Partial replacement of soybean meal with microalgae biomass on in vitro ruminal fermentation may reduce ruminal protein degradation

The objective of this study was to evaluate the effects of partially replacing soybean meal (SBM) with algal sources on in vitro ruminal fermentation. Using 6 fermenters in a 3 × 3 replicated Latin square with 3 periods of 10 d each, we tested 3 treatments: a control diet (CRT) with SBM at 17.8% of the diet dry matter (DM); and 50% SBM biomass replacement with either Chlorella pyrenoidosa (CHL); or Spirulina platensis (SPI). The basal diet was formulated to meet the requirements of a 680-kg Holstein dairy cow producing 45 kg/d of milk with 3.5% fat and 3% protein. All diets had a similar nutritional composition (16.0% CP; 34.9% NDF; 31.0% starch, DM basis) and fermenters were provided with 106 g DM/d split into 2 portions. After 7 d of adaptation, samples were collected for 3 d of each period for analyses of ruminal fermentation at 0, 1, 2, 4, 6, and 8 h after morning feeding for evaluation of the ruminal fermentation kinetics. For the evaluation of the daily production of total metabolites and for the evaluation of nutrient degrad-ability, samples from the effluent containers were collected daily. Statistical analysis was performed with the MIXED procedure of SAS with treatment, time, and their interactions considered as fixed effects; day, square, and fermenter were considered as random effects. Orthogonal contrasts (CRT vs. algae; and CHL vs. SPI) were used to depict the treatment effect, and significance was declared when P ≤ 0.05. Fermenters that received algae-based diets had a greater propionate molar concentration and molar proportion when compared with the fermenters fed CRT diets. In addition, those algae-fed fermenters had lower branched short-chain fatty acids (BSCFA) and isoacids (IA),


INTRODUCTION
Currently, one of the largest sources of protein in animal nutrition is soybean meal (SBM), mainly because of its good AA profile and high protein content and quality (Abraham et al., 2019;NASEM, 2021).However, relying on a major single ingredient is problematic because producers become too dependent on it.Price variability and possible environmental impact that soybean cultivation may cause (Zhang and Liu, 2020) are other concerns related to massive SBM utilization.Therefore, alternative protein sources are paramount.One emergent alternative protein source is algal biomass, due to its high protein content and quality (Becker, 2013).In addition, algae can be produced on vertical systems or on marginal lands, which would reduce the area required and avoid competition with other crops (Taelman et al., 2015).
Algae are a large group of autotrophic organisms that can vary from unicellular to multicellular, mostly aquatic and a few are terrestrial (Sahoo and Baweja, 2015).According to Becker (2013), just a few species are currently being cultivated on an industrial scale, which includes some green algae and cyanobacteria, such as Chlorella and Spirulina, respectively.Over the past decade, there has been a 150% increase in Europe in the number of new algae-producing facilities for dietary human supplements, especially for the production of Spirulina (142 t/yr) and Chlorella (30 t/yr), but only around 10% of the production is destined for livestock feed with possible growth in the near future (Araújo et al., 2021).
Evaluation of algal supplementation has shown promising results by increasing the average daily gain of beef steers fed Spirulina (Costa et al., 2016), and greater efficiency of nitrogen utilization was observed when supplemented to dairy animals (Lamminen et al., 2017(Lamminen et al., , 2019)).In addition, Lodge-Ivey et al. (2014) evaluated the replacement of SBM with Chlorella using in vitro ruminal fermentation.The authors reported that Chlorella-containing diets increased true OM degradability and total short-chain fatty acids (SCFA), especially iso-butyrate and iso-valerate.In a previous study from our group (data not published), we evaluated 50% and 100% replacement of SBM with either Chlorella pyrenoidosa (CHL) or Spirulina platensis (SPI) in diets with contrasting carbohydrate profiles, consisting of a diet with 42.5% of NDF and 25.7% of starch and a second diet with 26.8% of NDF and 40.6% of starch in a batch culture system.We observed that there was no interaction between the carbohydrate profile and algal species, which indicates that the effects of replacement were independent of the dietary carbohydrate profile.In addition, SPI reduced methane yield and modified the SCFA profile, especially increased in branched SCFA (BSCFA), which is a marker for dietary protein degradation.
Our prior study demonstrated some potential advantages of replacing SBM with either CHL or SPI, notably modulation in SCFA profile and reduction in methane yield, which can be relevant for the dairy industry.However, these findings should be further validated in a more robust system that allows greater volume of inoculum and longer fermentation periods.Therefore, the objective of the current study was to evaluate the partial replacement of SBM with either CHL or SPI in an in vitro continuous-culture dual-flow system using a diet formulated for a high-producing dairy cow.The hypothesis of this study is that CHL and SPI would increase SCFA production, especially BSCFA, and could be used as a partial replacement for SBM in dairy diets.

