| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Ohio State University Interdisciplinary Nutrition Program (OSUN), and
2 Department of Animal Sciences, The Ohio State University, Columbus 43210
3 The Institute of Rural Sciences, University of Wales Aberystwyth, Llanbadarn Fawr, Aberystwyth, UK
Corresponding author: J. L. Firkins; e-mail: firkins.1{at}osu.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Key Words: rumen protozoal nitrogen • real-time PCR • denaturing gradient gel electrophoresis • rRNA
Abbreviation key: DGGE = denaturing gradient gel electrophoresis, DNDF = potentially digestible NDF, HF = high forage diet, LF = low forage diet, rDNA = gene encoding ribosomal RNA.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Despite the importance and need for more research quantifying rumen pool size and duodenal flow of protozoal N, research in this area is hindered by the lack of an accurate protozoal marker (Firkins et al., 1998). Although protozoal N flow to the duodenum has been approximated to be about one-half that of the rumen pool (Jouany, 1996), comprising 20 to 40% of the total microbial N in the duodenum, most digesta flow studies ignore protozoal marker:N ratios, which can significantly decrease accuracy (Broderick and Merchen, 1992). Intentional ignoring of protozoa has been inappropriately justified based on results using indirect procedures, the combination of which appears to compound to underestimate duodenal flow of protozoal N (Firkins et al., 1987a). In contrast, indirect methods not relying on these potentially biased procedures tend to provide generally higher protozoal N flows (Steinhour et al., 1982; Shabi et al., 2000).
To assess effects of protozoal ecology on ruminal N metabolism, many researchers have counted total protozoa or have divided them into isotrichids and entodiniomorphids by their cilia pattern. However, protozoa vary considerably in biomass per cell (Williams and Coleman, 1992), so counts should be converted to a biomass basis for nutritional purposes. Although protozoal cells lyse in the abomasum, direct quantification of omasal flow of protozoal cells and N can be done (Ahvenjärvi et al., 2002), but these workers noted potentially significant underestimation of protozoal N because of potential efflux of very small protozoa through the pores of their filters. Besides capturing protozoa for direct quantification purposes, filtration has been used to provide a more representative sample of protozoa to be used for marker studies (Martin et al., 1994). Classical approaches used in many published reports used gravitational separation of protozoal cells from flocculating ruminal fluid primarily using separatory funnels (Firkins et al., 1987a; Shabi et al., 2000) or centrifugation (Punia and Leibholz, 1994; Ivan et al., 2000). These procedures could have significant amounts of contamination of bacterial matter (Sharp et al., 1998; Volden et al., 1999; Sylvester et al., 2004), so the accuracy of a direct protozoal marker could be decreased by the procedures used to harvest a reference protozoal fraction. This reference should accurately represent the protozoal populations in the rumen pool while minimizing contamination by plant matter and bacteria. Finally, even a more highly purified protozoal sample collected from the rumen might not represent the protozoa flowing to the duodenum because of possible selective retention of certain genera (Leng and Nolan, 1984; Dehority, 2003) and potentially different marker:N ratios among those populations.
Sylvester et al. (2004) recently developed an assay to measure protozoal biomass in ruminal and duodenal digesta by targeting and quantifying the DNA encoding 18S rRNA (rDNA). The assay includes a procedure with steps to concentrate protozoal cells by filtration while reducing contamination of nonassociated bacteria (i.e., bacteria not attached externally to or living inside protozoa). In experiment 1, our objective was to determine the degree of reduction of bacterial contamination in the protozoal standard derived by filtration compared with a traditional sedimentation procedure. In experiment 2, our objectives were to determine if protozoal standards representing predominant protozoal genera in ruminal fluid are appropriate for predicting protozoal N flows to the duodenum and to evaluate data obtained from our real-time PCR assay predicting protozoal N pool size in the rumen and flowing to the duodenum in lactating cows fed typical dairy diets compared with measurements from independent procedures.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
The samples were mixed and separated into equal volumes. Each volume was randomly assigned to either the sedimentation technique (Punia and Leibholz, 1994) or a revised filtration procedure described by Sylvester et al. (2004). Protozoa-rich fractions (2- and 5-h samples) were composited by cow and suspended in 30 mL of 0.9% saline. Subsamples (10 mL) of each protozoal sample were stored at 4°C for bacterial and protozoal enumeration, and the remaining 20 mL was lyophilized for N and RNA determination. Protozoal-free fluid from each protozoal harvesting method was retained and pooled; bacteria were collected by centrifuging at 14,000 x g for 25 min and were split into 2 equal and representative volumes. One half was lyophilized for subsequent chemical analyses, and the other half was stored at 4°C for bacterial enumeration.
