Differential Modulation of Maize Silage Odor: Lactiplantibacillus plantarum vs. Lactiplantibacillus buchneri Drive Volatile Compound Change via Strain-Specific Fermentation

Abstract

Volatile organic compounds (VOCs) are critical indicators of the metabolic status of whole-plant maize silage (WPMS). However, the impact of inoculating various strains of fermentation agents on VOC changes has not been systematically explored. This study aimed to determine how inoculation with Lactiplantibacillus plantarum and Lentilactobacillus buchneri modulates the VOC profile and odor of WPMS after 90 days. VOCs were extracted by headspace solid-phase microextraction and analyzed by gas chromatography-mass spectrometry (HS-SPME-GC-MS). Key VOCs were screened using the variable importance in projection (VIP) and substantiated by relative odor activity values (rOAV) and odor descriptions. A total of 82 compounds were identified, including 22 esters, 19 alcohols, 3 acids, 9 aldehydes, 2 ethers, 6 hydrocarbons, 4 ketones, 10 phenols, and 8 terpenoids. L. plantarum enhanced green/fruity odors while strain L. buchneri significantly reduced undesirable phenolic and aldehydic compounds. Six key VOCs influencing the odor of WPMS were selected: 4-ethyl-2-methoxyphenol and benzaldehyde, which contribute smoky, bacon, and bitter almond aromas, and (E)-3-hexen-1-ol, benzyl alcohol, (E, E)-2,4-heptadienal and methyl salicylate, which impart green, fruity, and nutty aromas. These findings highlight the effects and contributions of various strain additives on VOCs in WPMS, providing new theoretical insights for regulating the flavor profile of WPMS.

1. Introduction

Silage production rapidly chops green forage after harvest, compacts and seals it, and relies on anaerobic acid fermentation to suppress undesirable microorganisms and conserve nutrients for periods of feed shortage [1,2]. Whole-plant maize has become the predominant forage in dairy cow diets globally and a key feed component in dairy production because it is rich in fermentable carbohydrates, provides stable yields across environments, and exhibits favorable ensiling properties [3,4,5,6].

The fermentation quality of silage has a significant impact on the energy intake and milk production of dairy cows [7]. Prior studies report that elevated levels of certain aldehydes and biogenic amines reduce ruminant palatability [8], and that esters, terpenoids, short-chain fatty acids, and ketones formed during fermentation are associated with reduced feed intake [9]. Volatile organic compounds (VOCs) emissions from silage degrade air quality and contribute to atmospheric pollution [10]. Moreover, dietary nutrients like carbohydrates, fats, and proteins influence milk flavor compound formation via rumen metabolism in dairy cows [11,12]. Therefore, VOCs play an important role in influencing silage quality, animal feed intake, dairy product flavor, and the atmospheric environment.

VOCs in silage arise from microbial metabolism and their interactions along multiple pathways [13,14]. Variations in microbial communities and exogenous pathogen contamination can further alter silage quality and flavor profiles [15]. Because lactic acid bacteria (LAB) additives regulate microbial activity during ensiling, they are widely used in practice [1]. For example, Lentilactobacillus buchneri applied with potassium sorbate improved preservation of high-moisture corn silage [16]. The addition of Bacillus velezensis, Lentilactobacillus brevis, and cocci LAB strains to king grass silage promoted the lactic acid bacteria growth, lowered pH, and improved crude protein preservation [17]. Adding Lentilactobacillus buchneri to corn stover silage can also significantly reduce the content of butyric acid [18]. LAB additives have also been associated with improved feed intake [19]. In summary, adding lactic acid bacteria additives to silage feed can effectively enhance the fermentation performance of silage, thereby improving the feed intake of ruminants. Understanding VOC types, concentrations, and formation mechanisms is essential for evaluating odor profiles, informing strategies in raw material production, processing optimization, and downstream utilization [20]. However, most current research on whole-plant maize silage (WPMS) has focused on fermentation characteristics and nutritional quality, with little research on how different lactic acid bacteria additives affect the VOC composition and characteristic aroma of WPMS.

The application of microbial inoculants during corn silage fermentation may alter the VOC profile, leading to aroma characteristics distinct from those of natural fermentation while improving silage quality. Such improvements could further contribute to the sustainable development of dairy farming and the broader dairy industry. To validate this hypothesis, headspace solid-phase microextraction coupled to gas chromatography–mass spectrometry (HS-SPME-GC-MS) was employed to analyze the VOC profiles of whole-plant maize silage (WPMS) fermented with Lactiplantibacillus plantarum or Lentilactobacillus buchneri for 90 days. Since the odor impact of VOCs depends not only on concentration but also on odor activity [21], the relative odor activity value (rOAV) was applied to quantify the odor contribution of individual VOCs [22]. Multivariate analysis was further used to identify key VOCs in WPMS treated with different additives.

2. Materials and Methods

2.1. Chemicals and Reagents

For the standardised qualitative analysis of VOCs in maize silage, a comparison with reference standards for aromatic compounds. Reference standards were purchased as shown in

Table 1. Standard substance information for VOCs.