MATERIALS AND METHODS
All procedures using animals were approved by the University of Florida's Institutional Animal Care and Use Committee.

Experimental Design and Diets
Six fermenters of a dual-flow continuous-culture system were used in a replicated 3 × 3 Latin square design with 3 treatments and 3 fermentation periods.Treatments (diets) were randomly assigned within Latin square for each period.Diets were formulated to meet or exceed the requirements of a 680-kg Holstein dairy cow producing 45 kg/d of milk with 3.5% fat, 3.0% protein, and 4.8% lactose, and a DMI 25 kg/d (NASEM, 2021).In a previous study, we observed that the replacement of SBM with CHL or SPI can be carried out without negative effects on ruminal fermentation on in vitro batch culture system, regardless of the carbohydrate profile of the diet (data not published).For the current study, a corn silage-based diet was designed, where the control (CRT) diet had SBM as the main source of protein (17.8% of the diet DM), and the other 2 diets had a 50% replacement of SBM biomass with either CHL or SPI, which was the best replacement according to our previous study.Both algae sources were delipidated with cracked cell walls, The whole-plant corn silage was collected and dried in a forced-air oven (Heratherm, Thermo Scientific) at 60°C for 72 h.Subsequently, whole corn silage and the other feed ingredients, such as grass hay, corn meal, and soybean meal, were ground in a Wiley mill (model no.2; Arthur H. Thomas Co.) to pass a 2-mm screen.Each ingredient was homogenized, and a subsample was collected in a labeled bag and sent to Dairy One Laboratory (Ithaca, NY) for feed analysis.

Animals and Inoculum Collection
Three ruminally cannulated lactating Holstein dairy cows were used as ruminal content donors.Cows were kept at the Dairy Research Unit of the University of Florida (Alachua, FL) in a freestall barn with the other cows of the herd receiving a TMR containing 60% corn silage, 12.5% ground corn, 13% citrus pulp, 12% soybean meal, and 2.5% mineral and vitamin premix; containing 16% CP, 35% NDF, and 31% starch.On the first day of each experimental period, the ruminal content was collected from different anatomical areas from the rumen of the cows 2 h after the morning feeding and filtered through 4 layers of cheesecloth (grade 40) and stored in a prewarmed thermos.The inoculum was transported within 30 min of the collection to the laboratory and incubated in the prewarmed dual-flow in vitro system.

Dual-Flow Continuous-Culture System
The experiment was carried out in a dual-flow continuous-culture system similar to the one described by Hoover et al. (1976) and used in recent studies in our laboratory (Agustinho et al., 2022;Ravelo et al., 2022;Vinyard et al., 2023).Briefly, each fermenter was fed twice a day at a constant rate of 106 g/d DM (53 g each meal).This system simulates the ruminal fermentation efficiently due to constant temperature at 38.4 ± 0.44°C, agitation at 100 rpm, infusion of artificial saliva at 3.0 ± 0.11 mL/min and constant infusion of N 2 gas to maintain an anaerobic environment.The artificial saliva was formulated according to Weller and Pilgrim (1974) to allow the passage of liquid and solid flow.The outflow passage rate of the system across the experiment was 10.14 ± 0.37%/h.
Each experimental period was 10 d in length, for a total of 30 d of fermentation.Every period the inoculum was replaced with a fresh one collected from the same donors.The first 7 d of each period was considered as an adaptation for the microbiome of the inoculum to the diet and system (Ziemer et al., 2000), and the last 3 d were used for sampling and data collection.On the first 5 d of each period of fermentation, 15 N-nonenriched artificial saliva was used (Weller and Pilgrim, 1974).On d 5 of each experimental period, a 1-mL pulse dose of 0.0173 g/mL of 10.2% labeled ( 15 NH 4 ) 2 SO 4 (Sigma-Aldrich Co., St. Louis, MO) was added directly into each fermenter and nonenriched saliva was replaced with an 15 N-enriched saliva formulated to contain 0.077 g of 10.2% of labeled ( 15 NH 4 ) 2 SO 4 per liter until d 10 of the fermentation.The 15 N was used as a marker to estimate microbial protein synthesis (Calsamiglia et al., 1996).