Protozoal cell enumeration was done using the method outlined by Dehority (1993). Bacterial cell enumeration was done using epifluorescence microscopy on an Olympus BX 21 upright epifluorescent microscope (Melville, NY). To quantify bacterial contamination of harvested protozoa samples, we assumed that 1) bacteria that are associated with protozoa should be considered as an integral part of the protozoal fraction, and 2) dilutions used to count the bacteria consisted of nonadherent bacterial cells only (i.e., all protozoal cells and their associated bacteria were diluted to extinction). Protozoal cell samples harvested from each protozoal collection method were serially diluted, and the numbers of bacterial cells contaminating them were counted using the method described by Yu et al. (1995), with slight modifications. Briefly, samples were incubated overnight with Triton X-100 (0.1%, final concentration) to unclump bacterial cells that were stained with 4',6-diamidino-2-phenylindole, and quantitatively passed through a 0.2-µm polycarbonate filter. Filters were air-dried and mounted on slides, coverslips were mounted using nail polish at each corner, and slides were stored in the dark until enumeration was completed. Additional modifications included: 1) no addition of NaCl or Na4P2O7 because protozoal suspension includes NaCl, and 2) bacterial cells present within 40 random grids were counted instead of 8 to 20 grids. Bacterial cells were determined randomly from each sample using 2 replicate dilution series and 2 duplicate filter counts completed on 2 d.
Experiment 2
Feeding conditions.
Two multiparous Holstein cows (identified as 647 and 660), which were fitted with rumen and simple T duodenal cannulas anterior to the pancreatic duct, were used in a randomized complete block design to which the cows were considered the blocking criterion (Sylvester et al., 2004). Cows averaged 773 kg of BW, 24.5 kg/d DMI, and 35.1 kg/d of milk during the experiment. Cows were fed 1 of 2 control diets containing either 21 or 16% forage NDF [high forage (HF) and low forage (LF) respectively], which were part of a larger study (Lima et al., 2003). Diets were prepared once daily as a TMR, and animals were fed diets at 110% of ad libitum at 0600 and 1800 h for 14-d periods. Daily samples of feed offered and refused were composited by cow within sample type and dried at 55°C. Pellets containing 5% Cr2O3 and 95% soybean hulls were dosed into the rumen twice daily through the ruminal cannula at 1.1% of the diet DM starting on d 5 for use as a digesta flow marker as described previously (Harvatine et al., 2002a). Pellets were subsampled weekly, dried at 55°C, and composited by period.
Digesta sampling.
Ruminal fluid was collected (approximately 1000 mL), as described previously, on d 8 through 11 of each period at 0700, 0900, 1200, 1500, and 1700 h, and then pooled to represent a 24-h feeding cycle. Subsamples (50 mL) from each time point were treated with formalin (1% wt/vol final concentration) and pooled for use in DNA extraction, and the remaining 950 mL was used for protozoal harvesting. Protozoal collection, enumeration, and generic identification were completed as described for experiment 1.
A 200-mL solution of CoEDTA (Uden et al., 1980) was pulse dosed via the rumen cannula on d 11 of each period and mixed thoroughly. Ruminal fluid was collected (1000 mL) as described previously at 0, 20, and 40 min and at 1, 2, 4, 6, 9, 12, 18, 24, and 36 h postdosing; 50-mL aliquots were taken and frozen at –20°C for CoEDTA determination, and the remainder was returned to the rumen.
