The swine used in this study were selected from a commercial herd, and were the progeny of a Landrace× Yorkshire sow and a Duroc sire. The experimental samples were collected from a total of 9 sixth month old castrated pigs with a live weight of 95±5 kg. The experiment was conducted in the Ruminant Nutrition and Anaerobe Laboratory under the Department of Animal Science and Technology at Sunchon National University (SCNU), Republic of Korea. The experimental proposals and procedures for the care and treatment of the pigs were approved by the Animal Care Committee of SCNU. The animals were housed at the SCNU research farm which was well equipped with water, feed, and disposal systems, and individual pens (12.7 cm×7.6 cm) with fully slatted floors. The diets were fed to the pigs as mash ad libitum in the experimental period. The nutrient composition of the basal diet was in accordance with the suggestions of the nutrient requirements for swine from NRC (1998) are shown in Table 1.

Fresh pig fecal samples were collected directly from the rectum of the swine and transferred to a thermo flask at 39°C under vacuum condition. Collected feces were mixed with 10% (W/V) salt media under a constant flow of CO₂ (Jensen et al., 1995; Wang et al., 2004). The homogenate was filtered using sterile cheese cloth and transferred 100 ml of filtrate to sterile serum bottles (120 ml) containing one percent soybean meal (SM) and one tenth percent of phytogenic FA anaerobically O₂-free CO₂. The bottles were then sealed with sterile butyl rubber stoppers that were kept in place with metal screw caps. The filled serum bottles were subsequently placed in a HB-201SF shaking incubator (HANBAEK Scientific Co., Korea) set at 50 rpm and 37°C in anaerobic incubations for 12 and 24 h.

The phytogenic FAs used were red ginseng barn powder (Panax ginseng C. A. Meyer, FA1), persimmon leaf powder (Diospyros virginiana L., FA2), ginkgo leaf powder (Ginkgo biloba L., FA3), and oregano lippia seed oil extract (Lippia graveolens Kunth, OL, FA4). All treatments and Con were replicated in three. Total gas production, pH, NH₃-N, hydrogen sulfide (H₂S) and VFA and other metabolites were determined at 12 and 24 h while nitrite (NO₂), nitrate (NO₃), sulfate (SO₄) concentration and PCR-DGGE at 24 h of incubation.

Total gas production (press and sensor machine (Laurel Electronics, Inc., Costa Mesa, CA, USA)), H₂S and NH₃-N (VRAE Multi-Gas Monitor PGM-7840 (RAE Systems Inc., Sunnyvale, CA, USA)), pH (Pinnacle series M530p meter (Schott Instruments D-55122 Mainz, Germany)), and VFAs and other metabolites (high performance liquid chromatography (HPLC) (Agilent Technologies 1200 series, USA) according to the methods of Tabaru et al. (1988) and Han et al. (2005) were measured from each of the serum bottles at different.

Fermented samples (preserved) were centrifuged at 890×g for 15 min at 4°C and the supernatants were then filtered through 0.2-μm cellulose acetate filters. The filtered supernatants were then used for analysis of nitrite, nitrate and sulfate. The nitrite-N concentration was determined using a Griess reagent kit (G-7921, Molecular Probes (2003)) according to manufacturer’s method. The developed nitrite standard and equation were R² = 0.9989 and y = 0.1748x+0.0051, respectively. The absorbance of the nitrite was determined using the OD at 548 nm against reference with aforementioned equation. The nitrate concentration was measured based on the OD at 410 nm following the modified method described by (Jenkins and Medsker, 1964). The OD values were calculated from the equation developed from standard solution y = 0.005x–0.0057 (R² = 0.9973) and values were expressed as mg NO₃-N/L. For determination of the sulfate concentration, the modified turbidimetric method conducted according to the method described by (Kolmert et al., 2000) with absorbance at 420 nm. For calculation, the OD values were fitted with an equation, y = 0.0032x+0.0043 and standard, R² = 0.9911 to obtain an accurate sulfate concentrations.