Sampling and Data Collection
On d 5 of each experimental period, before the change in the saliva described earlier, samples of nonenriched saliva and background digesta from each fermenter were collected to determine the DM, ash, and 15 N abundance.Samples were stored at −20°C for further processing and analysis.
For the evaluation of the kinetics of the metabolic activity of each fermenter, samples from the ruminal content within each fermenter were collected and measured.The pH of the ruminal content of each fermenter was measured at 0, 1, 2, 4, 6, and 8 h after the morning feeding (0800 h) on d 8, 9, and 10 of the experimental periods using a portable pH meter (Thermo Scientific Orion Star A121, Thermo Fisher Scientific Inc.).The pH meter was calibrated on each experimental period using standard pH buffers 4, 7, and 10 (SB105, Thermo Fisher Scientific Inc.).
On d 8 to 10 of each experimental period at 0, 1, 2, 4, 6, and 8 h after the morning feeding (0800 h), approximately 10 mL of ruminal content was manually collected from each vessel.The samples were filtered through 4 layers of cheesecloth and stored in 15-mL tubes.These samples were acidified with 100 μL of 50% H 2 SO 4 (vol/vol) within 10 min of the collection and stored at −20°C for further processing and lactate, SCFA, and ammonia (NH 3 -N) analysis.The digesta effluent containers, which consisted of 2 separate containers, one to store the liquid effluent outflow and another to store the solid effluent outflow, were kept in a cold-water bath (4°C) on d 8 to 10 to inhibit further fermentation of the effluent outflow.The effluents of the liquid and solid outflows were pooled daily within each fermenter.Duplicate subsamples of 180 g of the pooled effluent from each fermenter were collected daily and stored at −20°C for further processing and nutrient analysis.At the same time, another sample of the pooled effluents was collected and filtered through 4 layers of cheesecloth.A 40-mL sample was then collected and acidified with 400 μL of 50% H 2 SO 4 (vol/vol) for lactate, SCFA, and NH 3 -N analysis.
At the end of the experimental period, the entire content from each fermenter vessel was collected for the isolation of the bacteria.The following procedure used for bacterial isolation was modified from that described by Krizsan et al. (2010).The total ruminal content from each vessel was collected and blended for 30 s with 200 mL of 0.9% saline solution using a household blender, squeezed through 4 layers of cheesecloth, and rinsed with another 200 mL of 0.9% saline solution.The filtrate was then centrifuged (Allegra X-15R Centrifuge, Beckman Coulter Life Sciences, CA) at 1,000 × g for 10 min at 4°C to remove the residual feed particles.The supernatant was then collected and centrifuged in an ultra-speed centrifuge (Sorvall LYNX 4000 Centrifuge, Thermo Scientific) at 11,250 × g for 20 min at 4°C for isolation of the bacterial pellet.The supernatant was carefully discarded to avoid loss of bacterial pellet, which was resuspended in 200 mL of McDougall's solution for removal of free ammonia (McDougall, 1948).The resuspended solution was then centrifuged (Sorvall LYNX 4000 Centrifuge, Thermo Scientific) at 16,250 × g for 20 min at 4°C.The bacterial pellet was harvested from the third centrifugation and stored at −20°C for further processing and analysis.