Duodenal samples (250 mL) were taken on d 9 to 12 of each period so that every 90-min interval in a 24-h period were represented (16 samples total), combined by cow within period, and frozen immediately for subsequent analyses.
Total ruminal digesta were quantitatively evacuated as previously described by Harvatine et al. (2002b) by fractionating the fluids from the solids at 0400 h (d 13) and 0800 h (d 14) of each period, before weighing, subsampling, and reconstitution. After the last evacuation, half of the respective cow’s ruminal content was mixed with half of the content from the cow on the dietary treatment to which that cow was being switched. Cows were immediately fed their new diets after return of the ruminal contents. An extra week of adaptation was provided for the first period. Although the adjustment period for the remainder of periods was only 8 d (the last day of the previous period plus 7 d of the subsequent period), we assumed that this mixing of ruminal contents would allow a fast enough adaptation of microbial populations between sampling periods. Bacteria (Klieve et al., 2003) and protozoa (Soliva et al., 2004) stabilized after about 1 wk of adaptation to a change in dietary conditions that was much harsher than ours was. Logistical limitations prevented us from having longer adaptation periods. Subsamples of evacuated ruminal contents were reconstituted and stored at –20°C for ruminal pool size measurements as described subsequently. In addition, approximately 2 kg of these samples was used for harvesting particle-associated bacteria by blending (Callison et al., 2001), and the remaining 2 kg was dried at 55°C to estimate ruminal DM pool size (55°C DM x total rumen wet weight) and ruminal fluid volume (total weight – DM pool size).
Laboratory analyses.
All dried and lyophilized digesta samples were ground to pass through a 2-mm screen using a Wiley mill (Arthur H. Thomas, Philadelphia, PA). Lyophilized bacteria and protozoal cells were ground using a mortar and pestle. All samples were analyzed for DM and OM according to AOAC (1990). Chromic oxide pellets and duodenal samples were analyzed for Cr by atomic absorption spectroscopy (Williams et al., 1962). Rumen, duodenal, protozoal, and bacterial samples were analyzed for Kjeldahl N and total purines (Zinn and Owens, 1986) as modified by Ushida et al. (1985), with standardization to a total RNA basis. Indigestible NDF (10-d in vitro), potentially digestible NDF (DNDF), and Co were determined as described previously (Harvatine et al., 2002b).
Real-time PCR assay for protozoal N and electrophoresis analyses.
Ruminal fluid, duodenal digesta, and harvested protozoal cells were analyzed to determine the protozoal rDNA concentration using a real-time PCR assay developed by Sylvester et al. (2004). Briefly, total genomic DNA was extracted from triplicate 0.5-mL ruminal and duodenal digesta samples and from approximately 1 million protozoal cells harvested and purified from ruminal fluid and counted by microscopy. The extracted DNA of each respective sample was amplified using PCR to generate the corresponding standards and individual standard curves for each sample for the real-time PCR assay, which included correction factors for recovery of genomic DNA during extraction from the samples.
From the same triplicate DNA extractions, 25 µL of genomic DNA per sample was composited for use in denaturing gradient gel electrophoresis (DGGE) to compare banding profiles of the protozoal rDNA from the rumen vs. duodenum. Two independent ciliate-specific PCR primer sets (Table 1
) were used to amplify 2 separate regions near the 5' and 3' ends of the 18S rDNA. Amplification conditions and primer set 1 (1320f and 1617r) are from Regensbogenova et al. (2004) as follows: One cycle of 94°C for 4 min, annealing at 60°C for 30 s, and extension at 72°C for 1 min; 35 cycles of 94°C for 1 min, annealing at 60°C for 30 s, and extension at 72°C for 1 min; and one cycle of 94°C for 4 min, 60°C for 30 s, and 72°C for 10 min. Amplification conditions for primer set 2 (316f and 539r) are the same except that annealing temperature was lowered to 56°C. Primer set 2 was previously designed and verified for real-time PCR quantification of protozoal rDNA (Sylvester et al., 2004). The reverse primer (539r) was modified for DGGE analysis by adding a GC clamp (Table 1). Reaction volumes for both primer sets were 100 µL containing 100 pmol of each primer, 125 µM of each dNTP mixture, 2 mM magnesium chloride, 0.05% BSA, 1x PCR buffer, and 5.0 U of platinum Taq DNA polymerase (Invitrogen, Carlsbad, CA). Amplicons were electrophoresed on 1% agarose gel and visualized after staining with ethidium bromide. When multiple bands were observed using primer set 1 (1320f-1617r), bands of expected molecular size were excised and gel-purified before use in DGGE analysis.