CAS	Compounds Name	Purity	Supplier	City	Country
119-36-8	Methyl salicylate	≥99%	Accela	Shanghai	China
4313-03-5	(E, E)-2,4-Heptadienal	98%	Adamas	Shanghai	China
1629-58-9	1-Penten-3-one	96%	Adamas	Shanghai	China
90-05-1	Guaiacol	99%	Adamas	Shanghai	China
928-97-2	(E)-3-Hexen-1-ol	97%	Aladdin	Shanghai	China
2785-87-7	2-Methoxy-4-propylphenol	98%	Aladdin	Shanghai	China
7786-61-0	2-Methoxy-4-vinylphenol	98%	Aladdin	Shanghai	China
2785-89-9	4-Ethyl-2-methoxyphenol	99%	Aladdin	Shanghai	China
93-51-6	Creosol	98%	Ark Pharma	Beijing	China
695-06-7	Gamma-caprolactone	98%	HWRK	Beijing	China
143-08-8	1-Nonanol	98%	Innochem	Beijing	China
584-02-1	3-Pentanol	99%	Innochem	Beijing	China
64-19-7	Acetic acid	99%	Innochem	Beijing	China
98-55-5	Alpha-terpineol	98%	Innochem	Beijing	China
140-11-4	Benzyl acetate	99%	Innochem	Beijing	China
100-51-6	Benzyl alcohol	98%	Innochem	Beijing	China
106-30-9	Ethyl heptanoate	99%	Innochem	Beijing	China
123-66-0	Ethyl hexanoate	99%	Innochem	Beijing	China
97-64-3	Ethyl lactate	98%	Innochem	Beijing	China
106-33-2	Ethyl laurate	99%	Innochem	Beijing	China
123-51-3	Isoamyl alcohol	99%	Innochem	Beijing	China
103-45-7	Phenethyl acetate	99%	Innochem	Beijing	China
60-12-8	Phenylethyl alcohol	98%	Innochem	Beijing	China
151-10-0	1,3-Dimethoxybenzene	99%	J&K	Shanghai	China
122-78-1	Benzeneacetaldehyde	97.5%	J&K	Shanghai	China
123-92-2	Isoamyl acetate	99%	J&K	Shanghai	China
25152-84-5	(E, E)-2,4-Decadienal	>90%	Macklin	Shanghai	China
2408-37-9	2,2,6-Trimethylcyclohexanone	97%	Macklin	Shanghai	China
78-92-2	2-Butanol	≥99.8%	Macklin	Shanghai	China
565-67-3	2-Methyl-3-pentanol	99%	Macklin	Shanghai	China
6032-29-7	2-Pentanol	≥99.5%	Macklin	Shanghai	China
1604-28-0	6-Methyl-3,5-heptadiene-2-one	98%	Macklin	Shanghai	China
61931-81-5	Cis-3-hexenyllactate	98%	Macklin	Shanghai	China
93-89-0	Ethyl benzoate	>99.5%	Macklin	Shanghai	China
2396-83-0	Ethyl hex-3-enoate	≥98%	Macklin	Shanghai	China
123-29-5	Ethyl nonanoate	98%	Macklin	Shanghai	China
97-53-0	Eugenol	99%	Macklin	Shanghai	China
19329-89-6	Isoamyl lactate	>99.5%	Macklin	Shanghai	China
91-20-3	Naphthalene	≥99.5%	Macklin	Shanghai	China
108-95-2	Phenol	≥99.6%	Macklin	Shanghai	China
79-77-6	Beta-ionone	95%	Mreda	Beijing	China
138-86-3	Limonene	95%	Mreda	Beijing	China
106-24-1	Geraniol	98%	Rhawn	Shanghai	China
78-70-6	Linalool	98%	Rhawn	Shanghai	China
13019-20-0	2-Methyl-3-heptanone	98%	Sigma-Aldrich	Shanghai	China
/	N-alkanes (C7~C40)	98%	Sigma-Aldrich	Shanghai	China
6728-26-3	(E)-2-Hexenal	98%	Sigma-Aldrich	Shanghai	China
2548-87-0	(E)-2-Octenal	≥95%	Sigma-Aldrich	Shanghai	China
3777-69-3	2-Pentylfuran	≥98%	Sigma-Aldrich	Shanghai	China
110-93-0	6-Methylhept-5-en-2-one	99%	Sigma-Aldrich	Shanghai	China
142-92-7	Hexyl acetate	99%	Sigma-Aldrich	Shanghai	China
97-54-1	Isoeugenol	99%	Sigma-Aldrich	Shanghai	China
124-07-2	Octanoic acid	>99%	Sigma-Aldrich	Shanghai	China
928-95-0	(E)-2-Hexen-1-ol	>95%	TCI	Shanghai	China
18409-17-1	(E)-2-octen-1-ol	>92%	TCI	Shanghai	China
111-27-3	1-Hexanol	98%	TCI	Shanghai	China
123-07-9	4-Ethylphenol	>97%	TCI	Shanghai	China
100-52-7	Benzaldehyde	>98%	TCI	Shanghai	China
17283-81-7	Dihydro-beta-ionone	>90%	TCI	Shanghai	China
2021-28-5	Ethyl 3-phenylpropanoate	>98%	TCI	Shanghai	China
106-32-1	Ethyl caprylate	>98%	TCI	Shanghai	China
628-97-7	Ethyl palmitate	>97%	TCI	Shanghai	China
66-25-1	Hexanal	>98%	TCI	Shanghai	China
142-62-1	Hexanoic acid	>98%	TCI	Shanghai	China
108-38-3	m-Xylene	>99%	TCI	Shanghai	China
124-13-0	Octanal	>98%	TCI	Shanghai	China
100-42-5	Styrene	>99%	TCI	Shanghai	China
589-98-0	3-Octanol	≥98%	Yuanye	Shanghai	China
122-97-4	3-Phenylpropanol	≥98%	Yuanye	Shanghai	China
432-25-7	Beta-cyclocitral	≥96%	Yuanye	Shanghai	China
17092-92-1	Dihydroactinidiolide	≥98%	Yuanye	Shanghai	China
101-97-3	Ethyl phenylacetate	≥99.5%	Yuanye	Shanghai	China
104-61-0	Gamma-nonanolactone	≥98%	Yuanye	Shanghai	China

.

