteria.     

Urease activity of adherent bacteria and rumen fluid bacteria

In experiments on six sheep fed on a low nitrogen diet (3.7 g N/day), urease (EC 3.5.1.5) activity (nkat X mg-1 bacterial dry weight) 3 h after feeding was found to be highest in the bacteria adhering to the rumen wall (13.25 +/- 2.10), lower in the rumen fluid bacteria (8.96 +/- 1.35) and lowest in the bacteria adhering to feed particles in the rumen (5.69 +/- 2.13). The urease activity of bacteria adhering to the rumen wall and of the rumen fluid bacteria of six sheep fed on a high nitrogen diet (21 g N/day) was significantly lower than in sheep with a low N intake and in both cases was roughly the same (3.81 +/- 1.37 and 3.76 +/- 1.02 respectively); it was lowest in bacteria adhering to feed particles in the rumen (1.92 +/- 0.90). It is concluded from the results that the urease activity of rumen fluid bacteria and of bacteria adhering to the rumen wall and to feed particles in the rumen is different and that it falls significantly in the presence of a high nitrogen intake. From the relatively high ureolytic activity of bacteria adhering to the rumen wall in the presence of a low nitrogen intake it is assumed that this is one of the partial mechanisms of the hydrolysis of blood urea entering the rumen across the rumen wall and of its reutilization in the rumen-liver nitrogen cycle in ruminants.     

Isolation and Presumptive Identification of Adherent Epithelial Bacteria (“Epimural” Bacteria) from the Ovine Rumen Wall

One hundred sixty-one strains of adherent bacteria were isolated under anaerobic conditions from four sites on the rumen epithelial surface of sheep fed hay or a hay-grain ration. Before isolation of bacteria, rumen tissue was washed six times in an anaerobic dilution solution, and viable bacteria suspended in the washings were counted. Calculation indicated that unattached bacteria would have been removed from the tissue by this procedure, but a slow and progressive release of attached bacteria also occurred. Nevertheless, a wide range of characteristic morphological types remained associated with the epithelium as demonstrated by scanning electron microscopy. Most of these types were represented among the isolates. Characterization and presumptive identification of the isolates showed that 95.0% belonged to previously described genera of functionally significant rumen bacteria, including Butyrivibrio sp. (31.1%), Bacteroides sp. (22.4%), Selenomonas ruminantium (9.9%), Succinivibrio dextrinosolvens (8.7%), Streptococcus bovis (8.1%), Propionibacterium sp. (4.3%), Treponema sp. (3.1%), and Eubacterium sp., Lachnospira multiparus, and Ruminococcus flavefaciens (2.5% each). Eight isolates (5.0%) were not identified. L. multiparus was recovered only from hay-fed animals; all other genera were obtained from animals fed either ration. All S. bovis strains and two strains each of Bacteroides sp. and Butyrivibrio sp. were aerotolerant; all other strains were strictly anaerobic. Bacteria representing the gram-positive, facultatively anaerobic flora associated with rumen wall tissue (R. J. Wallace, K.-J. Cheng, D. Dinsdale, and E. R. Ørskov, Nature (London) 279:424-426, 1979) were therefore not recovered by the techniques used; instead a different fraction of the adherent population was isolated. The term “epimural” is proposed to describe the flora associated with the rumen epithelium.     

Degradation of mimosine, 2,3-dihydroxy pyridine and 3-hydroxy-4(1H)-pyridine by bacteria from the rumen of sheep in Venezuela