|
Resulting bands were isolated from gels using a sterile pipette tip and transferred to sterile 0.5-mL tubes containing 30 µL of Tris-EDTA buffer. Tubes were subjected to 3 cycles of freezing at –80°C for 10 min followed by thawing in a water bath at 55°C for 15 min to allow diffusion of the amplicons. Amplicons were reamplified using primers and conditions previously described, purified, and sent for sequencing. The DNA sequences were checked for errors by manual evaluation of their respective electrophoretograms. Sequences were identified by comparing to GenBank (Benson et al., 2004) sequencing with the BLAST program (Altscul et al., 1990), and the highest identity to a known sequence was reported.
Calculations.
Protozoal standards of known volumes and concentrations of protozoal cells were lyophilized. After the mass of the dried residue was recorded, the concentrations of N, RNA, or rDNA copies could be quantitatively reconstituted to a cell number basis. The total ruminal pool of protozoal cell numbers was determined by multiplying cell concentrations x ruminal fluid volume, which was determined by rumen evacuation. This pool was multiplied by the concentrations of N, RNA, or rDNA copies per cell (using protozoal standards) to determine their respective ruminal pool sizes. Hereafter, these calculations will be referred to as the gravimetric method for independent comparison of results obtained using the real-time PCR assay. Ruminal pool size of protozoal N was calculated using the following 2 equations:
 | ([1]) |
 | ([2]) |
The ruminal pools of bacterial N were determined using the following 2 equations:
 | ([3]) |
 | ([4]) |
The average amount of Cr dosed daily was divided by the concentration of Cr in the duodenal samples to determine daily flow of DM to the duodenum. Ruminal fluid dilution rate was determined by nonlinear regression of CoEDTA dilution, and passage rate of DNDF was calculated as described previously (Harvatine et al., 2002b). These fractional rates were multiplied x 24 h/d to convert to a daily basis. Protozoal N flow to the duodenum (g/d) was determined using the following 3 equations:
 | ([5]) |
 | ([6]) |
 | ([7]) |
Bacterial N flow to the duodenum was calculated using the following equation:
 | ([8]) |
Statistical Analyses
Experiment 1.
Data were analyzed using Proc MIXED of SAS (SAS Institute, 1999) according to the following model:
 |
where Yij is the dependent, continuous variable; µ is the overall population mean; Mi is the fixed effect of the ith method (i = sedimentation or filtration); C is the fixed effect of the jth cow (j = 1, 2, 3); D is the fixed effect of the kth day (k = 1 or 2); Mi x Cj is the interaction between the ith method and the jth cow; and eij is the residual error, assumed independent and N(0,
).
Experiment 2.
Data were analyzed using Proc MIXED of SAS (SAS Institute, 1999) according to the following model:
 |
where Yijk is the dependent, continuous variable; µ is the overall population mean; Ti is the fixed effect of the ith dietary treatment (i = HF or LF); C is the fixed effect of the jth cow (j = 1 or 2); Mk is the fixed effect of the kth method (k = PCR or gravimetric); and eijk is the residual error, assumed independent. Variance homogeneity between different methods (i.e., N(0, )) was tested using the COVTEST option, with method repeated within cow across treatments and significance declared at P > 0.2. When variances between methods were deemed homogeneous, the repeated statement was removed, and SEM were pooled. Treatment means were separated using Fisher’s F-test-protected LSD, and significance was declared at P < 0.05 unless otherwise stated.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
), the sedimented protozoal sample appears to contain much more contamination with plant material than does the filtered protozoal sample. Such contamination might explain the relatively variable N:RNA ratios of protozoal preparations compared with bacterial samples (Firkins et al., 1987a; Volden et al., 1999).
