- Email: [email protected]
Profiling and characterization of odorous volatile compounds from the industrial fermentation of erythromycin
Environmental Pollution 255 (2019) 113130
Contents lists available at ScienceDirect
Environmental Pollution journal homepage: www.elsevier.com/locate/envpol
Profiling and characterization of odorous volatile compounds from the industrial fermentation of erythromycin* Xiaofang Yang a, *, Ruyuan Jiao a, Xinmeng Zhu a, Shan Zhao a, b, Guiying Liao c, Jianwei Yu a, Dongsheng Wang a, c a State Key Laboratory of Environmental Aquatic Chemistry, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing, 100085, China b Research and Development Center, Beijing Drainage Group Co., Ltd, Beijing, 100124, China c Faculty of Materials Science and Chemistry, China University of Geosciences (Wuhan), Wuhan, Hubei, 430074, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 22 April 2019 Received in revised form 14 August 2019 Accepted 27 August 2019 Available online 12 September 2019
Complaints caused by odors from the fermentation production of pharmaceuticals are common in China. The elimination of odor remains a challenge for the pharmaceutical industry to meet the increasingly strict environment regulations. Erythromycin is a representative antibiotic produced by microbial fermentation. The fermentation exhaust gas of erythromycin fermentation has an unpleasant odor, but the composition of the key odorants has not been identified. The major odorants from the fermentation production of erythromycin API were analyzed by electronic nose, olfactory measurements, gas chromatography-coupled ion mobility spectrometry (GC-IMS) and gas chromatography-mass spectrometry (GC-MS) analysis. Two compounds, 2-methylisoborneol (2-MIB) and geosmin, were identified as the major odorants of erythromycin fermentation. These had not been detected before using only GCMS analysis of exhaust gas. Aldehydes, including hexanal, octanal, decanal, and benzaldehyde, also contribute to the odor. The composition analysis of odorants using the fermentation broth headspace was more efficient and reliable, considering the significant dilution effect of exhaust gas. The concentration of 2-MIB and geosmin in the fermentation broth greatly exceeded their odor thresholds. The production of major odorants started in the early fermentation stage and became significant in the middle stage (30 e70 h). Due to the extremely low odor thresholds of 2-MIB and geosmin, advanced purification may require deodorization of erythromycin fermentation exhaust gas. © 2019 Elsevier Ltd. All rights reserved.
Keywords: GC-IMS HS-SPME-GCMS Fermentation odor Antibiotics MVOCs
1. Introduction Complaints about the fermentation odor from pharmaceutical manufacturers increased in the last decade in China. A large quantity of active pharmaceutical ingredients (APIs) of antibiotics are produced using fermentation process (Anonymous, 1965), for instance, erythromycin thiocyanate. Erythromycin and its derivatives are the most widely used macrolide antibiotics and antiinfective agents (Mironov et al., 2004). With the growing production of erythromycin in China, the consequent volatile organic compounds (VOCs) and odor pollution have been noticed.
* This paper has been recommended for acceptance by Charles Wong. * Corresponding author. No.18 Shuangqing Road, Haidian District, Beijing, 100085, China. E-mail addresses: [email protected] (X. Yang), [email protected] (D. Wang).
https://doi.org/10.1016/j.envpol.2019.113130 0269-7491/© 2019 Elsevier Ltd. All rights reserved.