Abstract The addition of Leucaena leucocephala herbage did not diet of sheep in Venezuela did not affect the concentration of volatile fatty acids (VFA) in the rumen, the degradation of rice straw incubated in sacco, or the numbers of rumen fungi or bacteria. However, feeding Leucaena increased the concentration of ammonia in the rumen. In addition, two products of the degradation of the toxic amino acid mimosine were detected in the rumen when Leucaena was fed. One of these products, 2,3-dihydroxy pyridine (2,3-DHP), was detected at concentrations of up to 1.1 μmol/ml. The other, 3-hydroxy-4(1H)-pyridone (3,4-DHP) was found at concentrations of up to 0.96 μmol/ml. The examination of bacterial cultures isolated from the rumen of the sheep under investigation showed that feeding Leucaena increased the relative proportions of short Gram-negative rods and decreased the proportion of long roads and coccobacilli present. Although the animals fed Leucaena showed a small loss in weight during the feeding trial, no evidence of Leucaena toxicity was seen. A total of 18 cultures capable of degrading 2,3-DHP or 3,4-DHP were isolated from the rumen of the sheep before Leucaena was fed. These included both Gram-positive and Gram-negative bacteria, and a Gram-positive sporeformer. It seems that 2,3-DHP and 3,4-DHP may be degraded by a much wider range of bacteria than has been recognised previously. The detection of these bacteria before Leucaena was fed suggests that they were indigenous members of the rumen microflora of sheep in Venezuela.     

The effect of Aspergillus oryzae fermentation extract on the growth of fungi and ciliate protozoa in the rumen

When added to the diet of sheep, 2 g/d, Aspergillus oryzae fermentation extract (AO) stimulated total and cellulolytic bacterial numbers in rumen fluid by 34 and 90% respectively. AO had no effect on the numbers of protozoa or fungal zoospores. AO did not affect hydrogen production by the rumen fungi Neocallimastix frontalis (RE1), N. patriciarum (CX) or Piromonas communis (P) in pure culture or protozoal activity in vitro , estimated from the rate of breakdown of [¹⁴C] leucine-labelled Selenomonas ruminantium. It was concluded that increases in ruminal fibre digestion observed previously in animals fed AO, were most likely due to a stimulation of bacteria rather than eukaryotes in the rumen microbial population.     

Rumen bacterial and fungal degradation of Digitaria pentzii grown with or without sulfur.

Sheep fed the forage Digitaria pentzii fertilized with sulfur were compared with those fed unfertilized forage for the rumen microbial population involved with fiber degradation. No differences were detected in the bacterial population as determined by anaerobic cultures on a habitat-simulating medium, xylan, or pectin, by 35S labeling techniques for microbial protein, or by transmission electron microscopic studies of bacterium-fiber interactions. Rumen volume and water flow from the rumen were not different for sheep fed each of the forages. Rumen fungi were prevalent in sheep fed sulfur-fertilized D. pentzii as shown by sporangia adhering to forage fiber and by colonies developing from zoospores in roll tubes with cellobiose plus streptomycin and penicillin. Fungi were absent or in extremely small numbers in sheep fed unfertilized forage. Nylon bag digestibility studies showed that the fungi preferentially colonized the lignified cells of blade sclerenchyma by 6 h and caused extensive degradation by 24 h. In the absence of bacteria in in vitro studies, extensive hyphal development occurred; other lignified tissues in blades (i.e., mestome sheath and xylem) were attacked, resulting in a residue with partially degraded and weakened cell walls. Nonlignified tissues were also degraded. Breaking force tests of leaf blades incubated in vitro with penicillin and streptomycin and rumen fluid from sheep fed sulfur-fertilized forage or within nylon bags in such sheep showed a residue at least twice as fragile as that from sheep fed unfertilized forage. In vitro tests for dry matter loss showed that rumen fungi, in the absence of actively growing bacteria, could remove about 62% of the forage material. The response of rumen fungi in sheep fed sulfur-fertilized D. pentzii afforded a useful in vivo test to study the role of these microbes in fiber degradation. Our data establish that rumen fungi can be significant degraders of fiber and further establish a unique role for them in attacking and weakening lignocellulosic tissues. The more fragile residues resulting from attack by fungi could explain the greater intake consistently observed by sheep eating sulfur-fertilized compared with unfertilized D. pentzii forage.     