Laboratory Analyses
Acidified ruminal samples that were collected for lactate, SCFA, and NH 3 -N concentration analyses were centrifuged (Sorvall LYNX 4000 Centrifuge, Thermo Scientific) at 10,000 × g for 15 min at 4°C.Approximately 2 mL of the supernatant was filtered through a 0.22-μm filter and analyzed for lactate and SCFA using a Merck Hitachi Elite LaChrome HPLC system (L2400, Hitachi, Tokyo, Japan) and a Bio-Rad Aminex HPX-87H column (Bio-Rad Laboratories, Hercules, CA) according to Amaro et al. (2023).Briefly, the column was used in an isocratic elution containing 0.015 M H 2 SO 4 in the mobile phase of HPLC with a UV detector (wavelength 210 nm; L2400, Hitachi) and a flow rate of 0.70 mL/min at 46°C.Indices functioning as markers of protein degradation were calculated, such as BSCFA that included iso-butyrate and iso-valerate, and the calculation of isoacids (IA) that included the BSCFA and valerate, according to Andries et al. (1987).
The remaining supernatant was then used to determine the NH 3 -N concentration in triplicate, according to Broderick and Kang (1980) and adapted to a plate reader by using 2 μL of the sample, 100 μL of phenol solution and 80 μL of hypochloride solution in each well of the microplate.Absorbance was measured in a spectrophotometer (SpectraMax Plus 384 Microplate Reader, Molecular Devices, San Jose, CA) at 620 nm.The inter-and intra-assay coefficient of variability of the analysis were 3.45% and 5.47%, respectively.
Background saliva, background digesta, bacterial pellet, and digesta samples were freeze-dried (FreeZone 6, Labconco).All of these samples and the feed ingredients were then analyzed for DM content, using an air-forced oven at 105°C for 12 h, according to method 930.15 (AOAC, 1990).Ash was determined using a muffle furnace (Isotemp Muffle Furnace, Fisher Scientific) at 600°C for 6 h, according to method 942.05 (AOAC, 1990).A dry subsample of about 0.5 mL (volume) of each of these samples was collected and placed into 2-mL impact-resistant microcentrifuge tubes (Fisher Scientific) and 0.5 mL (volume) of 2-mm zirconia beads (BioSpec Product, Bartlesville, OK) was added, the tube was properly labeled and closed.The sample was then pulverized in a homogenizer (Precellys 24, Bertin, Montigny-le-Bretonneux, France) at 5,500 × g for 10 s for determination of total nitrogen and 14 N and 15 N partitioning.
The samples were loaded into tin capsules (elemental microanalysis, Okehampton, UK) and weighed using a microscale (Excellence Plus XP Micro Balance Mettler-Toledo GmbH, Laboratory and Weighing Technologies, Mississauga, Canada).To avoid NH 3 -N contamination during the analysis, 35 μL of K 2 CO 3 solution (10 g/L) was added to the samples and allowed to dry overnight in a forced-air oven at 40°C.The percent 15 N in dried samples was determined using a mass spectrometer (IsoPrime 100, IsoPrime, Langenselbold, Germany), and the results were expressed as the fractional abundance of isotopic fractions ( 15 N/ 14 N).
For all samples, except for the bacterial pellet, NDF was analyzed following the procedure described by Mertens (2002) with the addition of thermostable α-amylase, sodium sulfite, and corrected for ash in an Ankom200 Fiber Analyzer (Ankom Technology, Macedon, NY).