Erythromycin API is biosynthesized as a secondary metabolite of microorganism growth (Mironov et al., 2004; Staunton and Wilkinson, 1997). The industrial strain of erythromycin Saccharopolyspora erythraea, formerly Streptomyces erythraeus, is an actinomycete (Smith et al., 1962). The cultivation of S. erythraea strains produces odor (Rezanka et al., 2008; Scholler et al., 2002; Wilkins, 1996), however, in the case of industry fermentation production of erythromycin API, the composition and contribution of major odorous compounds has not been clarified. Formation of off-odor compounds by actinomycetes is not uncommon and occurs in the food processing industry, drinking water supplies, aquaculture, and air-quality control (Azaria and van Rijn, 2018; Garcia-Alcega et al., 2017; Persson, 1983). More than 100 microbial volatile organic compounds (MVOCs) in a variety of chemical classes have been detected during actinomycete culture. These include alkanes, alkenes, alcohols, ketones, esters, aromatic hydrocarbons, terpenoids, and sulfur compounds (Citron et al., 2015; Jachymova et al., 2002;
2
X. Yang et al. / Environmental Pollution 255 (2019) 113130
Rezanka et al., 2008; Scholler et al., 2002). Many MVOCs are potent odorants. Geosmin and 1-octen-3-ol have strong musty or earth odors and sulfur-containing compounds often exhibit a rotten egg odor (Bingley et al., 2012; Citron et al., 2015; Scholler et al., 2002). The production of odorous compounds is determined by the physiological characteristics of the microbial strains and influenced by the nutritional and environmental factors of cultivation (Citron et al., 2015). Thus, although the composition of secondary metabolites is complex, MVOCs had been used for detecting pathogenic bacteria in clinical medicine and microbial activity (Gallegos et al., 2017; Garcia-Alcega et al., 2017; Juenger et al., 2012; Taylor et al., 2017). By far, very little information is available on odor MVOC emission during fermentation for antibiotic production. Erythromycin fermentation required extensive aeration. Odorous compounds are volatilized and diffused into the atmosphere with the emission of exhaust gas causing air pollution problems in nearby communities. The fermentation pharmaceutical industry must reduce or eliminate the odor problem to address the complaints from residents. The most often used deodorization techniques include wet scrubbers, photo-oxidation, ozone oxidation, plasma decomposition and adsorption (Andersen et al., 2012; Domeno et al., 2010; Kim et al., 2017; Kwong et al., 2008; Wang et al., 2012; Zhou et al., 2012). However, due to the composition and chemical properties of the odor pollutants were poorly understood, the performance of the deodorization techniques is unsatisfactory and unwarranted. Thus, characterization of the odorous compounds from antibiotic fermentation production is a key step for the development of improved odor abatement technologies. The odorous compounds volatilized from the industrial production of erythromycin were analyzed. The major odorants in the VOC matrix were identified using analytical techniques developed for studying air pollution and odor in the water and food industries. These techniques were applied to fermentation exhaust gas, exhaust gas condensation water, and fermentation broth. Odorous volatile profiles were obtained by electronic nose and gas chromatography-coupled ion mobility spectrometry (GC-IMS) analysis (Gallegos et al., 2017; Gerhardt et al., 2017) and odor concentration measurements. Qualitative identification of volatiles and quantitative analysis of target potent odorants were accomplished by gas chromatography-mass spectrometry (GC-MS) with different pre-concentration methods (Abalos et al., 2002; Cai et al., 2006; Domeno et al., 2010; Godayol et al., 2011; Ochiai et al., 2001). 2. Materials and methods 2.1. Sample collection Fermentation exhaust gas, exhaust gas condensation water and fermentation broth samples were collected from a commercial fermentation antibiotic factory in Northwest China. The size of a single fermenter for erythromycin production was 250 m3, and the ventilation conditions varied during the fermentation process. The average ventilation amount was about 40 m3 h1 of air per 1 m3 of fermentation broth. The exhaust gas samples were taken at the fermentation gas outlet and collected in 8-L polyester bags using a negatively evacuated sampling instrument (Sinodour Co., Ltd, Tianjin, China). The sampling of exhaust gas for olfactory measurement was performed on three consecutive days during a normal fermentation production cycle. It worth mentioning that the exhaust gas of 5 fermenters at different fermentation stages was mixed and discharged at the gas outlet. The water effluent from the off-gas outlet was collected as exhaust gas condensation water samples. The broth samples were taken at time intervals during the fermentation process. All the liquid samples, including the condensation water and the fermentation broth, were collected in