The Effect of Rumen and Post-Rumen Feeding of Carbohydrates on the Caecal Microflora of Sheep

S ummary . When sheep were bottle-fed with solutions of glucose, sucrose or maltose, so avoiding fermentation in the rumen, there was a marked increase in the number of caecal bacteria, with a higher percentage of viable bacteria, compared with caecal counts from sheep fed conventionally on carbohydrates. Bottle-feeding also resulted in a redistribution of bacterial types, 40–46% of the flora comprising Gram positive rods and a significant increase in butyric acid and decrease in propionic acid in the caecal contents. Feeding cellulose suspensions by bottle or conventionally failed to produce significant differences in counts of cellulolytic caecal bacteria. When sucrose was infused in increasing amounts into the abomasum, Streptococcus bovis was the principal organism, isolated from faeces.     

Lotus corniculatus condensed tannins decrease in vivo populations of proteolytic bacteria and affect nitrogen metabolism in the rumen of sheep

Condensed tannins in forage legumes improve the nutrition of sheep by reducing ruminal degradation of plant protein and increasing crude protein flow to the intestine. However, the effects of condensed tannins in forage legumes on rumen bacterial populations in vivo are poorly understood. The aim of this study was to investigate the specific effects of condensed tannins from Lotus corniculatus on four proteolytic rumen bacteria in sheep during and after transition from a ryegrass (Lolium perenne)-white clover (Trifolium repens) diet (i.e., low condensed tannins) to a Lotus corniculatus diet (i.e., higher condensed tannins). The bacterial populations were quantified using a competitive polymerase chain reaction. Lotus corniculatus was fed with or without ruminal infusions of polyethylene glycol (PEG), which binds to and inactivates condensed tannins, enabling the effect of condensed tannins on bacterial populations to be examined. When sheep fed on ryegrass-white clover, populations of Clostridium proteoclasticum B316T, Butyrivibrio fibrisolvens C211a, Eubacterium sp. C12b, and Streptococcus bovis B315 were 1.5 x 10(8), 1.1 x 10(6), 4.6 x 10(8), and 7.1 x 10(6) mL(-1), respectively. When the diet was changed to Lotus corniculatus, the average populations (after 8-120 h) of C. proteoclasticum, B. fibrisolvens, Eubacterium sp., and S. bovis decreased (P < 0.001) to 2.4 x 10(7), 1.1 x 10(5), 1.1 x 10(8), and 2.5 x 10(5) mL(-1), respectively. When PEG was infused into the rumen of sheep fed Lotus corniculatus, the populations of C. proteoclasticum, B. fibrisolvens, Eubacterium sp., and S. bovis were higher (P < 0.01-0.001) than in sheep fed Lotus corniculatus without the PEG infusion, with average populations (after 8-120 h) of 4.9 x 10(7), 3.8 x 10(5), 1.9 x 10(8), and 1.0 x 10(6), respectively. Sheep fed the Lotus corniculatus diet had lower rumen proteinase activity, ammonia, and soluble nitrogen (P < 0.05-0.001) than sheep that were fed Lotus corniculatus plus PEG. The Lotus corniculatus diet reduced rumen nitrogen digestibility (P < 0.05) and ammonia pool size and increased the flow of undegraded feed nitrogen to the abomasum. The nitrogen intake, rumen non-ammonia nitrogen pool size, rumen microbial non-ammonia nitrogen pool size, and abomasal microbial non-ammonia nitrogen fluxes were similar both in sheep fed only Lotus corniculatus and in sheep fed Lotus corniculatus plus PEG, but nonmicrobial non-ammonia nitrogen flux to the abomasum was higher (P < 0.01) for the sheep fed only Lotus corniculatus. Although condensed tannins in Lotus corniculatus reduced the populations of some proteolytic bacteria, total ruminal microbial protein and microbial protein outflow to the abomasum were unchanged, suggesting a species-specific effect of condensed tannins on bacteria in the rumen.     

