Use of real-time PCR to predict dry matter disappearance of individual feeds in a total mixed ration

Please cite this article in press as: Alexander, T.W., et al., Use of real-time PCR to predictdry matter disappearance of individual feeds in a total mixed ration. Anim. Feed Sci.Technol. (2008), doi:10.1016/j.anifeedsci.2008.06.011

ARTICLE IN PRESSG ModelANIFEE-11999; No. of Pages 10

Animal Feed Science and Technology xxx (2008) xxx–xxx

Contents lists available at ScienceDirect

Animal Feed Science andTechnology

journal homepage: www.elsevier.com/locate/anifeedsci

Use of real-time PCR to predict dry matter disappearanceof individual feeds in a total mixed ration

Trevor W. Alexandera,∗, Yuxi Wanga, Tim Reutera, Erasmus K. Okineb,Walter T. Dixonb, Tim A. McAllistera

a Agriculture and Agri-Food Canada Research Center, Lethbridge, Alta., Canadab University of Alberta, Edmonton, Alta., Canada

a r t i c l e i n f o

Article history:Received 7 December 2007Received in revised form 21 May 2008Accepted 26 June 2008Available online xxx

Keywords:Real-time PCRTotal mixed rationDigestibilityAssociative effectsAlfalfaCorn

a b s t r a c t

In ruminant nutrition, the accuracy of predicting available energyfrom a total mixed ration (TMR) is limited because digestibilitiesof individual feedstuffs vary from predicted values in response toassociative effects. To overcome this limitation, it was proposedthat plant-specific real-time PCR assays could be used to predictthe digestion of individual feeds within a TMR. To test this hypoth-esis, alfalfa and corn substrates (0.5 g dry matter) were incubatedin buffered rumen fluid at pH 6.8 or 5.5 for up to 48 h. Regressionanalysis between DNA disappearance and dry matter disappearance(DMD) for only alfalfa or corn was used to predict DMD in mixedrations containing 75:25, 50:50, and 25:75 ratios of alfalfa to corn,after plant-specific DNA quantification. The actual DM remainingwas compared against the total predicted DM remaining by sum-ming the individual values predicted for alfalfa and corn in themixed rations. The method was accurate and precise in predict-ing DM remaining (Concordance and Pearson correlations were0.89 and 0.91, respectively). Differences between coefficients of theactual and total predicted DM remaining ranged between −0.081and 0.087 or −0.08 and 0.09 when the pH was 6.8 or 5.5, respec-tively. Furthermore, by analyzing substrate incubated at pH 5.5, themethod was tested in the presence of an associative effect (reducedpH) known to inhibit fibre digestion but have limited consequence

Abbreviations: A100, substrate 100:0 alfalfa:corn; A75:C25, substrate 75:25 alfalfa:corn; A50:C50, substrate 50:50alfalfa:corn; A25:C75, substrate 25:75 alfalfa:corn; C100, substrate 0:100 alfalfa:corn; DM, dry matter; DMD, dry matter disap-pearance; PCR, polymerase chain reaction; RF, rumen fluid; TMR, total mixed ration.

∗ Corresponding author at: Agriculture and Agri-Food Canada Research Center, P.O. Box 3000, Lethbridge, Alta. T1J 4B1, Canada.Tel.: +1 403 317 3375; fax: +1 403 382 3156.

E-mail address: [email protected] (T.W. Alexander).

0377-8401/$ – see front matter. Crown Copyright © 2008 Published by Elsevier B.V. All rights reserved.doi:10.1016/j.anifeedsci.2008.06.011

Please cite this article in press as: Alexander, T.W., et al., Use of real-time PCR to predictdry matter disappearance of individual feeds in a total mixed ration. Anim. Feed Sci.Technol. (2008), doi:10.1016/j.anifeedsci.2008.06.011

ARTICLE IN PRESSG ModelANIFEE-11999; No. of Pages 10

2 T.W. Alexander et al. / Animal Feed Science and Technology xxx (2008) xxx–xxx

on starch utilization. The method predicted DMD in agreement withthe expected results. The DMD of alfalfa was significantly affected(P<0.05) by interactions between time and pH or between diet andpH, whereas DMD for corn was affected by time or diet (P<0.05). Thepresented method may lead to redefined models to more effectivelypredict the available energy of individual feedstuffs within a TMR.Crown Copyright © 2008 Published by Elsevier B.V. All rights reserved.

1. Introduction

The current feeding calculations for ruminants are based on net energy or metabolizable energysystems (Lopez et al., 2000). Both systems are dependent on the digestibility of feeds, which accountsfor most of the variations in predicting available energy for production parameters (Minson, 1982).Each system is limited by using a constant energy value for a single feed whether fed alone or incombination with other feeds (Lopez et al., 2000). None of the systems incorporates associative effectsbetween feeds, which are known to occur in total mixed rations (TMR; Kirkan and Krishnamoorthy,2007; Zhao et al., 2005).