duplicated in amber glass vessels with minimal headspace and stored under 4 C cold conditions. Field blanks and matrix spike samples were included for analysis. 2.2. On-site analysis of exhaust gas An electronic nose analyzer with 10 different metal oxide semiconductor gas sensors was used for on-site recognition of the exhaust gas odor characteristics (AIRSENSE PEN 3, Airsense Analytics GmbH, Germany). The 10 sensors are sensitive to different classes of compounds, as labeled in Fig. 1 (D'Imporzano et al., 2008; Franke et al., 2009). The sensors of the e-nose are positioned in a measurement chamber with 1.8 mL of volume. Each measurement was performed for 100 s and repeated three times for each sampling. The response values of the sensor array recorded in the 80e90 s were averaged to represent the odor characteristics of the sample. The concentrations of H2S, NH3, and CO2 of exhaust gas were measured at the sampling site using a portable multi-gas analyzer with photo-ionization detector (PID) (Eranntex, China). 2.3. Olfactory measurement of exhaust gas Olfactory measurements were performed using the national standard triangular odor bag method for exhaust odor (GB/T 14675-1993) (Ueno et al., 2009; Higuchi, 2013; Higuchi and Shigeoka, 2018). Briefly, the measurement was carried out with a panel consisted of six qualified odor panelists. For a positive measurement, the odor panelist needs to judge the presence of odor in one of three odor bags. The odor bag was 3-L and filled with odor free air. Odorous exhaust gas was injected into one of the three bags. If the panelist chose the correct odor bag that contained diluted sample, the sample was diluted further for the next round of testing until the panelist gave an incorrect judgement. The individual threshold of the panelist to the sample was determined from the dilution factors of the last correct judgement and the first incorrect judgement. The odor concentration was calculated from the individual thresholds of the panel. Thus, the odor concentration was expressed as a dimensionless dilution factor of the gas sample, by which the odor of the sample could not likely be sensed. All the samples were measured as soon as possible, normally within 36 h after sampling. 2.4. GC-IMS analysis Gas chromatography combined with ion mobility spectrometer (GC-IMS) (G.A.S, Germany) was used for rapid measurement of volatile compounds in the gas samples. The liquid samples were
Fig. 1. The e-nose response outputs (G/G0) of erythromycin fermentation exhaust gas.
X. Yang et al. / Environmental Pollution 255 (2019) 113130
measured using a static headspace sampling method with an automatic sampler unit (CTC, Switzerland) followed by GC-IMS analysis (FlavourSpec, G.A.S, Germany) (Denawaka et al., 2014). The GC system was equipped with a FS-SE-54-CB capillary column (15 m 0.53 mm 1 mm) and a heated splitless injector operated at 150 C. A 1 mL sample of the gas was injected into the heated injector for analysis. The liquid sample was incubated at 65 C for 15 min with speed agitation of 500 rpm and 500 mL of head space gas was sampled with a heated syringe. Chromatographic separation was performed under isothermal conditions at 40 C for gas and condensation water samples and 60 C for fermentation broth samples. High purity nitrogen was used as the carrier gas with a varied flow rate (5e100 mL/min for 40 C column measurements, 2e130 mL/min for 60 C column measurements). The drift tube was operated at a constant temperature of 45 C with a nitrogen flow of 150 mL/min. The IMS cell was operated in positive ion mode. Decanal, 2-nonanone, 2-octanone, 2-heptanone, 2-hexanone, 2pentanone, 2-butanone purchased from J&K and Supelco were used as external reference compounds for composition identification analysis. The detection of 2-MIB and geosmin was checked with standard substances purchased from Sigma-Aldrich. The GC-IMS spectra data were recorded as a 3D array including retention time (seconds), drift time (ms) and ion current signal intensity (mV) and displayed as a 2D topographic map. The spectra were normalized relative to the reaction ion peak (RIP) position to compensate for instrumental variability. The measurements determined under comparable analytical conditions were used to perform fingerprint profile alignment and multiple data analysis. The qualitative analysis was carried out using the GC-IMS library provided by G.A.S. The spectral data were extracted from the data set and applied to principal component analysis (PCA) for visualization of the differentiation of spectra features between samples. 