PCR-DGGE analysis reveals a distinct diversity in the bacterial population attached to the rumen epithelium

Bacteria attached to the rumen epithelium (or epimural community) are not well characterised and their role in rumen functioning is not totally understood. There is just one published report of a clone library from one cow that suggests that this epimural community differs from the bacteria associated with the rumen digestive contents. However, this time-consuming approach is not adapted for examining microbial population changes in groups of animals. In in vivo studies, when samples from several animals have to be analysed simultaneously, a simpler technique has to be used. In this study, a genetic fingerprinting technique, polymerase chain reaction-denaturing gradient gel electrophoresis (PCR-DGGE), was used to characterise the structure of the bacterial population attached to the rumen epithelium. This community was compared with that present in the solid and liquid phases of rumen content under two contrasting diets. Rumen samples were obtained from four forage-fed and four high-concentrate-fed (80 : 20, wheat grain : hay) 5-month-old lambs. After slaughter, samples from five epithelial sites and the solid and liquid digesta phases were taken for DNA extraction and analysis. Bacterial communities were profiled by PCR-DGGE using bacterial-specific 16S rDNA primers. Analysis of the fingerprint revealed that the epithelial community differed from those of rumen content in both diets. As expected, the nature of the feed influenced the bacterial communities from the solid and liquid rumen phases but no diet effect was observed in the rumen epithelial profiles suggesting a strong host effect on this bacterial population. Additionally, no differences were observed among the five epithelial sampling sites taken from each animal. The profile of the bacterial population attached to the rumen epithelium presented a high inter-animal variation, whether this difference has an influence in the function of this community remains to be determined.     

Effects of an Abrupt Change in Ration from All Roughage to High Concentrate upon Rumen Microbial Numbers in Sheep

When three sheep were abruptly changed from a ration of 100% orchardgrass hay to 60% cracked corn-40% orchardgrass hay, fed at equal dry-matter intakes, significant increases in concentration were observed in the rumen microbial population. Bacterial numbers (colony counts) per gram of rumen contents did not appear to have stabilized within 21 days after the ration change; however, protozoan numbers per milliliter plateaued after 5 days. The concentration of cellulose-digesting bacteria varied considerably between animals and decreased in all animals with the change. Changes were observed in total and molar percentages of volatile fatty acids, which were typical for the two types of rations. Although the concentration of protozoa increased after the ration change, only minor differences were observed in their percent generic distribution. A significant decrease in rumen volume was measured in two of the three sheep with the change in ration; however, fluid turnover rates were not significantly affected. Rates of rumen dry-matter turnover were slower with the concentrate ration, although rumen dry-matter digestion was increased. Calculation of total bacterial numbers based on total rumen volume completely negated the effect of ration change in one animal, whereas total numbers in the other two animals were still significantly different between rations and very similar between animals. Adjustment of total protozoa numbers did not alter the trends seen previously with concentration values.     

Effect of Dietary Extremes on Impala (Aepyceros melampus) Rumen Epimural Flora

The impala, classified as an intermediate feeder among wild ruminants, grazes exclusively on green grasses during the summer rains but turns to an essentially browse habit during the winter, when a diet of leaves, pods, and twigs is supplemented with dried grasses. Scanning electron microscope studies of bacteria attached to the rumen epithelia of animals feeding on digestible, carbohydrate-rich grazing diet showed a high level of surface colonization, with pockets of mixed morphological types occupying spaces vacated by desquamated epithelial cells. Rumen wall colonization in browsing animals on an indigestible, abrasive, low-plane diet showed denudation of distal papillar surfaces with sparse homogeneous adherent flora in the microprotected areas towards the bases. Forage characteristics associated with raised volatile fatty acid concentrations in ingesta, higher rate of epithelial sloughing, and minimum abrasion were interrelated with density and complexity of the adherent bacterial flora.     