2.2. Silage Preparation

The WPMS was harvested from a maize field in Tumd Left Banner (111°21′ E, 40°42′ N, Hohhot, China) on 8 September 2024. Cut the dried maize into 2–3 cm lengths using a forage chopper. The maize was treated with selected Lentilactobacillus plantarum (LP, 1 × 10⁶ CFU/g fresh weight), selected Lentilactobacillus buchneri (LB, 1 × 10⁶ CFU/g fresh weight), while an equivalent volume of sterile distilled water was applied as the control (CK). LP and LB were provided by the College of Grassland Science and Technology, China Agricultural University. After thoroughly mixing the silage ingredients, inoculate each batch with 1000 g of bacterial inoculant according to the manufacturer’s instructions. Place the inoculated mixture into vacuum-sealed silage fermentation containers (each measuring 25 × 30 cm). Each treatment was prepared in triplicate, and samples were collected after 90 days of ensiling at 25 °C to determine WPMS quality and VOCs.

2.3. Analysis of Silage Chemical Composition and Fermentation Parameters

WPMS samples were dried in an electric hot-air oven (GZX-9240MBE; Shanghai Boxun Medical and Biological Instruments Co., Ltd., Shanghai, China) at 65 °C for 72 h to determine dry matter (DM) content. Acid detergent fiber (ADF) and neutral detergent fibre (NDF) were determined using the ANKOM fiber analyzer (DELRAL, ANKOM Technology, New York, NY, USA), in accordance with methods NYT1459-2022 [23] and GB/T 20806-2022 [24], respectively. Ash content was determined using the GB/T 6438-2007 [25] method. Starch content was determined using the GB/T 42491-2023 [26] method. After homogenising the sample in distilled water (volume ratio 1:10, weight ratio 1:10), the pH value was determined using a calibrated pH meter (PHS-3C; Shanghai Yidian Scientific Instruments Co., Ltd., Shanghai, China). While the concentrations of lactic acid (LA), acetic acid (AA), propionic acid (PA), and butyric acid (BA) were quantified via high-performance liquid chromatography (HPLC; Agilent 1200, Santa Clara, CA, USA). The chromatography system was equipped with a Shodex RSpark KC-811 column (Shodex: New York, NY, USA; 8 mm × 300 mm), employing a 3 mmol per chloride solution as the mobile phase at a flow rate of 1.0 mL/min. Column temperature and detection wavelength were set at 50 °C and 210 nm, respectively.

2.4. Determination of VOCs by HS-SPME-GC-MS

2.4.1. Extraction of VOCs by HS-SPME

Fermented silage samples were opened and frozen in liquid nitrogen and then immediately chilled at −80 °C refrigerator until analysis. The VOCs in silage were extracted and analyzed by HS-SPME-GC-MS, as previously described, with a slight modification of the method of Liu et al. [27], the heating schedule has been modified. Grind the sample to less than 5 mm using a laboratory grinder, then place 2 g of the sample into a 20 mL headspace vial, with 10 μL internal standard (2-methy-3-heptanone, 0.80 mg/mL). The internal standard selected was 2-methyl-3-heptanone, dissolved in acetone solvent. Its concentration was determined through preliminary testing to align with the anticipated concentration range of the target analyte, ensuring it fell within the linear response range of the gas chromatography-mass spectrometry detector. The vial was preheated in a water bath at 60 °C for 5 min. Subsequently, the SPME fiber (50/30 μm DVB/CAR/PDMS, Supelco, Bellefonte, PA, USA) was extended, exposed to the vial headspace, and allowed to adsorb VOCs at 60 °C for 15 min. Following extraction, the fiber was inserted into the gas chromatograph injector for thermal desorption at 250 °C for 5 min.

2.4.2. GC-MS Analysis

The VOCs in silage samples were analyzed using an Agilent 8890 gas chromatograph (Santa Clara, CA, USA) equipped with a DB-WAX capillary column (30 m × 250 μm × 0.25 μm) (Santa Clara, CA, USA) and coupled to a 5977C mass spectrometer (Santa Clara, CA, USA). Helium was employed as the carrier gas at a constant flow rate of 1 mL/min. The injector temperature was maintained at 250 °C, and samples were introduced using a splitless injection mode. The GC temperature was initially held at 40 °C for 5 min, then ramped to 70 °C at 10 °C/min, followed by a further increase to 210 °C at 4 °C/min, where it was maintained for 1 min. Mass spectrometric analysis was performed using an electron impact (EI) ionization source with an electron energy of 70 eV and an ion source temperature of 230 °C. Full-scan data were acquired over a mass-to-charge ratio (m/z) range of 40–450 amu, with a solvent delay of 4 min.