Associative effects can result in deviation between the actual and predicted digestibility of TMR(Dixon and Stockdale, 1999). For example, negative associative effects may arise from the addition ofgrain to forage due to a reduction in rumen pH as a result of the rapid production of volatile fatty acidsfrom fermentable carbohydrates. Microbial digestion of fibre components in forages is optimal at apH range of 6.6–7.0 and typically decreases as the pH declines below 6.2 (Terry et al., 1969; Hiltnerand Dehority, 1983). Additionally, microbial populations may show substrate preference for one feedingredient over another, potentially altering the digestibility of individual feeds in a TMR (Mould etal., 1983). Therefore, the energy value of a feed component can change when mixed with other feeds.

Currently, there is no method available to directly quantify the digestion of individual feeds in aTMR. Consequently, there is no way of measuring whether an associative effect is specific to one ormore of the component feeds in a TMR. Such a method could contribute significantly to predicting theavailable energy of a mixed diet and lead to improved, cost-saving, formulations of diets. The ability tomeasure digestion of single dietary ingredients would be dependent on quantifying a marker uniqueto each feed substrate or plant. Low-copy plant genes (i.e., those present at less than 10 copies percell genome) have been detected in the ruminant digestive tract only in association with solid feedresidues, and an inverse relationship between plant cell degradation and persistence of plant cell DNAhas been observed (Alexander et al., 2004; Phipps et al., 2003). Thus, we hypothesized that uniqueplant DNA markers may provide the means to estimate the digestibility of specific plant cells within amixed diet. To explore this concept, an in vitro approach was used to determine if real-time PCR assaysspecific to alfalfa and corn could be used to estimate dry matter disappearance (DMD) of individualfeeds within a TMR during fermentation.

2. Materials and methods

2.1. Ruminal fermentation substrates

Dried alfalfa forage was ground to pass a 1-mm screen. Corn seed was ground to pass an 8-mmscreen. Dry matter (DM) contents of alfalfa and corn were determined after drying for 24 h at 105 ◦C.For the ruminal fermentation experiment, substrates were 100:0 (A100), 75:25 (A75:C25), 50:50(A50:C50), 25:75 (A25:C75), and 0:100 (C100) ratios of alfalfa:corn. Substrates were weighed (0.5 gDM) into filter bags (#57, ANKOM Technology, Macedon, United States).

2.2. Ruminal fermentation

The DaisyII Incubator in vitro fermentation system was used to test the DMD and DNA disappear-ance of each substrate, according to the manufacturer’s instructions (ANKOM Technology). Two buffer

Please cite this article in press as: Alexander, T.W., et al., Use of real-time PCR to predictdry matter disappearance of individual feeds in a total mixed ration. Anim. Feed Sci.Technol. (2008), doi:10.1016/j.anifeedsci.2008.06.011

ARTICLE IN PRESSG ModelANIFEE-11999; No. of Pages 10

T.W. Alexander et al. / Animal Feed Science and Technology xxx (2008) xxx–xxx 3

mixtures of pH 6.8 or 5.5 were prepared, with the latter simulating a known associative effect (i.e.reduced pH) due to rapid fermentation of easily digestible carbohydrates. To achieve the lower pH of5.5, less of the recommended amount of ANKOM Solution B was added to ANKOM Solution A (ANKOM,2008). Rumen fluid (RF) collected from a ruminally fistulated steer maintained on a barley silage andground barley grain diet was processed under anaerobic conditions. The steer was cared for accordingto the Canadian Standards on Animal Care Guidelines. Processed RF inoculant (400 mL) was addedto ANKOM fermentation jars containing either pH 6.8 or 5.5 buffers (1600 mL). Filter bags containingthe substrates were added to the fermentation jars and incubated at 39 ◦C. Triplicate substrate bagswere removed at 0, 8, 16, 24, and 48 h of incubation, rinsed under cold water and then frozen in liquidnitrogen. The pH of the medium was measured at each time point. The filter bags were freeze-driedand weighed to determine actual DMD, prior to DNA extraction.

2.3. DNA extraction and quantification

Substrates were removed from the bags and ground to a powder using a planetary micro mill(Retsch, Albisheim, Germany). DNA was extracted from each powder (30 mg) using a Qiagen DNeasyPlant Mini Kit (Qiagen Inc., Mississauga, Canada). The same DNA extraction method was used to isolateDNA from alfalfa and corn for use as real-time PCR standards. All DNA was quantified fluorometrically(excitation filter spectrum of 470–490 nm; emission filter spectrum of 515–525 nm) using the Quant-iTTM PicoGreen® dsDNA Assay Kit (Invitrogen, Burlington, Canada) with a VersaFluor fluorometer(BioRad, Mississauga, Canada).