2.5. GC-MS analysis The chemical components of the exhaust gas samples were preconcentrated using a three-stage cryogenic trapping concentrator (Entech7100, USA) and analyzed with GC separation (7890, Agilent, USA) and MS detection (5975C, Agilent, USA). A 400 mL gas sample was concentrated and injected into the GC-MS system with a DB-5 MS capillary column (60 m 0.32 mm 1.0 mm) (Agilent, USA). Headspace solid-phase micro-extraction (SPME) concentration followed by GC-MS analysis was used for liquid samples. DVB/Car/ PDMS Coated SPME fibers (55/30 mm, Supelco, USA) were selected (Godayol et al., 2011). Before use, the SPME fiber was conditioned in the GC injection port according to Supelco instructions. Headspace SPME extractions were carried out at 60 C for 20 min. After extraction, the SPME fiber was thermally desorbed at 250 C for 3 min in a splitless injection port of the GC for analysis. After each run, the SPME fiber was cleaned by reheating in the injection port. A splitless injection technique was used. GC conditions were as follows: carrier gas He 1.5 mL min1, injection temperature 100 C for gas samples and 250 C for SPME, EI ionization 70 eV and the ion source temperature 230 C. The GC temperature was programmed from 35 C (holding for 2 min) to 150 C at 5 C/min, then to 220 C at a rate of 15 C/min and held at 220 C for 7 min. The full-scan mode between 15 and 350 amu was conducted for qualitative analysis and the SIM mode was used for quantitative analysis.
3
3. Results and discussion 3.1. Odor characteristics of fermentation exhaust gas The fermentation odor is difficult to describe due to the complex composition of the odorous compounds and the combination effects between odorants on olfaction (Li et al., 2018; Scholler et al., 2002). However, the fermentation exhaust gas of erythromycin had a strong musty or earthy odor, which could be easily recognized. The olfactory intensity of the exhaust gas was high, and the odor concentrations (dimensionless dilution factor) varied from 4121 to 231739 (Table 1). The high odor concentration values indicated that the odorants contained in the exhaust gas could still be sensed at low concentrations after high dilution. The conductivity responses of the e-nose sensor array were used directly for odor-quality profiling. Sensors Nos. 2, 6, 7, 8, 9 gave high responses to the exhaust gas (Fig. 1). Sensors 2 and 6 were highly sensitive to a broad range of compounds. Sensors 7, 8, and 9 were selective for sulfur-containing compounds, oxygen-containing organics and terpenes, which usually have low odor threshold concentrations (OTCs) in the range of 1e10 ppb (Nagata and Takeuchi, 1990). Thus, the response intensity of sensor 2 represented the odor intensity, and the other sensors were used to recognize the major groups of odorous compounds. The e-nose responses indicated that the odor intensity was high, and the exhaust gas may contain a certain content of oxygen-containing organics and/or sulfur compounds. 3.2. Chemical profile of the odorous volatiles of fermentation using GC-IMS and GC-MS The topographic plots of GC-IMS spectra in Fig. 2 indicate a complex composition of the volatiles from gas, water and broth samples. The analytes were separated depending on their retention index (RI) (y-axis) and ion mobility (x-axis) (Fig. 2). About 20 compounds were identified among the complex matrix of volatiles in all the samples by analyzing the GC-IMS spectra (Table 2). The most frequently detected compound classes were alcohols, aldehydes and ketones. Several common VOCs were identified in all the gas, water and broth samples. These were 1-propanol, 1-butanol, acetone, 3-pentanone, hexanal, benzaldehyde, and 2methylisoborneol (2-MIB). Compounds with low molecular weight or high vapor pressure, including ethanol, acetone, 3pentanone, appeared in the low RI region (RI < 700 under the conditions in this study) and showed strong signal intensity. Compounds with medium RI and high RI (RI > 950), including hexanal, benzaldehyde, and 2-MIB, appeared in the upper part of the plots, which could be considered a fingerprint of the samples. Geosmin and 2-MIB were detected in the headspace of liquid samples, and these compounds had not been previously reported as fermentation odorants. The 2-MIB is a secondary metabolite of actinomycetes and has a strong musty odor (Persson, 1980b; Scholler et al., 2002). It exhibits a camphoraceous odor at high aqueous concentrations (>10 mg,L1) (Persson, 1980a). Geosmin is also a known microbial metabolite with an earthy odor and a very low OTC (ca. 50 ng,m3 in air) (Agus et al., 2011; Gerber and Lechevalier, 1965). The presence of 2-MIB and geosmin explained the strong musty or earthy odor character of erythromycin exhaust