An investigation of the relationship between fecal and rumen bacterial concentrations in sheep

With no acceptable method for collecting fresh rumen fluid from zoo ruminants, it was proposed that fecal bacterial concentrations may be correlated with rumen bacteria. If so, fecal bacterial concentrations could be used to study both the effects of diet on rumen bacteria as well as rumen abnormalities. Total and cellulolytic bacterial concentrations were determined in whole rumen contents and feces of sheep using a most‐probable‐number (MPN) assay. In a Latin square design, four crossbred ewes were fed diets of 100% long or chopped orchardgrass hay (OH) and 60% ground or whole shelled corn plus 40% chopped OH. In a second trial, the sheep were fed a pelleted complete feed at varying levels of intake i.e., control at 2.0% of body weight and at 1.8, 1.6, and 1.2% of body weight. Higher total rumen bacterial concentrations (P<0.01) were found on the high concentrate diets as compared with the high forage diets. Grinding the corn also increased total bacterial concentrations (P<0.05). Fecal concentrations of total bacteria were higher (P<0.01) with the high concentrate diets. Chopping the forage decreased the concentration of fecal cellulolytic bacteria (P<0.05) but had no effect on their concentration in the rumen. An inverse linear relationship (P<0.01) was observed between total bacterial concentrations in the feces and diet intake. Although relationships were observed between the rumen and feces for total and cellulolytic bacterial concentrations, they were dependent on diet, particle size, and level of intake. Thus, fecal bacterial concentrations cannot be used to reliably predict rumen bacterial concentrations. Zoo Biol 27:100–108, 2008. © 2008 Wiley‐Liss, Inc.     

Comparative Analysis of Microbial Profiles in Cow Rumen Fed with Different Dietary Fiber by Tagged 16S rRNA Gene Pyrosequencing

The ruminal microbiome of cattle plays an important role not only in animal health and productivity but also in food safety and environment. Microbial profiles of rumen fluid obtained from dairy cows fed on three different fiber/starch diet compositions were characterized. Tagged 16S rRNA gene pyrosequencing and statistical analysis revealed that the dominant ruminal bacteria shared by all three sample groups belonged to phyla Bacteroidetes, Firmicutes, and Proteobacteria. However, the relative abundance of these bacterial groups was markedly affected by diet composition. In animals fed with a high fiber diet, the fibrolytic and cellulolytic bacteria Lachnospiraceae, Ruminococcaceae, and Fibrobacteraceae were found in highest abundance compared with animals fed other diets with lower fiber content. The polysaccharide-degrading Prevotellaceae and Flavobacteriaceae bacteria were most abundant in the rumen of cows fed on diet with the highest starch content. These data highlight the ruminal microbiome’s ability to adapt to feed composition and also provide a basis for the development of feed formulation systems designed to improve livestock productivity.     

Evidence for existence of lectins on the ruminal bacteria from steers fed roughage and concentrate diets

Mixed rumen bacteria, isolated by centrifugation from the rumen of steers fed a roughage (R) or concentrate (C) diet, were used to determine if lectins are present on rumen bacteria, based on hemagglutination (HA) and HA inhibition assays in vitro. Rumen bacteria from steers fed either diet agglutinated erythrocytes from cattle, sheep, pigs, and rats, suggested that lectins exist on rumen bacteria. Bacterial HA titers from steers receiving the R diet were much higher (p<0.001) than those from steers fed the C diet, depending on the erythrocyte source used. Centrifugation at 20,000xg for 30 min fractionated the rumen bacteria into upper (U) and lower (L) layers. The HA titers of the U bacterial fractions were significantly higher (p<0.001) than those of the L fractions. A remarkable reduction or complete disappearance of HA titers following treatment of rumen bacteria with protease, trypsin, sodium dodecyl sulfate (SDS) or sodium periodate indicates that rumen bacterial lectins are probably glycoproteins. Lectin specificity for saccharides (galactose, lactose, N-acetyl-D-galactosamine, methyl-alpha-D-galactopyranoside and methyl-beta-D-galactopyranoside) and glycoproteins (mucin, fetuin, and thyroglobulin) was found in the RU, RL, and CU bacterial fractions; no specific binding was determined in the CL fractions. The potential role of lectins in mediating the attachment of rumen bacteria to feed particles, rumen epithelia and other microorganisms is discussed.