2.4. PCR analyses and plant DNA disappearance

Oligonucleotides used in PCR are described in Table 1. Conventional PCR was used to detect thepresence of Fibrobacter succinogenes and Streptococcus bovis after 8, 16 and 24 h of ruminal fermentationof mixed rations. PCR conditions have been described previously (Tajima et al., 2001). Each PCR mixture(50 �L) contained: 1× HotStarTaq Master Mix (Qiagen Inc.), 20 pmol of each primer and 200 ng oftemplate DNA. The PCR were performed with a Mastercycler ep (Eppendorf, Mississauga, Canada).

Primers and probe for real-time PCR analysis of corn DNA were designed using Beacon Designer 4.0software (PREMIER Biosoft Int., Palo Alta, United States); those for alfalfa were previously published(Alexander et al., 2007). Alfalfa DNA was quantified using a PCR assay specific to the Medicago sativaacetyl-CoA carboxylase gene (Shorrosh et al., 1994; Accession No. L25042). Corn DNA was quantifiedusing a PCR assay specific to the maize alcohol dehydrogenase 1 gene (Sachs et al., 1986; Accession No.X04050). An iCyler iQ system (BioRad) was used for PCR. Each PCR (50 �L) contained: 1× Universal

Table 1Oligonucleotides used in PCR

Target Sequence (5′ → 3′) Gene Fragment size (bp)

F. succinogenesa (Fc) GGTATGGGATGAGCTTGC 16S rDNA 445(Rd) GCCTGCCCCTGAACTATC

S. bovisa (F) CTAATACCGCATAACAGCAT 16S rDNA 869(R) AGAAACTTCCTATCTCTAGG

Alfalfab (F) GATCAGTGAACTTCGCAAAGTAC Acetyl-CoA carboxylase 91(R) CAACGACGTGAACACTACAAC(Pe) TGAATGCTCCTGTGATCTGCCCATGC

Corn (F) CACTCTCAGCTACTTTCCTTCTCC Alcohol dehydrogenase 1 128(R) TGAACTCAGCGTCCTTATGTGG(P) ACACCCACCAGCACAGCAACACCC

a From Tajima et al. (2001).b From Alexander et al. (2007).c Forward primer.d Reverse primer.e Probe (labelled at the 5′-end with the 6-carboxyfluorescein (6-FAM) and at the 3′-end with 6-carboxytetramethylrodamine

(TAMRA).

Please cite this article in press as: Alexander, T.W., et al., Use of real-time PCR to predictdry matter disappearance of individual feeds in a total mixed ration. Anim. Feed Sci.Technol. (2008), doi:10.1016/j.anifeedsci.2008.06.011

ARTICLE IN PRESSG ModelANIFEE-11999; No. of Pages 10

4 T.W. Alexander et al. / Animal Feed Science and Technology xxx (2008) xxx–xxx

Master Mix (Applied Biosystems, Streetsville, Canada), 7.5 pmol of each primer, 2.5 pmol of probe,and 200 ng of unknown template DNA. The standards for alfalfa and corn real-time PCR included2.0 × 105, 1.0 × 105, 2.0 × 104, 1.0 × 104, 2.0 × 103, 1.0 × 103, 200, and 100 pg of alfalfa or corn DNA perPCR, respectively. In all instances, PCR assays showed efficiencies between 90% and 105%.

Microbial DNA likely diluted plant DNA to varying extents determined by the length of fermen-tation and/or type of fermentation substrate. To account for this variation, the amount of alfalfa orcorn DNA quantified in 200 ng of extracted DNA was used to calculate the total amount (DM basis)of each feed-specific type of DNA remaining at every time point for individual substrates through-out fermentation. These data were used to determine total DNA disappearance for alfalfa and cornsubstrates.

2.5. Statistical analyses

Regression between DNA disappearance (independent variable) and DMD (dependent variable) forthe substrates containing only alfalfa or corn (A100 and C100, respectively) were analyzed at pH 6.8and 5.5 using Proc NLIN (SAS Institute, 1999). Both DNA disappearance and DMD were calculated as theproportion of each quantified at single time points (8, 16, 24, and 48 h) compared to each quantifiedat time 0 h. Segmented regression models were used to fit DMD according to DNA disappearancefor alfalfa (Shuai et al., 2003). The models derived from regression were used to predict the DMD ofalfalfa and corn in substrates containing mixed amounts of each (A75:C25, A50:C50, and A25:C75).Predicted alfalfa and corn DMD in each of the mixed substrates were compared against each otherusing Proc MIXED procedure of SAS Institute (1999). The model for analyzing predicted DMD includedthe fixed effects of diet, time, pH, and the interactions between these effects. The repeated statementwas applied to time of sampling. Various error structures were tested and the one giving the lowestakaike information criterion (AIC) was chosen for analysis.