Table 1 Odor concentration and inorganic components detected in fermentation exhaust gas. Sample quantities
OC Min. value
OC Max. value
OC Geometric average
Relative humidity (%)
CO2 (mg/m3)
H2S (mg/m3)
NH3 (mg/m3)
26
4121
231739
24023
>85
>1300
0.52 ± 0.36
1.00 ± 0.43
4
X. Yang et al. / Environmental Pollution 255 (2019) 113130
gas. Geosmin and 2-MIB were produced and accumulated in broth and volatilized during ventilation. They were diluted to a low concentration in the exhaust gas, and then partly re-condensed into a water phase in the pipelines. The differences in the signal intensity supported the deduction that the concentration of 2-MIB and geosmin in the off-gas was much lower than in the headspace of liquid samples, and the concentration of geosmin was probably lower than that of 2-MIB (Fig. 2). The GC-MS qualitative analysis results are shown in Table 2. Various volatile substances of different chemical groups were detected, including alcohols, ketones, and hydrocarbons. The GCMS results confirmed the presence of 2-MIB and geosmin in the headspace of fermentation broth, however, the terpenoids were not detected in the exhaust gas. This was probably because the concentrations of 2-MIB and geosmin in the exhaust gas were lower than the method detection limits. In addition, various hydrocarbons, mainly alkanes and alkylated benzenes, were only detected in exhaust gas using GC-MS but not in the headspace of liquid samples. These hydrocarbons may come from the ambient air that was used for broth aeration. No hydrocarbon was detected by GC-IMS due to the positive ion measurement mode that was used (Denawaka et al., 2014). The chemical concentrations of volatiles detected in exhaust gas are plotted in Fig. 3. The total VOCs concentration of fermentation exhaust gas was rather low compared with other industrial waste gases, such as paint industry and petrochemical industry (Domeno et al., 2010; Kamarulzaman et al., 2019; Qi et al., 2019; Schiavon et al., 2017). However, the odor concentration of fermentation offgas was very high (Table 1), suggesting the presence of odorants with extremely low OTCs. The air odor thresholds of 2-MIB and geosmin are at the ng$m3 level (106 ppm), which means that 50 ng m3 of 2-MIB and geosmin in air could still be smelled by humans. Thus, the presence of 2-MIB and geosmin even at very low chemical concentrations could cause odor problem. Combining the information obtained from the different analytical methods, the major odorants identified from erythromycin fermentation are 2MIB and geosmin. Aldehydes, such as hexanal, octanal, decanal, and benzaldehyde also contributed to the odor. Although ketones and esters were frequently detected, their contribution to odor are generally less significant than aldehydes. The composition analysis of fermentation odorous volatiles using broth headspace instead of exhaust gas is more efficient and reliable, considering the subsequent dilution effect of air ventilation. The low odor threshold characters of the major odorants pose a challenge for deodorization of the fermentation off-gas. The odor activity value of a compound is determined by dividing the chemical concentration of the compound by its OTC (Shen et al., 2012; Wu et al., 2017). It means the concentration of 2-MIB and geosmin has to be reduced to extremely low level to significantly deodorize the off-gas. However, most of the currently used technologies, such as scrubbing combined O3 oxidation or photooxidation, zeolite or carbon materials adsorption, and biological technologies could only partly remove the odor or otherwise operated at-all-cost (Estrada et al., 2013). New techniques or advanced integrated techniques are needed for better purification performance. 3.3. Variation of major odorants during fermentation A rapid profile analysis of the organic volatiles during the fermentation period was performed by displaying the GC-IMS spectra fingerprint information in a gallery plot (Fig. 4). The composition of volatiles detected in the headspace of broth varied with fermentation time. Some compounds that were detected in the broth at 16th hours were not detected in the broth samples
Fig. 2. Topographic plots of GC-IMS spectra of (a) exhaust gas, (b) exhaust gas condensation water, and (c) fermentation broth of erythromycin. RI is the retention index of the compounds.