Calculations for Disappearance of Nutrients and N Metabolism
Nutrient disappearance in the digesta effluent was estimated according to Soder et al. (2013), as follows: where ND is the percentage of nutrient disappearance in a DM basis, NI is the nutrient intake (g/d), NE is the nutrient outflow in the effluent (g/d), NS is the nutrient in the saliva (g/d), and NB is the nutrient in the bacterial pellet (g/d).
Total N (TN) present in the digesta effluent corresponds to the N remaining from the microbial fermentation, and it was subdivided into NH 3 -N flow (ANF), dietary N flow (DNF), and bacterial N flow (BNF).The NAN is a combination of the DNF and BNF.These parameters were calculated according to Calsamiglia et al. (1996) and Bach and Stern (1999): where ANF is the NH 3 -N flow (g/d), NH 3 is the NH 3 -N concentration in the effluent (mg/dL), TE is the volume of total effluent flow (mL), NAN is the nonammonia nitrogen flow (g/d), TN is the total nitrogen in the effluent (g), BNF is the bacterial nitrogen flow (g/d), NAN 15 N is the percentage of atom excess of 15 N in NAN, which is the % atom 15 N in NAN effluent sample minus the % atom 15 N in the background sample (%), and B 15 N is the percentage of atom excess of 15 N in the bacterial pellet.
In addition, the flow of dietary N and microbial efficiency indicators were determined according to Bach and Stern (1999): where DNF is the dietary nitrogen flow (g/d), NAN is the nonammonia nitrogen flow (g/d), BNE is the bacterial nitrogen in effluent (g/d), BE is the bacterial efficiency (g of bacterial N/kg of tOMD), BNF is the bacterial nitrogen flow (g/d), tOMD is the true OM degraded (kg), NE is the efficiency of nitrogen use (%), BNF is the bacterial nitrogen flow (g/d), and TN is the total nitrogen available (g/d).

Statistical Analysis
The experiment was conducted in a duplicated 3 × 3 Latin square, where the fermenter within Latin square was the experimental unit.The residual variance for each continuous dependent variable was modeled and the covariance structure that generated the smallest corrected Akaike's information criterion was individually selected for each variable.Normality of residuals and homogeneity of variance were examined for each continuous dependent variable using the Shapiro-Wilk test from the UNIVARIATE procedure of SAS 9.4 (SAS Institute Inc., Cary, NC) and maximum studentized residue of ± 4 was allowed.
Statistical analysis for the nutrient disappearance, N metabolism, and SCFA variables in the daily outflow were performed using the MIXED procedure of SAS 9.4, using the model where Y ijklm is the observation ijklm, μ is the overall mean, T i is the fixed effect of treatment (i = 1 to 3), D j is the random effect of day of collection (j = 1 to 3), P k is the random effect of period (k = 1 to 3), S l is the random effect of square (l = 1 to 2), S(F) km is the random effect of fermenter (m = 1 to 6) nested within Latin square (l = 1 to 2), and e ijklm is the random residual.The evaluation of the kinetics of SCFA, NH 3 -N, and lactate was carried out using the following model: where Y ijklmn is the observation ijklmn, μ is the overall mean, T i is the fixed effect of treatment (i = 1 to 3), U j is the fixed effect of time (j = 1 to 6), T i × U j is the fixed effect of the interaction of the treatment (i = 1 to 3) and time (j = 1 to 6), D k is the random effect of day of collection (k = 1 to 3), P l is the random effect of period (l = 1 to 3), S m is the random effect of square (m = 1 to 2), S(F) nm is the random effect of fermenter (n = 1 to 6) nested within Latin square (m = 1 to 2), and e ijklmn is the random residual.The treatment effect was evaluated by orthogonal contrasts to depict the effect of partial replacement of SBM with algae (CRT vs. CHL + SPI), in addition, the comparison of algae species was characterized (CHL vs. SPI).Significance was declared at P ≤ 0.05, and tendency was declared at 0.05 < P ≤ 0.10.