The VOCs in silage samples were identified by comparing retention indices (RI, determined by n-alkanes C9–C40), the mass spectrometry database provided by the National Institute of Standards and Technology (NIST20) and authentic standards. The concentration of each VOC was quantified by comparing the peak area of the target compound with that of the internal standard [28]. Results are reported as the mean value of three independent replicates. The aroma description was adapted from online websites (https://www.thegoodscentscompany.com and https://www.femaflavor.org/flavor-library accessed on 20 July 2025).

2.4.3. Calculation of rOAV

Odor activity value (OAV) is commonly used to evaluate the contribution of individual aroma compounds to the overall flavor profile [29]. The rOAV [30] can be assessed by the relative concentration of VOCs in terms of its contribution to the total aroma. The rOAV is calculated as follows:

$$\mathrm{rOAV}={\displaystyle \frac{{\mathrm{C}}_{\mathrm{i}}}{{\mathrm{OT}}_{\mathrm{i}}}}$$

(1)

where OT_i represents the odor threshold of the compound in water; and C_i denotes the relative concentration of the VOCs in maize silage. The Compounds with rOAV value above 1 are typically considered to make a significant contribution to the aroma profile, and compounds between 0.1 and 1 play a supplementary role in the overall aroma profile [31,32]. Odor threshold is found in the book Compilations of Odour Threshold Values in Air, Water and Other media (Second Enlarged and Revised Edition) [33]. Given the ethical scrutiny surrounding animal testing and the difficulties inherent in live testing, the odour threshold evaluated by humans represents the most accurate characterisation achievable within the constraints imposed.

2.5. Statistical Analysis

The significance level between fermented silage inoculated with different additive groups was determined using one-way analysis of variance (ANOVA), followed by Duncan’s multiple range test, with SPSS 27.0 software (IBM Corp., Chicago, IL, USA). Data analysis and table generation were performed using Microsoft Office 2021 (Microsoft, Redmond, WA, USA). PLS-DA was conducted using MetaboAnalyst6.0 on https://www.metaboanalyst.ca/home.xhtml (accessed on 29 July 2025) to select VOCs with VIP values > 1. The image was drawn using the OmicStudio tool on https://www.omicstudio.cn/tool (accessed on 29 July 2025). All experiments were conducted with at least three replicates.

3. Results

3.1. Chemical Composition of WPMS

The chemical composition of WPMS from different additive groups is shown in

Table 2. Chemical composition of WPMS treatments.

Items	CK	LP	LB	p-Value
DM (% FW)	28.76 ± 0.78 a	25.09 ± 0.59 b	25.95 ± 0.86 b	0.002
NDF (% DM)	36.46 ± 2.73 a	31.54 ± 0.66 b	32.24 ± 0.60 b	0.022
ADF (% DM)	20.07 ± 1.32 a	17.56 ± 0.25 b	17.72 ± 0.28 b	0.014
Ash (% DM)	3.24 ± 0.06 a	3.53 ± 0.12 a	3.22 ± 0.29 a	0.156
Starch (% DM)	37.52 ± 1.78 a	33.33 ± 0.60 b	31.69 ± 1.27 b	0.004

CK: no addition; LP: screened Lentilactobacillus plantarum; LB: screened Lentilactobacillus buchneri; FW: fresh weight; DM: dry matter; ADF: acid detergent fiber; NDF: neutral detergent fiber. Different lowercase letters denote statistically significant differences among the treatments (p < 0.05).

. There were significant differences in DM, ADF, NDF and starch content among the different additive groups (p < 0.05), while there were no significant differences in ash content (p > 0.05). After 90 days of silage, the DM content of the CK group was 28.76% FW, significantly higher than that of the LP and LB groups (p = 0.002). The NDF contents in the LP and LB groups were 31.54% DM and 32.24% DM, respectively, both significantly lower than that observed in the CK group (p = 0.022). Similarly, ADF levels in the LP and LB treatments were 17.56% DM and 17.72% DM, also showing significant reductions compared with the CK group (p = 0.014). In contrast, the ash content of the CK, LP, and LB groups was 3.24% DM, 3.53% DM, and 3.22% DM, with no statistically significant differences detected (p > 0.05). Compared with the CK group, the starch content in the LP and LB groups are 33.33% DM and 31.69% DM, which was significantly reduced (p = 0.004). This indicating that the addition of Lentilactobacillus plantarum and Lentilactobacillus buchneri significantly reduces the starch content and chemical composition of WPMS.

3.2. Fermentation Parameters of WPMS

The fermentation parameters of WPMS from different treatment groups after 90 days are shown in

Table 3. Fermentation parameters of WPMS treatments.

Items	CK	LP	LB	p-Value
pH	4.21 ± 0.05 a	3.81 ± 0.08 c	3.96 ± 0.05 b	<0.001
LA (% DM)	5.44 ± 0.16 b	6.03 ± 0.21 a	5.69 ± 0.03 b	0.008
AA (% DM)	1.60 ± 0.12 b	1.88 ± 0.03 ab	2.14 ± 0.25 a	0.017
PA (% DM)	0.17 ± 0.02 a	0.12 ± 0.01 b	0.14 ± 0.01 b	0.01
BA (% DM)	0.01 ± 0.01 a	0.01 ± 0.00 a	0.01 ± 0.00 a	0.422

CK: no addition; LP: screened Lentilactobacillus plantarum; LB: screened Lentilactobacillus buchneri; DM: dry matter content, LA: lactic acid, AA: acetic acid, PA: Propionic acid, BA: butyric acid. Different lowercase letters denote statistically significant differences among the treatments (p < 0.05).