|
). The net result is a 33-fold reduction in bacterial contamination using our filtration procedure.
|
Using calculations described in the footnotes of Table 2
, the 47.5 and 60.2% contribution of bacterial N and RNA in sedimented protozoal samples provides clear evidence that previous data using this or similar protozoal harvesting techniques could be profoundly in error. The higher values than the roughly 20% of total small subunit rRNA of bacterial origin (Sharp et al., 1998) reflect differences in the ratios of N:rRNA in bacteria vs. protozoa. Our values are also higher than in other studies because the current calculation is represented as a percentage of total microbial N, rather than as a percentage of total N (which includes other sources besides bacteria and protozoa), but also because we used the more pure filtration-derived protozoa containing higher N and RNA concentrations in the protozoal biomass. Although Martin et al. (1994) did not find detectable diaminopimelic acid in their filtered protozoal samples using the original filtration procedure, Volden et al. (1999) estimated bacterial N (again using diaminopimelic acid) to be up to 23% of the total N in the protozoal sample. Microscopic evaluations of the successive washes (data not shown) confirmed that our serial dialysis washing procedure (Sylvester et al., 2004) markedly reduced the contamination of bacteria not associated with protozoa compared with the original filtration method (Martin et al., 1994).
Experiment 2
Our objectives were to provide quantitative, independent measurements to evaluate our protozoal real-time PCR assay (Sylvester et al., 2004). Numerous samples were collected and analyzed for validation of various steps (data not shown), thus restricting our ability to sample only 2 dairy cattle fed different diets. A more complete analysis of microbial differences between those 2 diets would clearly take more animals but would also require more streamlined, routine procedures to measure protozoal N than could be accomplished in the current study. Statistical analyses were performed with the primary objective to prove homogeneity of variance for the intention of providing reasonable assessments of variation for future researchers, who would need a valid assessment of experimental variation with which to estimate number of animal replications needed to provide suitable statistical power (Berndtson, 1991).
The N and RNA concentrations of protozoa collected using the current protozoal filtration method and bacteria harvested directly from the ruminal fluid (i.e., not obtained from the protozoal filtrate as in experiment 1) are shown in Table 3. The high N and RNA concentrations in the protozoal samples indicate successful removal of contaminating plant material (which has lower N and RNA concentrations). The slightly lower values for bacterial N than those seen in some reports (Clark et al., 1992) probably are a result of salt contamination from saline in our final wash (other studies used distilled water) or from greater deposition of stored carbohydrate from these relatively high-grain diets. Although our blending procedure could cause more plant contamination, which could decrease N, we note that several researchers in the Clark et al. (1992) database blended samples to detach bacteria.
|
). For cows fed HF, pool sizes were nearly identical (27.4 and 28.1 g) for the real-time PCR and gravimetric procedures. In contrast, for cows fed LF, protozoal N pool sizes were lower for the real-time PCR method than the gravimetric method. This apparent discrepancy might be due to a discrepancy in rDNA copies per protozoan cell (see later discussion). Variances were not different (P > 0.2) between the 2 methods, suggesting that each method estimated means with the same precision.
|
) are also consistent with predictions based on a mechanistic model of protozoal ecology (Dijkstra et al., 1998). For cattle fed 17.1 kg/d of DMI, their model simulated an average of 21.6% protozoal N in the total microbial N, with a range from 10.1 to 27.6%. This percentage declined as DMI simulations increased, so at our higher DMI (24.5 kg/d), presumably with more rapid ruminal passage rates, we would expect a further predicted decline in protozoal N percentage from that model.