Partial SBM Replacement with Algae Biomass
Data for daily fermentation metabolite production is presented in Table 3. Partial replacement of SBM with algae biomass increased daily molar concentration of propionate (P < 0.01) by 15% and reduced the molar concentration of butyrate (P = 0.02), valerate (P = 0.10), and BSCFA (P = 0.09) by 9%, 15%, and 14%, respectively.Greater propionate molar concentration generated a reduction of 14% in the acetate-topropionate ratio (P < 0.01) in the fermentation vessels that had the partial replacement of the SBM with the algae biomass.There was no statistical difference for the daily molar concentration of lactate, acetate, isobutyrate, iso-valerate, IA, and total SCFA.For the daily molar proportion of SCFA, a similar pattern was observed, where fermenters that received a partial re-placement of SBM with algae biomass had a greater molar proportion of propionate (P < 0.01) and a reduction on butyrate (P = 0.06) and valerate (P = 0.05) molar concentration.
The fermentation metabolite kinetics are presented in Table 4 and Figure 1.Fermenters receiving a diet with partial replacement of SBM with algal biomass had a reduction in average NH 3 -N molar concentration (P < 0.01) and an increase in propionate molar concentration (P = 0.03), as well as a reduction in the average molar concentration of iso-butyrate (P = 0.02), butyrate (P = 0.02), valerate (P < 0.001), iso-valerate (P < 0.001), BSCFA (P < 0.01), IA (P < 0.001), and acetate-to-propionate ratio (P = 0.07).There was an 8% increase in molar concentration of propionate (P = 0.03) for fermenters receiving the partial replacement of SBM with algal biomass, as well as a reduction in butyrate (P = 0.04), valerate (P < 0.01), iso-valeric (P < 0.001), BSCFA (P < 0.001), and IA (P < 0.001).There was no statistical difference for pH, molar concentration of lactate, acetate, and total SCFA, or the molar proportion of acetate.Depiction of treatmentby-time interaction of molar concentration of SCFA is presented in Figure 1.There was no treatment-by-time interaction and there was a main effect of time for all variables evaluated.
The nitrogen metabolism variables are presented in Table 5.There was a tendency that fermenters receiving the partial replacement of algal biomass reduced the NH 3 -N concentration (P = 0.06) and flow (P = 0.07) by around 14% as well as an increase in NAN (P = 0.08) of about 4%.There was no statistical difference for TN, BNF, DNF, or measurements of nitrogen utilization efficiency.Data for the degradability of nutrients is presented in Table 6.Fermenters receiving the diets with partial replacement of SBM with algal biomass had an increase of about 12% in NDF degradability (P < 0.001) compared with the control.There were no statistical differences for the degradability of other nutrients.

Comparison Between CHL and SPI
The variables of daily metabolite production are presented in Table 3. Partial replacement with CHL reduced the molar concentrations of iso-butyrate (P = 0.08), iso-valerate (P = 0.02), BSCFA (P = 0.01), and IA (P = 0.01) when compared with partial replacement with SPI.There were no statistical differences for molar concentration of lactate, acetate, propionate, butyrate, valerate, total SCFA, or acetate-to-propionate ratio.When compared with partial replacement of SBM with SPI, CHL increased the molar proportion of acetate (P = 0.02) and reduced the molar concentrations of iso-valerate (P = 0.08), BSCFA (P < 0.01), and IA (P < 0.01).There were no statistical differences for molar proportion of propionate, iso-butyrate, butyrate, or valerate.
Nitrogen metabolism variables are presented in Table 5.When compared with SPI, fermenters receiving the replacement of SBM with CHL had lower NH 3 -N molar concentration (P < 0.01) and flow (P < 0.01) as well as greater daily NAN (P < 0.01) and bacterial (P = 0.05) flow.For those fermenters, the efficiency of nitrogen utilization (P = 0.05) and percentage of nitrogen captured (P = 0.05) were also greater than fermenters receiving the partial replacement of SBM with SPI.There were no statistical differences for TN, DNF, or for ruminal protein degradation parameters.In addition, there were no statistical differences for the degradability of nutrients (Table 6).