. The pH, LA, AA, and PA levels of WPMS were significantly influenced by the type of additive applied (p < 0.05). The pH values for the CK, LP, and LB groups were 4.21, 3.81, and 3.96, respectively, with the CK group exhibiting a significantly higher pH than both additive treatments (p < 0.05). LA content was significantly greater in the LP group compared with the CK and LB groups (p < 0.05). The AA concentration in the LP group exceeded that of the CK group, although this difference was not statistically significant (p > 0.05), and did not differ significantly from the LB group (p > 0.05). Both LP and LB treatments resulted in significantly lower PA levels than the CK group (p < 0.05), suggesting that inoculation with Lentilactobacillus plantarum or Lentilactobacillus buchneri can effectively reduce propionic acid accumulation.

3.3. Analysis of VOCs by HS-SPME-GC-MS

After 90 days of fermentation, HS-SPME-GC-MS analysis was performed on WPMS from the different additive groups. A total of 82 VOCs were identified (see Supplementary Table S1 for detailed information), including 3 acids, 19 alcohols, 9 aldehydes, 22 esters, 2 ethers, 6 hydrocarbons, 4 ketones, 10 phenols and 8 terpenoids. As shown in Figure 1a, the 73 compounds were identified in the CK group, 77 compounds in the LP group and 75 compounds in the LB group. This indicates that adding Lentilactobacillus plantarum to fermented WPMS results in a more varied aroma profile. Among the 82 VOCs, two VOCs were only present in the CK group, including (E, E)-2,4-heptadienal and 3-ethyltoluene. Four compounds were exclusively present in the LP group: Isobutyl alcohol, 2-methyl-3-pentanol, 2,3-butanediol, and heptyl acetate. Octanoic acid, ethyl nonanoate, dihydroactinidiolide, and naphtalene were uniquely present in the LP group. The percentage of different types of VOCs detected in WPMS from different additive groups, as shown in Figure 1b. It is evident that aldehydes account for the largest percentage in the CK group (12%), while in the LB group they account for only 9%. Esters account for the largest percentage in the LP group (27%), while alcohols account for the largest percentage in the LB group (25%), and in the CK and LP groups they account for only 21%. Proportional differences were observed visually; statistical testing by class was not performed. Figure 1c illustrates the changes in volatile compound content across the various additive groups. Overall, the content of various volatile compounds in WPMS differed significantly between the different additives. The total content of all volatile compounds was highest in the LP group and lowest in the CK group.

3.4. Analysis of Key VOCs in WPMS from Different Additives Groups

Unsupervised principal component analysis (PCA) was conducted to evaluate the variation in volatile compounds among WPMS samples from different additive treatments, as illustrated in Figure 2a. The results indicated that PC1 (88.5%) and PC2 (9.4%) accounted for a cumulative 97.9% of the variance, suggesting that inoculation with Lentilactobacillus plantarum and Lentilactobacillus buchneri significantly altered the volatile compounds in the silage. To investigate the contribution of these compounds to the odor of WPMS further, VIP was calculated in the supervised PLS-DA model. In general, volatile compounds with VIP > 1 are regarded as making a significant contribution to the overall aroma profile of the sample [34,35]. As shown in Figure 2b, nine key volatile compounds with a VIP > 1 were identified. These include 4-Ethyl-2-methoxyphenol, benzaldehyde, acetic acid, benzyl alcohol, ethyl hex-3-enoate, (E, E)-2,4-heptadienal, beta-selinene, (E)-3-hexen-1-ol and methyl salicylate. However, these key volatiles should be further validated to determine their contribution to the overall aroma of WPMS in different additive groups.

3.5. rOAV Analysis of Volatile Organic Compounds in WPMS from Different Additive Group

The overall contribution of individual volatile compounds to WPMS depends on their rOAV, which is the ratio of their concentration to the odor threshold in water (Supplementary Material Table S2). Compounds exhibiting rOAV > 1 are typically regarded as having a substantial influence on the overall aroma profile, while those with 1 > rOAV > 0.1 play a supplementary role. Calculations identified 48 volatile compounds with rOAV> 0.1, 32 of which had values greater than 1, as shown in

Table 4. rOAV > 0.1 of VOCs in WPMS.