The bacterial N pool sizes were not different (P > 0.5) among diets and averaged 142 and 189 g for LF and HF diets, respectively, determined by real-time PCR, compared with 133 and 189, respectively, determined by the gravimetric procedure (Table 4). This method still assumes that all purines of dietary origin are completely degraded in the rumen. Although as much as 20% of the feed purines escape ruminal degradation, the much lower concentrations of N and RNA of feed compared with bacteria dilute this error to 5% or less (Calsamiglia et al., 1996). Future researchers should consider this and other issues with regard to validation of the purine assay in their laboratories (Firkins and Reynolds, 2005). The SE were higher relative to the means for the bacterial N pool sizes, as might be expected for a procedure that combines the errors from both the real-time PCR and purine assays.
Variances between methods were different (P < 0.06) when comparing duodenal flow of protozoal N; therefore, separate SEM are given for each method and were used to determine treatment differences (Table 4). The duodenal flow of protozoal N was lower (P < 0.01) when cows were fed LF vs. HF when determined using the real-time PCR method. Similarly, when protozoal N pool size obtained using the gravimetric method was multiplied by either the passage rates per day of the ruminal fluid [6] or DNDF [7], these estimated duodenal flows of protozoal N were lower for LF than HF. The passage rate of DNDF was chosen because it explained much more of the difference in ruminal digestibilities of NDF in similar diets (Harvatine et al., 2002b) and because we reasoned that protozoa would be chemotactically attracted more to the DNDF than the indigestible NDF. Certain genera such as Epidinium, Isotricha, and Polyplastron are capable of temporary attachment to various substrates (Ankrah et al., 1990; Williams and Coleman, 1992), thus allowing them to associate with solid material and evade passage with the faster-passing ruminal fluid. For cows fed LF, the protozoal N flow determined using the real-time PCR procedure (16.1 g/d) was similar to that estimated using the passage rate of DNDF (17.0 g/d). Both were much lower than that estimated (44.4 g/d) assuming protozoa pass entirely with the ruminal fluid. For cows fed HF, the protozoal N flow determined using the real-time PCR procedure was in between that estimated using the passage rate of DNDF and that using the fluid dilution rate. Dijkstra et al. (1998) modeled the passage of protozoa to be half that of the solids dilution rate. Passage rate determined using rare earth markers, which was not measured in this study but has been used extensively in published reports, can migrate to small particles and fluid (Firkins et al., 1998) and often is up to 2-fold higher than indigestible NDF. Using that model’s parameters, we would therefore expect protozoal N passage (half of the solids passage rate) to approach the passage rate of indigestible NDF (half of rare earth-determined passage rate). Thus, passage of protozoal N to the duodenum using the gravimetric method are in agreement with our expectations (similar to or higher than that estimated using [7]) based on this model of protozoal metabolism. The decreasing pool size and duodenal flow of the higher grain diet (LF) could be because of varying feed intake during the sampling period for cow 660 (data not shown) but also a result of lower ruminal pH (average of 6.18 for LF vs. 6.34 for HF; data not shown). At times during the feeding cycle, decreasing ruminal pH would be expected to be inhibitory to protozoa (Dehority, 2003). Moreover, the faster passage rates of DNDF for LF than HF (0.039 vs. 0.026/h; data not shown) are indicative of a less firm ruminal mat to retain small feed particles and probably the protozoa feeding on them.
The net generation times of protozoal N calculated using the real-time PCR assay were 10.7 and 17.3 h for LF and HF, respectively (pool size/duodenal flow x 24 h/d), explaining why relatively slow-growing protozoa represented a higher concentration of the total microbial N for HF. Although our average apparent generation times are somewhat shorter than those that others have determined (Dehority, 2003), those values generally were for sheep fed at lower maintenance intakes (Michalowski, 1998) or from various other indirect procedures (Dehority, 2003). With decreasing transfer interval as the criterion directly controlling the measurement of generation time, Dehority (1998) reported in vitro generation times for Epidinium (which was >9% in cow 647) and Entodinium species (>98% in cow 660) of 12 h (data from Sylvester et al., 2004). Williams and Withers (1993) determined apparent net generation times of 9 to 10 h in sheep. These net generation times are protozoa-doubling time corrected for cell lysis and recycling (Williams and Withers, 1993), both of which likely change with nutrient availability (Coleman, 1979; Dijkstra et al., 1998). Faster dilution rates for dairy cattle vs. sheep or non-lactating cattle (for most of the protozoa literature) should decrease recycling as fluid retention time approaches protozoal generation interval, as occurs for efficiency of bacterial growth (Firkins, 1996). The fluid dilution rates averaged 0.124 and 0.135/h for LF and HF (P > 0.10; data not shown); fluid retention times (8.1 and 7.4 h; reciprocals of fluid dilution rates) are somewhat shorter than what would be expected to maintain protozoal populations in the rumen, providing further support that protozoal N flow to the duodenum calculated using the gravimetric method with [6] is overestimated.