Partial SBM Replacement with Algae Biomass
In our study, partial replacement of SBM with algal biomass increased the propionate molar concentration and proportion, while decreasing the butyrate, BSCFA, and IA concentrations and proportions.In addition, there was a reduction in NH 3 -N concentration kinetics that increased NAN flow.Propionate is one of the main SCFA produced by gastrointestinal microbes by 3 different pathways, including the succinate, acrylate, and propanediol pathways (Reichardt et al., 2014).In the rumen, the succinate and acrylate pathways, also called the randomizing and direct pathways, are the  main metabolic pathways representing about 85% and 15% of propionate production, respectively (Owens and Basalan, 2016;Louis and Flint, 2017).
According to Blackburn and Hungate (1963), most propionate is produced by succinate-utilizing bacteria.Succinate is a major end product of the carbohydratefermenting bacteria in the rumen (Sijpesteijn and Elsden, 1952), and an increase in propionate may be due to an increase in succinate production in diets containing algae.It is well established that Chlorella and Spirulina have around 15% of carbohydrates in their composition (Oliveira et al., 1999;Weber et al., 2022).Chlorella have a cellulosic cell wall, whereas Spi-rulina have a peptidoglycan cell wall; both polymers are chemically characterized as carbohydrates (Raji et al., 2020) and with the potential to enter the microbial fermentation process and be converted into succinate and consequently to propionate.
In contrast to algal cell walls that do not have lignin in their composition, SBM (plant cells) cell walls have lignin, a polyphenol polymer that is not degraded by ruminal microorganisms.This inhibits the fermentation of the SBM structural carbohydrates (Knudsen, 1997;Raffrenato et al., 2017), which could explain the greater propionate molar concentration and proportion with diets containing algal biomass. 2 Total N = total N flow (g/d) = NH 3 -N + NAN (Bach and Stern, 1999).7 ENU = efficiency of N use (%) = (g of bacterial N/g of available N) × 100 (Bach and Stern, 1999). 8 Microbial efficiency = g of bacterial N flow/kg of OM truly digested (Calsamiglia et al., 1996).9 Percentage (%) of N that was captured in the microbial biomass.
10 N supply %.Along with greater propionate production from ruminal microorganism fermentation, reductions in BSCFA and IA were also observed.According to Andries et al. (1987), BSCFA and IA are synthesized during the microbial fermentation in the rumen by the oxidative decarboxylation of the branched-chain AA valine, isoleucine, and leucine; they are also end products of the degradation of the AA proline, arginine, lysine, and methionine.During ruminal fermentation, those SCFA may also be used by cellulolytic microorganisms as a precursor for the synthesis of long-chain fatty acids and microbial AA, such as valine, isoleucine, proline, and leucine (Bryant, 1973).
Consequently, changes in the molar concentration of those SCFA may be attributed to a lower AA degradation or greater synthesis of microbial protein and longchain fatty acids.However, from the nitrogen metabolism parameters evaluated, there was a tendency for diets containing algae to reduce NH 3 -N concentration and flow as well as a tendency to increase NAN flow.It is not clear if that the reduction in BSCFA is due to a decrease in AA degradation or synthesis of microbial compounds.Although there is a numerical difference in the RUP, where the control diet had a lower RUP when compared with the Chlorella-containing diet, which could suggest a reduction in the degradation of protein in Chlorella-containing diets.