To determine the accuracy of the real-time PCR method, the actual and predicted DM remaining at8, 16, and 24 h of fermentation of the mixed rations (A75:C25, A50:C50, and A25:C75) were compared.The predicted DM remaining in mixed rations was calculated by first using the regression equationsfrom digestion of only alfalfa (A100) or only corn (C100) to measure DNA-predicted DMD of eachfeedstuff in each of the mixed rations. From these data, the predicted amount of alfalfa and corn DMremaining at each measured point of digestion were calculated. Predicted DM weights of alfalfa andcorn were summed to estimate total predicted DM remaining in each of the mixed rations. The actualand predicted DM remaining were compared via Concordance correlation (Lin, 1989) and analysisusing Proc MIXED. For all statistical analyses, differences were considered significant if the P-valueswere less than or equal to an alpha value of 0.05.

3. Results

3.1. Regression analyses of control diets

The relationships between DNA disappearance and DMD from the control diets containing exclu-sively alfalfa (A100) or corn (C100) are shown in Fig. 1. From these data, equations were derived toestimate DMD from DNA disappearance as quantified by real-time PCR. For alfalfa, a single equa-tion did not provide the most accurate predictions for DMD. Instead, segmented regression was usedto define the best models for prediction of DMD (Shuai et al., 2003). A linear equation was derivedfrom time points 8 and 16 h (for predicting DMD at 8 h) and a quadratic equation from time points16 to 48 h (for predicting DMD at times 16 and 24 h). The regression models for alfalfa digested atpH 6.8 (linear: y = 1.1x − 68.8; quadratic: y = 2.9x2 − 573.2x + 28,075) and pH 5.5 (linear: y = 0.4x − 10.4;quadratic: y = 0.6x2 − 118.1x + 5723.9) differed, however, the trends were similar. The majority of DNAwas degraded in the first 8 h at both pH’s (0.88 at pH 6.8 versus 0.87 at pH 5.5). Despite this, significantamounts of alfalfa DNA were quantified at 48 h for each pH-treatment (57.8 �g at pH 6.8 versus 166.1 �gat pH 5.5; data not shown). Regression models for corn digested at pH 6.8 (y = −0.01x2 + 2.8x − 119.3)and pH 5.5 (y = 0.01x2 − 0.8x + 15.0) were both quadratic. After 24 h of incubation, the amount of DNA

Please cite this article in press as: Alexander, T.W., et al., Use of real-time PCR to predictdry matter disappearance of individual feeds in a total mixed ration. Anim. Feed Sci.Technol. (2008), doi:10.1016/j.anifeedsci.2008.06.011

ARTICLE IN PRESSG ModelANIFEE-11999; No. of Pages 10

T.W. Alexander et al. / Animal Feed Science and Technology xxx (2008) xxx–xxx 5

Fig. 1. Plotting of plant DNA disappearance against dry matter disappearance (DMD) during in vitro fermentation. Substratescomprising alfalfa only (A100) or corn only (C100) were incubated in rumen fluid at pH 6.8 or 5.5 (n = 3). Both DNA disappearanceand DMD were calculated as the proportion of each moiety quantified at 8, 16, 24, and 48 h of incubation compared to the initialamount of each at time 0 h. Residual DNA associated with alfalfa or corn was estimated using real-time PCR.

disappearance for corn (0.73 at pH 6.8 and 0.70 at pH 5.5) was less than that for alfalfa (0.99 at pH 6.8and 0.98 at pH 5.5).

3.2. Method accuracy

The accuracy of the real-time PCR method was tested by comparing the actual gravimetric andpredicted DM remaining at 8, 16, and 24 h of fermentation of the mixed rations (Fig. 2). The predictedDM remaining was calculated by summing individual values from alfalfa and corn, as described inSection 2. Differences in actual and predicted coefficient values ranged from −0.08 to 0.087 for pH6.8 and from −0.08 to 0.09 for pH 5.5. When data from the pH 6.8 and 5.5 data sets were combined,comparisons of actual and predicted DM remaining over time produced Concordance and Pearsoncorrelations of 0.89 and 0.91, respectively.