X. Yang et al. / Environmental Pollution 255 (2019) 113130
5
Table 2 Number of identified VOCs and frequently detected compounds by GC-MS and GC-IMS. Samples
Exhaust gas
Analytical method
GC-MS
Alcohols Ethanol 1-propanol 1-butanol 1-pentanol 1-hexanol Aldehydes methyl-propanal Hexanal Octanal Decanal benzaldehyde Ketones Acetone 2-butanone 3-pentanone 4-methyl-2-pentanone Terpenoids 2-MIB Geosmin
2 ✓ ✓
Hydrocarbons Carboxylic acids and esters Heterocyclics S-compounds Total a b c
1
2 ✓ ✓
18 1 1 1 26
ConWatera
Broth
Exhaust gas
ConWatera
Broth
5
6 ✓ ✓
Odor threshold value (mg$m3)b
GC-IMS 5
3
✓ ✓ ✓ ✓ 5 ✓ ✓
✓
✓ ✓ 2 ✓ ✓
6 ✓ ✓ ✓ ✓ 2 ✓ ✓
2 ✓ ✓
2 ✓ ✓
3
1
4 ✓ ✓ ✓
✓ ✓ ✓ ✓ 3 ✓ ✓
✓ ✓ 4
✓ 5 ✓
✓ 5 ✓
✓ 5 ✓
✓ ✓ 1 ✓
✓ ✓ 1 ✓
✓ ✓ 2 ✓ ✓
✓ 6 ✓ ✓
4
17
14
20
14
✓ ✓
0.52 0.094 0.038 0.1 0.006 0.00035 0.00028 0.00001 0.00040 4.5c 42 0.44 0.028 0.17 0.006e0.015c 0.00005 0.001e0.01c
1 2 2 22
ConWater, condensation water of exhaust gas. Odor thresholds measured by the triangle odor bag method (Nagata and Takeuchi, 1990). Odor thresholds in water (mg$L1) (Persson, 1980b; Young et al., 1996).
Fig. 3. Concentration of volatiles identified in fermentation exhaust gas.