Fermented samples (preserved) were extracted using Wizard Genomic DNA Purification Kits (Promega, USA). 16S rDNA PCR amplification was performed using 27F and 1492R universal primers. DGGE was performed using a D-Code Universal Mutation Detection System (Bio-Rad, Hercules, CA, USA). To amplify the V3 region of the bacterial 16S rDNA amplicons, 341F-GC and 518R were used (Nübel et al., 1996). Amplicons of the V3 region of 16S rDNA were used for sequence-specific separation by DGGE according to the specifications of Muyzer and Smalla (1998). The DGGE gel was scanned at 400 dpi and similarity indices were calculated for pairs of DGGE profiles. The number of DGGE bands and similarity indices were calculated from the densitometric curves of the scanned DGGE profiles with Molecular Analyst 1.12 software (Bio-Rad) using the Pearson product-moment correlation coefficient (Häne et al., 1993; Simpson et all, 1999) from the Central Microbiology Laboratory of SCNU in Korea.

A total of 10 purified dominant bands were sent to Macrogen, Seoul, Korea for sequencing. Results were compared to Gen-Bank database using the BLAST tool of the National Center for Biotechnology Information (NCBI) and EzTaxon.

Means of triplicates from treatments and control were analyzed by one-way analysis of variance (ANOVA) using the General Linear Models (GLM) procedures of the Statistical Analysis Systems Institute (SAS, 2002). The effects of the FAs on total gas, pH, NH₃-N, H₂S, VFA, NO₂, concentrations were compared to the controls NO₃, and SO₄ significant differences between treatment means were examined using Duncan’s multiple comparison tests. Variability in the data was expressed as pooled mean values ±standard error (SE) and significance was declared at p<0.05 for all variables measured.

Predominant formation of malodorous substances occurs in the large intestine due to microbial degradation of several substrates (Jensen and Hansen, 2006). Total gas (TG) production, pH, ammonia-nitrogen (NH₃-N), hydrogen sulfide (H₂S), nitrite (NO₂⁻), nitrate (NO₃⁻) and sulfate (SO₄⁻⁻) concentrations of in vitro fermented swine fecal slurry added with soybean meal and different phytogenic feed additives was shown in Table 2. Results revealed that TG production increased as incubation time became longer. FA3 had the highest (p<0.05) total gas production among all treatments and control incubated at 12 and 24 h. This agrees with Robinson et al. (1989) findings wherein TG production is directly proportional to incubation time due to increased fermentation over time. Also, TG was found higher compared to our previous results obtained wherein the FAs added but starch was used as substrate (Alam et al., 2012). Moreover, result was also partially similar to Patra et al. (2006) results in which total gas production as well as odorous compounds increased within the duration of incubation period. However, increase in gas production is not an indicator of odor intensity because it depends on what types of odorous compounds are produced.

The pH values on all treatments increased as incubation time became longer while FA4 decreased. The current result obtained on pH value was opposite of our previous research obtained in Alam et al. (2012). This is due to different substrate used during in vitro fermentation. Reduced pH of FA4 from 12 to 24 h incubation was supported by Risley et al. (1992) and Kempen (2001), stating that lowering the pH of urine and subsequent slurry can be beneficial for reducing odor and NH₃-N emissions. Maintaining the proper acid-base balance and buffering capacity of the diet and the intestinal contents may influence the final pH. In addition, Bailey et al. (2002) stated that anaerobic incubation with additive added fermenta caused significant time-dependent decreases in pH from 0 to 24 h. This trend of lowering pH is probably beneficial when it comes to reducing odorous compounds.

NH₃-N concentration of FA2 was lowest (p<0.05) after 12h of incubation followed by FA3, FA1 and control, respectively while FA4 was not detected. Moreover, only control was detected with NH₃-N concentration after 24 h of incubation. Naidu et al. (2002) stated that a novel probiotic effectively reduced sulfur and NH₃-N compounds in vitro. NH₃-N gas concentration reduction might be due to the antimicrobial effect of the FA. Macfarlane and Macfarlane (2003) showed that NH₃-N and odorous branched chain fatty acids are produced from the fermentation of amino acids. Killeen et al. (1998) reported that FAs contain subfractions which have partially antagonistic properties on intestinal urease activity and NH₃-N formation under in vitro conditions. In addition, Duffy et al. (2001) confirmed the existence of active components in extracts which lowered or reduced intestinal urease activity and enzymes involved in the metabolic urea cycle.