Number	CAS	Compounds	OT A	Odor Descriptions B	rOAV			VIP-Value
					CK	LP	LB
Alcohols
1	123-51-3	Isoamyl alcohol	4	Burnt, cocoa, floral	13.27	12.37	18.73	0.29
2	111-27-3	1-Hexanol	5.6	Banana, flower, grass	44.86	39.20	29.26	0.77
3	928-97-2	(E)-3-Hexen-1-ol	110	Green	7.33	7.87	5.78	1.30
4	589-98-0	3-Octanol	78	Citrus, moss, mushroom	0.26	0.21	0.29	0.11
5	78-70-6	Linalool	0.22	Coriander, floral, lavender	513.19	668.39	650.57	0.50
6	111-87-5	1-Octanol	125.8	Bitter almond, fat, floral	0.33	0.40	0.39	0.17
7	18409-17-1	(E)-2-Octen-1-ol	40	Green, citrus, vegetable	0.51	0.55	0.46	0.06
8	143-08-8	1-Nonanol	45.5	Fat, floral, green	0.72	0.79	0.71	0.03
9	100-51-6	Benzyl alcohol	2546.21	Boiled cherries, moss, roasted bread	0.11	0.11	0.13	2.07
10	60-12-8	Phenylethyl Alcohol	564.23	Fruit, honey, lilac	0.16	0.15	0.13	0.19
Aldehydes
11	66-25-1	Hexanal	5	Apple, fat, green	2.72	3.23	2.48	0.06
12	6728-26-3	(E)-2-Hexenal	88.5	Green, leafy, fruity	0.21	0.18	ND	0.09
13	124-13-0	Octanal	0.59	Citrus, fat, green	11.78	14.81	14.09	0.04
14	2548-87-0	(E)-2-Octenal	3	Dandelion, fat, fruit	49.89	63.47	37.84	0.84
15	4313-03-5	(E,E)-2,4-Heptadienal	15.4	Fat, nut	5.59	ND	ND	1.85
16	100-52-7	Benzaldehyde	750.89	Bitter almond, burnt sugar, cherry	4.13	4.11	2.54	4.56
17	432-25-7	beta-Cyclocitral	5	Tropical, saffron, herbal	7.32	7.48	6.32	0.02
18	122-78-1	Benzeneacetaldehyde	6.3	Berry, geranium, honey	43.12	45.84	34.44	0.37
19	25152-84-5	(E,E)-2,4-Decadienal	0.03	Coriander, deep fried, fat	204.08	471.6	131.96	0.17
Esters
20	123-92-2	Isoamyl acetate	0.15	Apple, banana, pear	116.59	111.80	238.25	0.18
21	123-66-0	Ethyl hexanoate	5	Apple peel, brandy, fruit gum	35.67	29.15	27.00	0.91
22	142-92-7	Hexyl acetate	115	Apple, banana, grass	1.20	1.09	1.19	0.53
23	106-30-9	Ethyl heptanoate	1.9	Brandy, fruit, wine	5.44	4.99	ND	0.05
24	106-32-1	Ethyl caprylate	19.3	Apricot, brandy, fat	4.63	3.15	3.53	0.85
25	93-89-0	Ethyl benzoate	55.56	Camomile, celery, fat	0.98	1.05	0.81	0.15
26	695-06-7	gamma-Caprolactone	260	Herbal, coconut, sweet	0.12	0.15	0.13	0.11
27	140-11-4	Benzyl acetate	364	Fruit	0.20	0.22	0.19	0.11
28	119-36-8	Methyl salicylate	40	Almond, caramel, peppermint	6.03	7.20	5.71	1.06
29	104-61-0	gamma-Nonanolactone	9.7	Coconut, creamy, waxy	6.90	7.61	4.78	0.16
Hydrocarbons
30	3777-69-3	2-Pentylfuran	5.8	Butter, floral, fruit	2.76	2.40	2.38	0.05
31	100-42-5	Styrene	65	Sweet, balsam, floral	4.82	4.96	5.60	0.84
32	91-20-3	Naphtalene	6	Pungent, dry, tarry	ND	1.49	ND	0.22
33	30364-38-6	Dehydro-ar-ionene	2.5	Licorice	8.25	9.58	10.19	0.04
Ketones
34	1629-58-9	1-Penten-3-one	23	Fish, green, mustard	0.59	0.63	0.42	0.03
35	2408-37-9	2,2,6-trimethylcyclohexanone	100	Floral	0.14	0.14	0.13	0.02
36	110-93-0	6-Methylhept-5-en-2-one	68	Citrus, mushroom, pepper	0.46	0.55	0.53	0.16
Phenols
37	90-05-1	Guaiacol	0.84	Burnt, smoke, woody	73.29	75.50	58.14	0.08
38	93-51-6	Creosol	21	Spicy, clove, vanilla	9.09	10.57	7.56	0.49
39	2785-89-9	4-Ethyl-2-methoxyphenol	89.25	Spicy, smoky, bacon	23.06	20.02	19.10	4.75
40	97-53-0	Eugenol	0.71	Sweet, spicy, clove	110.75	129.32	131.95	0.36
41	123-07-9	4-Ethylphenol	13	Leather, spice, stable	35.81	36.76	27.68	0.91
42	7786-61-0	2-Methoxy-4-vinylphenol	12.02	Clove, curry, spice	2.05	2.20	1.22	0.07
43	2628-17-3	4-Vinylphenol	10	Chemical, sweet	0.59	0.50	0.40	0.05
Terpenoids
44	138-86-3	Limonene	34	Citrus, mint, orange	0.13	0.29	0.30	0.13
45	23726-93-4	Damascenone	0.06	Apple, rose, honey	229.87	316.27	251.57	0.07
46	17283-81-7	Dihydro-beta-ionone	1	Violet-like	10.82	16.65	10.92	0.14
47	106-24-1	Geraniol	6.6	Geranium, lemon peel, peach	0.68	0.85	0.71	0.02
48	79-77-6	beta-Ionone	0.01	Floral, woody	1009.11	1197.32	1176.78	0.02

^A: Odor threshold in water (μg/kg) is found in the book Compilations of Flavour Threshold Values in Water and Other Media (Second enlarged and revised edition). ^B: Odor descriptions adapted from https://www.thegoodscentscompany.com or https://www.femaflavor.org/flavor-library (accessed on 20 July 2025). ND: not detected. CK: no addition; LP: screened Lentilactobacillus plantarum; LB: screened Lentilactobacillus buchneri.