3.3. Predicted alfalfa and corn DMD

Results for individual DMD of alfalfa and corn are shown in Fig. 3. Throughout fermentation ofthe pH 6.8 treatment, media pH did not decrease to below pH 6.3. In contrast, media pH for the pH5.5 treatment was always less than pH 5.7 (data not shown). For alfalfa, predicted DMD was affectedby an interaction between time and pH (P<0.0001). At 8 h, there was no difference in DMD betweentreatments pH 6.8 or 5.5. However, at 16 and 24 h, DMD was reduced (P<0.0001 at both times) at pH 5.5compared with pH 6.8. F. succinogenes was detected by PCR at every time point from the diet samplesincubated at pH 6.8, whereas it was not detected in DNA extracted from the samples at pH 5.5 (data notshown). In contrast to alfalfa, time did affect (P<0.0001) predicted DMD from corn, which increasedthroughout fermentation, but was not altered by pH. S. bovis was detectable by PCR in all of the dietsamples, and at every time point, irrespective of media pH (data not shown).

Diet formulation affected predicted DMD of both mixed ration substrates. An interaction betweenpH and diet existed for alfalfa (P=0.036). Each of the DNA-predicted DMD values for the mixed diets(A75:C25, A50:C50, A25:C75) was greater at pH 6.8 compared to pH 5.5. Within the pH 6.8 treatment,alfalfa DMD was greater in diet A25:C75 compared to both A75:C25 (P=0.005) and A50:C50 (P<0.0001).Within the pH 5.5 treatment, DNA-predicted DMD did not differ among the diets, indicating that alfalfa

Please cite this article in press as: Alexander, T.W., et al., Use of real-time PCR to predictdry matter disappearance of individual feeds in a total mixed ration. Anim. Feed Sci.Technol. (2008), doi:10.1016/j.anifeedsci.2008.06.011

ARTICLE IN PRESSG ModelANIFEE-11999; No. of Pages 10

6 T.W. Alexander et al. / Animal Feed Science and Technology xxx (2008) xxx–xxx

Fig. 2. Plotting of the actual gravimetric versus DNA-predicted total dry matter remaining during in vitro fermentation ofmixed rations. Mixed rations containing 75:25 (A75:C25), 50:50 (A50:C50), and 25:75 (A25:C75) ratios of alfalfa:corn wereincubated in rumen fluid at pH 6.8 or 5.5 (n = 3). Real-time PCR assays specific to alfalfa or corn were used to estimate dry matterdisappearance of each plant substrate. From these data, the amount of individual alfalfa and corn dry matter remaining werecalculated and summed, in order to predict total dry matter remaining.

was digested to the same extent regardless of its concentration in the diets. For corn, diet compositionalone influenced DMD (P<0.0001), which was greatest for ration A75:C25, followed by A50:C50 andA25:C75.

4. Discussion

4.1. Regression analyses of control diets

For alfalfa, disappearance of DM and DNA in the first 8 h likely reflected loss due to soluble plantfractions and digestion of the more readily digestible components of the alfalfa plant such as theleaves. Theoretically, DNA remains protected within plant cells and associated with solid digesta untilenzymatic or physical digestion of the cell wall and membrane (Alexander et al., 2002). In legumes, theblades contain a greater proportion of thin-wall cells compared to the stem (Wilson, 1991) and it hasbeen shown that digestion of alfalfa leaves is more rapid than stems (Albrecht et al., 1987). The slowerrate of disappearance of DM and DNA in the later time points probably resulted from digestion of themore lignified plant cell walls associated with the stem. When alfalfa is at the mid-flowering stage ofmaturity, approximately 0.25 of the leaf blade mass is neutral detergent fibre compared to 0.4–0.55

Please cite this article in press as: Alexander, T.W., et al., Use of real-time PCR to predictdry matter disappearance of individual feeds in a total mixed ration. Anim. Feed Sci.Technol. (2008), doi:10.1016/j.anifeedsci.2008.06.011

ARTICLE IN PRESSG ModelANIFEE-11999; No. of Pages 10

T.W. Alexander et al. / Animal Feed Science and Technology xxx (2008) xxx–xxx 7

Fig. 3. DNA-predicted dry matter disappearances of alfalfa and of corn in mixed rations. Mixed rations containing 75:25(A75:C25), 50:50 (A50:C50), and 25:75 (A25:C75) ratios of alfalfa:corn were incubated in rumen fluid at pH 6.8 or 5.5 (n = 3).Real-time PCR assays specific to alfalfa or corn were used to measure DNA disappearance of each plant substrate. Dry matterdisappearance was estimated from DNA disappearance.

neutral detergent fibre content in the stems (Buxton et al., 1995). The excess fibre in the stems may havelimited microbial access to the interior of plant cells, resulting in slower DNA disappearance duringthe later hours of digestion (Fig. 1). Different digestibilities of stems and leaves likely explains whydividing the DNA and DM regressions into two equations resulted in more accurate DMD predictions.