H₂S concentration was not detected in FA2, FA3 and FA4 at 12 and 24 h of incubation. Meanwhile, FA1 H₂S concentration was only detected at 12 h but it was detected at 12 and 24 h incubation in control. FA2 to FA₂S. Fakhoury et al. (2000) stated that H₂S had the highest correlation of malodorous sulphur compounds which may be reduced by adding selective FAs to the diet. Moreover, Le et al. (2005) reported that a change in pH may also change the release of other odorous compounds such as hydrogen sulfide. The decreased in pH of FA4 explains the reduced to undetectable amounts of H₂S.

NO₂-N concentration was better in FA supplemented groups than control while NO₃-N concentration was better in FA4. After 24 h of incubation, NO₂-N and NO₃-N concentration was highest (p<0.05) FA4. NH₃-N is nitrified by bacterial enzymes and converted to NO₂ and NO₃. Ward (1996) explained that oxidation of NH₃-N to NO₃ is called nitrification which is the result of actions of the nitrosifyers (NH₃-oxidizing, NO₂ producing bacteria) and nitrifying bacteria (NO₂-oxidizing, NO₃ producing bacteria). Bengtson (2010) also showed the NH₃-N, NO₂, and NO₃ correlation and their conversion through N-cycle. This explains that the conversion of NO₂ or NO₃ from NH₃-N is beneficial for odor reduction during incubation period. FA4 might be a good FA for reducing nitrogen compounds through conversion to NO₂ and NO₃.

Evaluation of SO₄ production is an indirect method used to evaluate H₂S. In the metabolic pathway, the conversion of H₂S to SO₄ is referred to as chemolithotrophic oxidation of sulfur (Kelly, 1982) is favorable to odor reduction; therefore, higher SO₄ production is better. SO₄ concentration was highest (p<0.05) in FA4, FA1 and FA3 followed by FA2 and control, respectively. Hao et al. (1996) reported that sulfur-containing compounds are produced by bacteria through two processes: reduction of SO₄ and metabolism of sulfur-containing amino acids. SO₄ reduction proceeds via either assimilatory or dissimilatory pathways. In the first process, bacteria produce enough reduced sulfur for cell biosynthesis, while in the second process SO₄ is utilized as a terminal electron acceptor and large quantities of sulfide are produced. It was observed in the present experiment that FA treatments were more effective in reducing H₂S to SO₄ than control at 24 h.

Gibson et al. (1988) stated that SO₄-reducing bacteria (SRB) may play an important role in the terminal stages of fermentation in the colon and in particular, their ability to utilize hydrogen can prevent excessive amounts of this gas from accumulating in the colonic lumen. FA added in this experiment might help in the stimulation of those beneficial bacteria, such as sulfate-producing bacteria which reduces the sulfide. This effect might be due to FA’s enhancement of beneficial microbes and anti-microbial action against undesirable microbes (Spoelstra, 1980). Based upon our findings, we confirmed that selected FA groups produced more SO₄ than control. So, FA4 as well as the other FAs are suitable to reduce H₂S by converting it to SO₄.

Jensen and Hansen (2006) and Le et al. (2005) explained that the predominant formation of malodorous substances occurs in the large intestine due to microbial degradation of several substrates or residual feed components. Substrate fermentation depends on microbial activity which may be modified by antimicrobial FAs. VFA and other metabolites production of in vitro fermented swine fecal slurry added with soybean meal and different phytogenic feed additives were shown on Table 3. Formate was only detected in 12 h FA4 in vitro fermenta. Lactate production was highest (p<0.05) in FA3 but propionate and butyrate were found highest (p<0.05) in control and FA2 after 12 h of incubation, respectively. After 24 h of incubation, control has the highest (p<0.05) in lactate and propionate production while FA3 has the highest (p<0.05) butyrate production. Total VFA production was found highest (p<0.05) in FA2 followed by FA3, FA4, FA1 and Con, respectively after 12 h incubation while after 24 h incubation, FA3 was found highest (p<0.05) followed by FA2, FA4, FA1 and Con, respectively. van Beers-Schreurs et al. (1998) explained that the quantity of VFA depends on the amount and composition of the substrate and on the type of microbes present in the cecum. Low VFA and other metabolites concentration indicates a low amount of substrate fermented and microbial activity in the cecum of pigs.