Unfortunately, the low pH of the abomasum completely lyses protozoal cells, so there was no independent, count-based procedure for us to verify our duodenal flow data. As discussed in our companion paper (Sylvester et al., 2004), low pH in the abomasum appeared to have no deleterious effects on recovery of protozoal rDNA (typically about 108 copies/mL) in our current study. Further research quantifying omasal flow of protozoal cells would help evaluate our technique under more diverse dietary conditions.
Duodenal flow of bacterial N did not differ (P > 0.5) among diets (Table 4). The N:RNA ratios in bacteria were about 9.5 and 16.5% lower for bacteria than protozoa for LF and HF, respectively (calculated from data in Table 3). The standard method used in most research papers assumes negligible protozoal N or RNA in duodenal digesta because only a bacterial standard is used. Thus, using only bacterial N:RNA ratios would bias the determination of total microbial N progressively with more protozoal N flowing to the duodenum and with more divergent N:RNA ratios in bacteria vs. protozoa. Using N:RNA data in protozoa and bacteria from Firkins et al. (1987a, 1987b) and an estimated 41% of microbial N as protozoal N (Firkins et al., 1987a), we calculated that total microbial N flow and microbial N pool could be underestimated by as much as 13 and 20%, respectively, if protozoal N:RNA is ignored. Without reliable protozoal markers, the degree of error in studies ignoring protozoal contribution of microbial markers is hard to assess.
DGGE banding patterns of ruminal and duodenal digesta.
The counts or generic distributions of protozoa were not significantly affected by diet, although differences in protozoal genera among animal were detected (Sylvester et al., 2004). The most notable difference was that cow 660 contained up to 15% Epidinium, but none were observed in cow 647. There were numeric increases (20%; data not shown) in protozoal counts on the HF diet that support the larger protozoal N pool sizes (Table 4).
The rDNA ranged from 1010 to 6210 copies per protozoal cell (data not shown). For cows fed HF, the mean rDNA copies per protozoal cell were very similar when determined from either ruminal fluid or from the filtered protozoa samples, averaging 6210 and 6140 for cow 647 and 1790 and 1870 for cow 660, respectively. When fed LF, however, mean rDNA copies per protozoal cell were somewhat lower for the ruminal fluid compared with the protozoal standard, averaging 1940 and 4370 for cow 647 and 1010 and 2730 for cow 660. To our knowledge, no data for ruminal protozoa have been reported, although the nonrumen protozoan Tetrahymena maintained about 9000 copies of rDNA per cell (Blomberg et al., 1997) and the ratio of rDNA/total genomic DNA should increase during rapid growth (Larson et al., 1991). If rDNA copies per cell were adjusted to the lower values for ruminal fluid, ruminal pool size and duodenal flow would be increased for LF. We are investigating how substrate availability and time after feeding affects rDNA copies per cell to explain these effects and for potential improvements in accuracy of our assay. However, our previous work seemed to rule out effects of PCR inhibitors or differences in amplification efficiencies with varying amounts of template (Sylvester et al., 2004).
Denaturing gradient gel electrophoresis is a culture-independent method that has been used to study ecology of ruminal bacteria (Kocherginskaya et al., 2001) and nonrumen protozoa (Rasmussen et al., 2001). More recently, Regensbogenova et al. (2004) have adapted DGGE analysis to ruminal protozoa. This analysis is semiquantitative, subject to PCR biases, and requires cloning for identification of bands (Zoetendal et al., 2004). However, sequencing was done to identify bands, and PCR bias seems much less likely for protozoa than for bacteria because the latter are much more diverse (Kocherginskaya et al., 2001), whereas ruminal protozoa are monophyletic (Mackie et al., 2000; Dehority, 2003). Mackie et al. (2000) concluded that DGGE is a very useful procedure to monitor microbial population shifts.