The reduced DNA disappearance of corn compared to alfalfa may have resulted from differences inparticle size. Corn was ground to pass through an 8-mm screen whereas alfalfa was passed through a 1-mm screen. Larger particles would be less accessible to microbial digestion. Therefore, accessibility ofdigestive enzymes to DNA protected within cells of corn plant material would have also been reduced.

4.2. Method accuracy

According to Lin (1989), the Pearson correlation measures how far the observations deviate from thebest-fit line (measure of precision), while the Concordance correlation coefficient measures how farthe best-fit line deviates from the one-to-one line (measure of accuracy, or correctness). In the presentstudy, these correlations indicated that the real-time PCR method was both accurate and precise inpredicting individual DMD of alfalfa and corn in mixed rations, as indicated by analyzing their summedvalues of DM remaining. Additionally, statistical analysis revealed that the actual and predicted DMremaining were not different from each other. Combined, these results demonstrated that quantifyingplant-specific DNA by real-time PCR could be applied to estimate the DMD of each substrate in mixedrations.

Please cite this article in press as: Alexander, T.W., et al., Use of real-time PCR to predictdry matter disappearance of individual feeds in a total mixed ration. Anim. Feed Sci.Technol. (2008), doi:10.1016/j.anifeedsci.2008.06.011

ARTICLE IN PRESSG ModelANIFEE-11999; No. of Pages 10

8 T.W. Alexander et al. / Animal Feed Science and Technology xxx (2008) xxx–xxx

4.3. Predicted alfalfa and corn DMD

In vitro DMD studies are typically conducted in buffered media to maintain pH above pH 6. However,we measured digestion deliberately under two pH treatments designed to elicit associative effects, inorder to test further the application of our method.

Differences in digestion of alfalfa between the two pH treatments appeared to be related to theactivity of fibrolytic bacteria. At 8 h of incubation, DMD did not differ with pH. This was likely a reflec-tion of digestive activity being directed primarily against the easily digestible leaf fraction of alfalfa(Albrecht et al., 1987). However, digestion of the fibrous material at later incubations was reducedfor the pH 5.5 treatment and we were unable to detect F. succinogenes at this low pH. This result isin agreement with earlier reports that the digestive activity of fibrolytic bacteria is inhibited at pHbelow 6.2 (Hoover, 1986; Russell and Dombrowski, 1980). Growth and fermentation of cellulose by F.succinogenes is substantially reduced below pH 5.7 (Russell, 1987). Negative associative effects due tothe addition of readily fermentable carbohydrates (RFC) to forage diets has been previously shown invitro (Mertens and Loften, 1980; Hiltner and Dehority, 1983) and in vivo (Joanning et al., 1981; Millerand Muntifering, 1985). Reduction in ruminal fibre digestion has been suggested to arise from severalfactors including a decrease in pH due to the rapid fermentation of RFC (Hoover, 1986). It was evidentin this study that a pH of 5.5 decreased DMD of alfalfa.

In contrast to alfalfa, pH did not affect the DMD of corn. This difference was likely due to the higherstarch content of corn and, therefore, the bacterial population involved in corn DM digestion. S. bovis,detected at all time points for both pH treatments in the present study, is an amylolytic bacterium thatcan actively ferment starch at a pH value as low as 4.5 (Russell, 1987). Activity of S. bovis and additionalstarch-utilizing bacteria resistant to low pH would have been expected to contribute to DMD of cornat both pH 6.8 and 5.5. This is consistent with previous reports in which pH as low as 4.6–4.9 did notdiminish the activity of amylolytic digestive microorganisms (Russell, 1991).

It was evident that diet composition affected digestion of both alfalfa and corn. While it was notclear why this occurred, it did appear that the microflora associated primarily with the digestion ofalfalfa or of corn may have contributed in some manner to digestion of the other feedstuff. Preferencefor one feed ingredient over another has previously been reported to affect the digestion of a TMR(Mould et al., 1983).

One of the limitations of the current methodology for measuring DMD of mixed rations is that it isnot possible to measure independently the extent to which individual substrates are being digested.In this study, we detected differences in DMD of both alfalfa and corn, which were dependent on theration formulation (alfalfa versus corn) and media pH (for alfalfa). By adjusting rumen fluid to pH5.5, we were able to induce a known associative effect that negatively impacts fibre digestion due toshifts in microbial populations. Our model using DNA quantification for alfalfa predicted less digestionat pH 5.5 compared to pH 6.8, validating its ability to identify associative effects. In agreement withprevious studies, the corn digestion models showed no effects of pH on corn digestion (Russell andDombrowski, 1980; Russell et al., 1979).