. These 32 volatile compounds include 4 alcohols, 8 aldehydes, 7 esters, 4 hydrocarbons, 6 phenols, and 3 terpenoids. The top 10 compounds include beta-ionone (1009–1197), linalool (513–668), damascenone (229–316), (E, E)-2,4-decadienal (131–471), isoamyl acetate (116–238), eugenol (110–131), guaiacol (58–75), (E)-2-octenal (37–49), 1-hexanol (29–44), benzeneacetaldehyde (34–45). These volatiles mainly contribute to the floral and fruity aromas in WPMS.

The boxplots of the changes in the concentration of nine key VOCs are shown in Figure 3a–i, and combined with the rOAV values and odor descriptions calculated for each substance, an analysis was performed. In the CK, LP and LB groups, 4-ethyl-2-methoxyphenol exhibited strong spicy, smoky and bacon-like odors. Excessive levels could adversely affect the odor of WPMS. Among all volatile compounds, benzaldehyde (4.13, 4.11, and 2.54) had the highest concentration, reaching 3104.84 ± 114.29 μg/kg in the CK group. At high concentrations, it performs bitter almond odor. However, when WPMS was fermented with the addition of LP and LB, the concentrations of these two VOCs and their rOAV values were lower than in the CK group. This indicates that fermentation with bacterial agents can inhibit the production of undesirable odors. Acetic acid typically has a pungent, acidic odor. However, due to its high solubility in water, it contributes very little to the overall odor of WPMS. (E)-3-Hexen-1-ol and beta-selinene often contribute to the fresh, green aroma, with the highest concentrations found in the LP group (p < 0.05). Benzyl alcohol and ethyl hex-3-enoate contribute to the fruity aroma. Methyl salicylate and (E, E)-2,4-heptadienal contribute to aromas reminiscent of nuts, fats and caramel, primarily in the CK and LP groups.

Ultimately, six key VOCs influencing WPMS flavor were selected by combining VIP > 1 and rOAV values > 0.1 (Figure 4a), including 4-ethyl-2-methoxyphenol, benzaldehyde, benzyl alcohol, (E)-3-hexen-1-ol, methyl salicylate and (E, E)-2,4-heptadienal. As the odor thresholds of ethyl hex-3-enoate and beta-selinene in water are unknown, further research is required to establish these thresholds and assess their specific contributions to WPMS flavor. As shown in Figure 4b, a molecular sensory odor wheel constructed for the key VOCs. Of these compounds, 4-ethyl-2-methoxyphenol and benzaldehyde negatively impact the overall flavor of WPMS by imparting smoky, bacon and bitter almond odors. Conversely, benzyl alcohol, (E)-3-hexen-1-ol, methyl salicylate and (E, E)-2,4-heptadienal have a positive impact by providing green, fruity and nutty aromas.

4. Discussion

4.1. Effects of Different Additives on the Chemical Composition of WPMS

The primary nutritional components of silage feed mainly include DM, carbohydrates, crude fiber, etc. Studies have demonstrated that silage inoculated with LAB exhibits superior nutritional value compared to untreated silage [36]. LB converts lactic acid into more acetic acid through heterotrophic fermentation [37]. Acetic acid inhibits spoilage caused by yeast and mold [1]. Consequently, LP and LB showed lower DM (% FW) than CK, without mass balance measurements, we do not infer treatment effects on DM losses. Yang et al. improved the fermentation quality of alfalfa by inoculating LP to increase organic acid content [38]. The observed reduction in NDF content post-fermentation may result from plant-derived hemicellulases and acid hydrolysis of hemicellulose [39]. The NDF and ADF contents of LP and LB inoculated groups were both lower than those of the CK group, consistent with the findings of Lara et al. [39].

4.2. Effects of Different Additives on the Fermentation Parameters of WPMS

pH serves as a key indicator of silage acidity [40]. After 90 days of fermentation, the pH values of WPMS treated with LP and LB were both reduced (p < 0.05). Typically, lower pH and higher lactic acid content signify superior fermentation quality [41]. The LP group displayed the lowest pH and highest lactic acid content, attributable to LP being a homofermentative LAB [42] that predominantly produces lactic acid under anaerobic conditions [43], rapidly lowering pH and suppressing harmful microorganisms [44]. In contrast, LB as an atypical fermentation LAB, converts most of the lactic acid into acetic acid during fermentation [45], causing the acetic acid concentration to rise significantly. Poor fermentation results in the conversion of lactic acid to butyric acid [46], thus, low butyric acid content across treated groups indicates successful fermentation. Propionic acid content is very low in high-quality silage, and the addition of LP and LB to corn silage resulted in a decrease in propionic acid content [35,40], confirming that LAB inoculants enhance fermentation quality.

4.3. Effects of Different Additives on the VOCs of WPMS

The odor of silage feed can more directly reflect changes in its quality, and volatile metabolomics has been widely used for flavor chemistry detection [47]. The VOCs detected in WPMS feed are the result of multiple simultaneous chemical reactions, which are catalyzed by endogenous and exogenous enzymes from naturally occurring or inoculated microorganisms and are influenced by changes in fermentation conditions, such as anaerobic conditions and acidification [48].