Negative image and similarity index of 16S rRNA DGGE amplified using total genomic DNA extracted from in vitro fermenta of swine fecal slurry added with different phytogenic feed additives was shown in Figure 1. These bands represented dominant species in the community that were enriched by SM and respective FAs. There were three clearly different clusters: I (FA2 and FA3), II (Con and FA1), and III (FA4). The similarity indices in profiles of clusters I, II, and III, were 88%, 100%, and 36%, respectively. FA4 had the lowest similarity among communities. In the previous research of Alam et al. (2012), the same FAs were added using starch as substrate but the bacterial community and similarity index were different from this research. This means that the substrate affects the bacterial community, similarity indices as well as identified dominant bands. FAs provide a subfraction for the bacteria that can utilize SM and affects their colonization of the large intestine as in vitro. The similarity index covered the range of different percentages for the clusters. The different ranges of the three clusters indicate that different FAs had an effect on the microbial communities during in vitro fermentation after 24 h of incubation. This supports the idea that different FAs create a strong selection pressure on the microbial community in vitro. Among the three clusters, cluster III (FA4) had the lowest similarity (36%) which was evidence of a different microbial community. It has been shown that some FAs can beneficially affect the host by selectively stimulating the growth and or activity of one or a limited number of bacteria in the intestine or gut (Tannock, 1999). Overall, different clusters indicate that the bacterial communities were changed with the addition of FAs which might influence the intestinal bacterial communities involved in fermentation. Our results are consistent with the in vitro fermentation parameters and the concepts of Mao et al. (2008).

Identified dominant bands from 16S rRNA DGGE amplified using total genomic DNA extracted from in vitro fermenta of swine fecal slurry added with different feed additives was shown in Table 4. Identity ranges from 96 to 100% sequence similarity with 170 to 196 bp. Most of the identified bands were previously isolated from different digestive parts of different animals. Identified dominant bands were almost similar our previous result obtained (Alam et al., 2012). This is due to the same FAs added which obviously enhanced some bacterial species during fermentation process. There were three dominant bands which were present only in the FA4 treatment group and were not found in other groups. They were identified closest to Eubacterium limosum (ATCC 8486T), Uncultured bacterium clone PF6641 and Streptococcus lutetiensis (CIP 106849T). The three distinct strains present in the FA4 fermented group probably modified the fermenta environment in some way, either utilizing or enhancing the odor reducing bacteria. FA4 had different bacterial diversity compared to control and other treatments and thus explains having lowest odorous compounds. However, the threshold for similarity to be considered as the same species is 97%, according to Stackebrandt and Goebel (1994); therefore, it appears that the sequence of band no. 5 with similarity of <97%, likely represents some different, not yet cultured pig intestinal or fecal bacteria, that has not previously been identified. This is similar to the situation in the human intestine, where most species analyzed by DGGE and sequencing have not been previously identified (Zoetendal et al., 1998). FAs may provide a subfraction for bacteria that can utilize odor compounds, and these odor compound-utilizing bacteria may then accelerate the metabolism of the other intermediate products such as NO₂, NO₃ and SO₄. The reported substantial reduction effect of certain FAs compared to others can be explained by the fact that it elicits detrimental effects on the microbial activities which release malodor from swine manure as well as function as masking agents (Varel and Miller, 2001). Currently, the precise mechanism of how the above additives reduce the odorous compounds is not well documented. A suggested explanation reported by Kim et al. (2008) is a transformation of the microbial species related to odor generation in the slurry. It may be due to changes in pH, characteristics of the substrates, and microbial activities. FA4 of present experiment seems to function similarly. Moreover, herbs and spices are known to exert antimicrobial actions in vitro against important pathogens, including fungi (Özer et al., 2007). The antimicrobial mode of action is considered to arise mainly from the potential of the hydrophobic essential oils to intrude into the bacterial cell membrane, disintegrating membrane structures, and causing ion leakage. As such, FAs may also be beneficial to swine by reducing pathogens along with odor reduction.