We realised that there may be limitations to utilizing real-time PCR for digestion studies. PlantDNA may be degraded or heavily diluted in microbial DNA, to an extent that is not quantifiable. Inthese cases, the method would overestimate DMD. Using higher copy genes for quantification mayhelp reduce this error. High-copy plant genes were detectable in bovine feces when low-copy plantgenes were not (Phipps et al., 2003). The genes selected for real-time PCR in this study were bothlow-copy genes (Shorrosh et al., 1994; Sachs et al., 1986) and adequate amounts remained after 48 h ofincubation to estimate DMD. Additionally, in situations where different diets containing the same plant(e.g. corn silage and corn grain) require testing, the DNA-based assays would not differentiate the diets.Some genes, however, are differentially expressed in specific plant tissues, thus RNA quantificationcould perhaps overcome this limitation. In vitro digestibility studies are commonly used in ruminantnutrition to estimate the feeding value of feeds (Dijkstra et al., 2005). Thus, our method offers thepotential for independent comparison of in vitro ruminal digestion of feed components, with possibleapplication to the entire digestive tract of ruminants and monogastric animals as well. Although ourmodel was based upon alfalfa and corn, we anticipate that the principles applied in this paper areapplicable to additional plant species. By quantifying digestion of individual feed components in the

Please cite this article in press as: Alexander, T.W., et al., Use of real-time PCR to predictdry matter disappearance of individual feeds in a total mixed ration. Anim. Feed Sci.Technol. (2008), doi:10.1016/j.anifeedsci.2008.06.011

ARTICLE IN PRESSG ModelANIFEE-11999; No. of Pages 10

T.W. Alexander et al. / Animal Feed Science and Technology xxx (2008) xxx–xxx 9

presence of associative effects, improved models to more accurately predict the available energy inmixed rations may be achieved and lead to more economical diet formulations.

Acknowledgements

Natural Sciences and Engineering Research Council of Canada (NSERC) fellowship to Trevor W.Alexander and Tim Reuter is gratefully appreciated. The authors thank Toby Entz for his help withstatistical analyses and Katherine Jakober for help in preparing the manuscript.

References

Albrecht, K.A., Wedin, W.F., Buxton, D.R., 1987. Cell-wall composition and digestibility of alfalfa stems and leaves. Crop Sci. 27,735–741.

Alexander, T.W., Reuter, T., McAllister, T.A., 2007. Qualitative and quantitative polymerase chain reaction assays for an alfalfa(Medicago sativa)-specific reference gene to use in monitoring transgenic cultivars. J. Agric. Food Chem. 55, 2918–2922.

Alexander, T.W., Sharma, R., Deng, M.Y., Whetsell, A.J., Jennings, J.C., Wang, Y., Okine, E., Damgaard, D., McAllister, T.A., 2004.Use of quantitative real-time and conventional PCR to assess the stability of the cp4 epsps transgene from Roundup Readycanola in the intestinal, ruminal, and fecal contents of sheep. J. Biotechnol. 112, 255–266.

Alexander, T.W., Sharma, R., Okine, E.K., Dixon, W.T., Forster, R.J., Stanford, K., McAllister, T.A., 2002. Impact of feed processingand mixed ruminal culture on the fate of recombinant EPSP synthase and endogenous canola plant DNA. FEMS Microbiol.Lett. 214, 263–269.

ANKOM Technology, 2008. In vitro True Digestibility Using the DAISYII Incubator. Available at:http://www.ankom.com/09 procedures/Daisy%20method.pdf.

Buxton, D.R., Mertens, D.R., Moore, K.J., 1995. Forage quality for ruminants: plant and animal considerations. Prof. Anim. Sci. 11,121–131.

Dixon, R.M., Stockdale, C.R., 1999. Associative effects between forages and grains: consequences for feed utilisation. Aust. J. Agric.Res. 50, 757–773.

Dijkstra, J., Kebreab, E., Bannink, A., France, J., Lopez, S., 2005. Application of the gas production technique to feed evaluationsystems for ruminants. Anim. Feed Sci. Technol. 123–124, 561–578.

Hiltner, P., Dehority, B.A., 1983. Effect of soluble carbohydrates on digestion of cellulose by pure cultures of rumen bacteria. Appl.Environ. Microbiol. 46, 642–648.

Hoover, W.H., 1986. Chemical factors involved in ruminal fiber digestion. J. Dairy Sci. 69, 2755–2766.Joanning, S.W., Johnson, D.E., Barry, B.P., 1981. Nutrient digestibility depressions in corn silage-corn grain mixtures fed to steers.