Comparison Between Chlorella and Spirulina
When contrasting with fermenters fed Spirulinabased diets, fermenters fed Chlorella-based diets had a lower molar concentration of BSCFA and IA, along with lower NH 3 -N concentration and flow, and greater NAN, bacterial nitrogen flow, and efficiency of nitrogen utilization.As mentioned earlier, BSCFA and IA are markers of the degradation of AA from feed sources and can be used to synthesize long-chain fatty acids and microbial AA.Changes in their concentration may be due to a change in the rate of dietary AA degradation or a change in the rate of synthesis of microbial long-chain fatty acids and AA.
Similarly, in our previous experiment (data not published), where diets containing 50% and 100% replacement of SBM with Chlorella or Spirulina using a batch culture were compared, Chlorella-based diets had a lower molar concentration of BSCFA and IA, regardless of the carbohydrate profile of the diet, corroborating with the findings of our current study.Costa et al. (2016), evaluated the supplementation with Chlorella or Spirulina to cattle fed ad libitum speargrass hay and observed that Chlorella-supplemented animals had a lower concentration of iso-butyrate, iso-valerate, and valerate when compared with Spirulina-supplemented animals.
Interestingly, the CHL diet resulted in a numerical reduction in RDP when compared with the SPI diet, which corroborates our results of BSCFA and IA.This suggests that Chlorella protein may be more resistant to proteolysis by ruminal microorganisms, thereby reducing the production of BSCFA and IA.Kose et al. (2017), evaluated the in vitro protein digestibility of Chlorella and Spirulina biomass using intact algae cells and hydrolyzed cells with pancreatic enzymes that would damage the cell wall.Regardless of the cell wall disruption, Chlorella had around 30% lower in vitro protein degradability compared with Spirulina, which corroborates with our results.Similarly, Muys et al. (2019), evaluated the nutritional value of Chlorella and Spirulina biomass from several producers and production batches.They also observed that intact Chlorella cells had a lower in vitro protein degradability when compared with intact Spirulina cells.
In humans and monogastric animals, differences in protein digestibility may be explained by differences in cell wall composition.For instance, Chlorella cells have a cellulosic cell wall, whereas Spirulina cells have a peptidoglycan cell wall (Becker, 2013).The former will make it harder for the proteolytic enzymes to access the proteins.However, in the rumen environment, cellulolytic enzymes are also present and can disrupt the cell wall of Chlorella and proteolytic enzymes can have access to the algae protein.Nevertheless, even with cellulolytic enzymes present, our results suggest that the digestibility of protein may be reduced in Chlorella biomass.We could speculate that those results could be due to the interference of another component in Chlorella that could reduce the activity or efficiency of proteolysis by proteolytic enzymes.In addition, further evaluation of algal protein digestion in ruminants should be carried out and evaluation of abomasal and small intestine digestion and absorption of AA should be considered.

CONCLUSIONS
The replacement of SBM with algae biomass reduces the molar concentration of BSCFA and IA, which are markers of protein degradation, suggesting that algaebased diets may reduce protein degradation in the rumen along with an increase in NDF degradability.Those results are mainly driven by a reduction in molar concentration of BSCFA and IA by the CHL diet.Additionally, when compared with Spirulina, our results demonstrate that Chlorella protein is less degradable in the rumen, reducing NH 3 -N concentration and flow,

Table 2 .
Lobo et al.:DIETARY ALGAE BIOMASS Nutritional composition (% DM) of the soybean meal (SBM), Chlorella pyrenoidosa (CHL), and Spirulina platensis (SPI) Chlorella cells have a circular shape and diameter varying between 2 and 12 μm, whereas Spirulina have cells in a spiral shape measuring 4 μm in length, 6 to Lobo et al.: DIETARY ALGAE BIOMASS

Table 3 .
Lobo et al.:DIETARY ALGAE BIOMASS Effect of partial replacement of soybean meal (SBM) with Chlorella pyrenoidosa (CHL) or Spirulina platensis (SPI) in the experimental diet on fermentation metabolites of daily effluents (24 h) in a dual-flow continuous-culture system

Table 4 .
Lobo et al.:DIETARY ALGAE BIOMASS Effect of partial replacement of soybean meal (SBM) with Chlorella pyrenoidosa (CHL) or Spirulina platensis (SPI) in the experimental diet on fermentation metabolites in a dual-flow continuous-culture system collected at 0, 1, 2, 4, 6, and 8 h after morning feeding Control vs. algae = contrast among control against Chlorella and Spirulina; CHL vs. SPI = contrast between Chlorella against Spirulina.

Table 5 .
Lobo et al.: DIETARY ALGAE BIOMASS Effect of partial replacement of soybean meal (SBM) with Chlorella pyrenoidosa (CHL) or Spirulina platensis (SPI) in the experimental diet on N metabolism in a dual-flow continuous-culture system Control vs. algae = contrast among control against Chlorella and Spirulina; CHL vs. SPI = contrast between Chlorella against Spirulina.

Table 6 .
Effect of partial replacement of soybean meal (SBM) with Chlorella pyrenoidosa (CHL) or Spirulina platensis (SPI) in the experimental diet on ruminal degradability of organic matter (OMD), crude protein (CPD), and neutral detergent fiber (NDFD) in a dual-flow continuous-culture system Lobo et al.: DIETARY ALGAE BIOMASS