Distinct DGGE banding patterns were obtained, with primer set 1 producing 5 major bands (Figure 2) and primer set 2 producing 11 major bands (Figure 3). Because a threshold amount of rDNA needs to be amplified for bands to be distinguished, these bands represent the most abundant protozoal species. Sequence analysis supports the diversity of abundant phylotypes represented by DGGE bands and generally matches well with protozoal counts (Sylvester et al., 2004). For example, band 4 in Figure 2 and band 6 in Figure 3 are seen for cow 647 but barely visualized for cow 660, which is expected based on the high abundance of Epidinium counts in 647 but not in 660 (Sylvester et al., 2004). As sequence identity decreases below 99%, the band identity becomes progressively difficult to resolve because of the high degree of homology between rumen protozoa. More 18S rDNA sequencing of functionally important (but not previously sequenced) protozoal species for deposition in GenBank (Benson et al., 2004) would help further characterize protozoal populations from the 
94% sequence identity closer to true species identification.
|
|
The DGGE banding profiles generally were much more similar between the ruminal and duodenal samples within each respective animal than when compared across animals. Similar distributions of protozoal populations in the rumen and omasum (Towne and Nagaraja, 1990; Punia and Leibholz, 1994; Michalowski, 1998) further support our conclusion based on DGGE profiles that the filtration method accurately represents the actual protozoal populations of both the ruminal fluid and duodenal digesta; thus, the protozoal standard collected from the rumen appears to be satisfactory to adjust copies of protozoal rDNA to grams of protozoal N for both sites (Table 4).
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Received for publication November 2, 2004. Accepted for publication February 23, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
-subunit rRNA genes from Kinetoplastida (protozoa) used in denaturing gradient gel electrophoresis. Microb. Ecol. 42:109–115.[Medline]
This article has been cited by other articles:
|
|
 |
|
  J. T. Sylvester, S. K. R. Karnati, B. A. Dehority, M. Morrison, G. L. Smith, N. R. St-Pierre, and J. L. Firkins Rumen ciliated protozoa decrease generation time and adjust 18S ribosomal DNA copies to adapt to decreased transfer interval, starvation, and monensin J Dairy Sci, January 1, 2009; 92(1): 256 - 269. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L. Firkins, Z. Yu, and M. Morrison Ruminal Nitrogen Metabolism: Perspectives for Integration of Microbiology and Nutrition for Dairy J Dairy Sci, June 1, 2007; 90(13_suppl): E1 - E16. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. F. Brito, G. A. Broderick, and S. M. Reynal Effects of Different Protein Supplements on Omasal Nutrient Flow and Microbial Protein Synthesis in Lactating Dairy Cows J Dairy Sci, April 1, 2007; 90(4): 1828 - 1841. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. F. Brito, G. A. Broderick, J. J. O. Colmenero, and S. M. Reynal Effects of Feeding Formate-Treated Alfalfa Silage or Red Clover Silage on Omasal Nutrient Flow and Microbial Protein Synthesis in Lactating Dairy Cows J Dairy Sci, March 1, 2007; 90(3): 1392 - 1404. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. K. R. Karnati, J. T. Sylvester, S. M. Noftsger, Z. Yu, N. R. St-Pierre, and J. L. Firkins Assessment of Ruminal Bacterial Populations and Protozoal Generation Time in Cows Fed Different Methionine Sources J Dairy Sci, February 1, 2007; 90(2): 798 - 809. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L. Firkins, A. N. Hristov, M. B. Hall, G. A. Varga, and N. R. St-Pierre Integration of Ruminal Metabolism in Dairy Cattle J Dairy Sci, March 1, 2006; 89(e_suppl_1): E31 - E51. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