J. Anim. Sci. 53, 1095–1102.Kirkan, D., Krishnamoorthy, U., 2007. Rumen fermentation and microbial biomass synthesis indices of tropical feedstuffs deter-

mined by the in vitro gas production technique. Anim. Feed Sci. Technol. 134, 170–179.Lin, L.I.-K., 1989. A Concordance correlation coefficient to evaluate reproducibility. Biometrics 45, 255–268.Lopez, S., Dijkstra, J., France, J., 2000. Prediction of energy supply in ruminants with emphasis on forages. In: Givens, D.I.,

Owen, E., Axford, R.F.E., Omed, H.M. (Eds.), Forage Evaluation in Ruminant Nutrition. CAB International, Oxfordshire, pp. 63–94.

Mertens, D.R., Loften, J.R., 1980. The effect of starch on forage fiber digestion kinetics in vitro. J. Dairy Sci. 63, 1437–1446.Miller, B.G., Muntifering, R.B., 1985. Effect of forage:concentrate on kinetics of forage fiber digestion in vivo. J. Dairy Sci. 68,

40–45.Minson, D.J., 1982. Effect of chemical composition on feed digestibility and metabolizable energy. Nutr. Abstr. Rev. 52, 591–615.Mould, F.L., Ørskov, E.R., Mann, S.O., 1983. Associative effects of mixed feeds. I. Effects of the type and level of supplementation

and the influence of the rumen fluid pH on cellulolysis in vivo and dry matter digestion of various roughages. Anim. FeedSci. Technol. 10, 15–30.

Phipps, R.H., Deaville, E.R., Maddison, B.C., 2003. Detection of transgenic and endogenous plant DNA in rumen fluid, duodenaldigesta, milk, blood, and feces of lactating dairy cows. J. Dairy Sci. 86, 4070–4078.

Russell, J.B., Sharp, W.M., Baldwin, R.L., 1979. The effect of pH on maximum bacterial growth rate and its possible role as adeterminant of bacterial competition in the rumen. J. Anim. Sci. 48, 251–255.

Russell, J.B., Dombrowski, D.B., 1980. Effect of pH on the efficiency of growth by pure cultures of ruminal bacteria in continuousculture. Appl. Environ. Microbiol. 39, 604–610.

Russell, J.B., 1987. Effect of extracellular pH on growth and proton motive force of Bacteriodes succinogenes, a cellulolytic ruminalbacterium. Appl. Environ. Mircrobiol. 53, 2379–2383.

Russell, J.B., 1991. Resistance of Streptococcus bovis to acetic acid at low pH: relationship between intracellular pH and anionaccumulation. Appl. Environ. Mircrobiol. 57, 255–259.

Sachs, M.M., Dennis, E.S., Gerlach, W.L., Peacock, W.J., 1986. Two alleles of maize alcohol dehydrogenase 1 have 3′ structural andpoly(A) addition polymorphisms. Genetics 113, 449–467.

SAS Institute, 1999. SAS/STAT User’s Guide, Version 8.1. SAS Institute, Cary, NC.Shorrosh, B.S., Dixon, R.A., Ohlrogge, J.B., 1994. Molecular cloning, characterization, and elicitation of acetyl-CoA carboxylase

from alfalfa. Proc. Natl. Acad. Sci. U.S.A. 91, 4323–4327.Shuai, X., Zhou, Z., Yost, R.S., 2003. Using segmented regression models to fit soil nutrient and soybean grain yield changes due

to liming. J. Agric. Biol. Environ. Stat. 8, 240–252.Tajima, K., Aminov, R.I., Nagamine, T., Matsui, H., Nakamura, M., Benno, Y., 2001. Diet-dependent shifts in the bacterial population

of the rumen revealed with real-time PCR. Appl. Environ. Microbiol. 67, 2766–2774.

Please cite this article in press as: Alexander, T.W., et al., Use of real-time PCR to predictdry matter disappearance of individual feeds in a total mixed ration. Anim. Feed Sci.Technol. (2008), doi:10.1016/j.anifeedsci.2008.06.011

ARTICLE IN PRESSG ModelANIFEE-11999; No. of Pages 10

10 T.W. Alexander et al. / Animal Feed Science and Technology xxx (2008) xxx–xxx

Terry, R.A., Tilley, J.M.A., Outen, G.E., 1969. Effect of pH on cellulose digestion under in vitro conditions. J. Sci. Food Agric. 20,317–320.

Wilson, J.R., 1991. Organization of forage plant tissues. In: Jung, H.G., Buxton, D.R., Hatfield, R.D., Ralph, J. (Eds.), Forage CellWall Structure and Digestibility. American Society of Agronomy, Inc., Crop Science of America, Inc., Soil Science Society ofAmerica, Inc., Madison, pp. 1–27.

Zhao, G.-Y., Li, Y.-X., Ren, J.-B., Li, Y.-J., Guo, D.-S., 2005. The influence of associative effects on the in vitro-estimated utilizablecrude protein (uCP) of feeds for ruminants. Arch. Anim. Nutr. 59, 149–154.