Esters dominated across treatments and are plausibly formed via esterification between fermentation acids (e.g., lactic, acetic) and alcohols [37], consistent with Chen et al. [49]. Among the ester s with rOAV > 1 were identified as methyl salicylate, isoamyl acetate, hexyl acetate, ethyl heptanoate, ethyl benzoate, ethyl caprylate, ethyl hexanoate, and gamma-nonanolactone. The LP group exhibited significantly higher ester content than CK and LB groups, enhancing fruity, sweet, and floral aromas [50], positively impacting silage flavor. VIP combined with rOAV highlighted methyl salicylate as a principal contributor, with strong caramel and peppermint aromas, and its content is significantly increased in the LP group, indicating that the addition of LP enhances the flavor of WPMS.

Alcohol compounds are generally produced by yeast or bacteria through the decomposition of soluble sugars in plant materials or lactic acid produced by lactic acid bacteria as a carbon source [39]. Ethanol has been reported as one of the primary volatile organic compounds emitted into the atmosphere from corn silage [10], was not detected at 90 days, alcohols and acids undergo esterification reactions to form esters, thereby enhancing fruity aromas. (E)-3-Hexen-1-ol and benzyl alcohol contribute herbal and fruity aromas. These alcohols are key contributors to the odor of WPMS, with LP addition augmenting their impact. Ketones, ethers, and hydrocarbons also increased with inoculation, implying broader diversification of odor-relevant minors rather than a single compound effect.

Organic acids are important indicators for evaluating fermentation quality [49]. Acetic and hexanoic acids appeared in all groups; octanoic acid occurred only with LP and is associated with sour odor [49]. The LB group had higher acetic acid content because Lentilactobacillus brevis converts lactic acid into more acetic acid, providing a distinct acidic flavor characteristic. Low-intensity odors can be masked by those of higher intensity [51]. With rOAV < 1 for all acids, they exert negligible flavor influence.

Nine aldehydes were detected; benzaldehyde was most abundant (bitter-almond note) [49]. Both LP and LB reduced benzaldehyde, with a larger decrease under LB. Given that aldehydes often derive from lipid oxidation pathways [52] and possess low odor thresholds, their modulation by inoculants is odor-relevant. VIP and rOAV shortlisted benzaldehyde and (E, E)-2,4-heptadienal as key contributors; the LB group suppression of these compounds aligns with attenuation of almond-like and fatty/green notes.

Oxidative degradation of lignocellulosic fibers yielded phenolic compounds like 4-ethyl-2-methoxyphenol, 4-ethylphenol, 2-methoxy-4-vinylphenol, and 2-methoxy-4-propylphenol, which decreased in the LB group but partially increased in the LP group. Microbial conversion of phenolic acids to volatile phenols has been reported in the fermented process [53]. Moreover, 4-ethylphenol and eugenol have been proposed as spoilage markers in silage [54]. VIP and rOAV analyses revealed 4-ethyl-2-methoxyphenol significantly influences WPMS flavor, imparting spicy, smoky, and bacon-like notes. LB addition mitigated these undesirable odors, proving more effective than LP in reducing adverse flavor contributions.

Terpenoids are typically derived from plants themselves. Beta-selinene and alpha-selinene were detected in silage for the first time. LP and LB inoculations increased terpenoid concentrations, imparting citrus, herbaceous, and woody aromas [55], that enhance silage aromaticity. Damascenone, dihydro-beta-ionone, and beta-ionone may be formed due to the oxidative degradation of carotenoids during silage fermentation [47]. However, beta-selinene is considered to be a key compound affecting the odor of WPMS. Its unknown odor threshold precludes precise assessment of its odor contribution. Future research should systematically determine odor thresholds for additional volatiles in aqueous matrices to evaluate their contributions to WPMS flavor.

This study offers important insights into how LP and LB influence the chemical composition, fermentation characteristics, and VOC profiles of WPMS. However, several limitations should be acknowledged. First, the research was conducted under controlled laboratory conditions with a single maize variety and specific fermentation duration (90 days), which may not fully represent field-scale variations influenced by environmental factors, such as temperature fluctuations or microbial contamination. Second, while HS-SPME-GC-MS effectively identified VOCs, the analysis relied on rOAV, but odor thresholds for certain compounds, like beta-selinene, remain unknown, hindering precise assessment of their flavor contributions. Gas chromatography-olfactometry (GC-O) serves as an effective tool for identifying odor-active compounds, potentially revealing additional flavor compounds. Future research may employ this technique to ensure more comprehensive flavor assessments. Additionally, the study focused solely on two LAB strains without comparing them to other additives or combinations, potentially limiting generalizability. No sensory evaluations or animal feeding trials were performed to validate impacts on palatability and intake. Future research should address these gaps through larger scale trials, microbial metagenomics, and threshold determinations to enhance practical applications in dairy farming.

5. Conclusions

This study demonstrates that microbial inoculation can shape the odor-relevant volatile profile of WPMS. Integrating VIP and rOAV analyses, six compounds emerged as the principal odor-active contributors, including 4-ethyl-2-methoxyphenol and benzaldehyde (smoky/bacon-like and bitter-almond notes), benzyl alcohol, (E)-3-hexen-1-ol, methyl salicylate, and (E, E)-2,4-heptadienal (green/fruity/nutty notes). Relative to natural fermentation, Lentilactobacillus plantarum increased green/fruity-associated contributors, whereas Lentilactobacillus buchneri reduced benzaldehyde and phenolic-type contributors that are associated with undesirable notes. These inoculant-specific effects provide a clear answer to the central research question and indicate that inoculant choice is a viable lever to steer the odor profile of WPMS. The present findings offer a compositional basis for subsequent GC-O/sensory validation and intake studies aimed at translating odor modulation into practical feeding benefits.