with longer fermentation time (Fig. 4a), and some compounds such as 1-pentonal and 2-MIB achieved their maximum concentrations during the middle stage of fermentation (Fig. 4b). Other compounds continuously increased as fermentation progressed. Without identification of all the volatiles, the gallery plot of GC-IMS spectra can provide fingerprint information on chemical composition of VOCs, especially oxygenated VOCs. To extract more information from the GC-IMS spectra, principal component analysis (PCA) was carried out using the relative peak heights of 53 selected signals for each sample as variables. Two components could explain 55% of the total variance. The observations, namely the samples taken at different fermentation stages, were grouped as shown in the score plot (Fig. 5a). The broth at the
initial fermentation stage (16 h) was clearly separated from other samples, and the samples with longer fermentation times were grouped more closely. These results show that the emission pattern of volatiles is related to the fermentation stage. The correlation of each principal component with the variables is shown in the loading plot in Fig. 5b. Clearly, 2-MIB, acetone and several unidentified variables (07, 56, 67, 70) are closely correlated to observations 32 and 45, which were broth after 32 and 45 h of fermentation. However, broth with a longer fermentation time was better correlated with the geosmin and benzaldehyde content. The variations of major odorant concentrations in broth with fermentation time were measured by HS-SPME-GC-MS. Fig. 6 shows that the concentration of 2-MIB was higher than 30 mg L1 for all the fermentation broths, which was 3000 times higher than its odor threshold in water. The maximum concentration of 2-MIB was 150 mg L1 at about 60 h of fermentation. The concentration of geosmin ranged from 25 to 55 mg L1 and significantly exceeded its OTC (0.001e0.01 mg L1). The concentrations of other aldehydes showed peak values during the middle fermentation stage. These results indicated that the production of major odorants began in the early stage of fermentation and became significant in the middle (30e70 h) stage of fermentation. The excretion of terpenoids is correlated to the differentiation of microbial strains (Bentley and Meganathan, 1981; Scholler et al., 2002). During differentiation, excretion of terpenoids occurred with the growth of aerial mycelium and sporulation. The compound 2-MIB is produced by the addition of a methyl group to a monoterpene precursor and geosmin is produced from the sidechain cleavage of a sesquiterpene precursor (Bentley and Meganathan, 1981). Therefore, the excretion of odorous compounds coinciding with desired secondary metabolites is vital and currently unavoidable for Saccharopolyspora erythraea cultivation. Due to the extremely low odor thresholds of 2-MIB and geosmin, new techniques or advanced purification processes are needed for
6
X. Yang et al. / Environmental Pollution 255 (2019) 113130
Fig. 4. Gallery plot of GC-IMS spectra for volatiles from erythromycin broth at different cultivation times. The original gallery plot was divided into two plots (a) and (b)to fit in the display.
Fig. 5. Principal component analysis of erythromycin broth headspace volatiles measured using GC-IMS. (a) Score plot showing the correlation between broths with different fermentation hours, and (b) loading plot showing the correlation of the compounds detected from each broth.
X. Yang et al. / Environmental Pollution 255 (2019) 113130
7
References
Fig. 6. Variations of odorant concentrations of erythromycin broth with fermentation time measured using SPME-GC-MS.
the satisfactory deodorization of erythromycin fermentation exhaust gas. In addition, the formation of odorous compounds is also influenced by the culture conditions, such as medium composition, temperature, pH and oxygen content (Dionigi et al., 1996). The generation of MVOCs could vary significantly among different fermentation processes. Further investigation of odorants production during the fermentation process is required to increase understanding of odor emission patterns and the identification of major fermentation odorants. 4. Conclusions The fermentation exhaust gas of erythromycin had a strong musty or earthy odor. The olfactory intensity of the exhaust gas was high, and the odor concentrations varied from 4121 to 231739. Geosmin and 2-MIB were identified as the major odorants of erythromycin fermentation. Aldehydes, including hexanal, octanal, decanal, and benzaldehyde also contributed to the fermentation odor. A rapid profile analysis of the organic volatiles using GC-IMS spectra and principal component analysis (PCA) indicated that the emission pattern of volatiles is related to the fermentation stage. The production of major odorants was initiated in the very early stage of fermentation and became significant in the medium stage (30e70 h) of fermentation. The extremely low odor threshold values of the major odorants were responsible for the strong odor associated with erythromycin fermentation. More effective treatment processes are needed for the deodorization of erythromycin fermentation exhaust gas. Conflicts of interest No conflict of interest. Acknowledgments Financial support from the National Natural Science Foundation of China (51778604, 21577160) is gratefully acknowledged. The authors acknowledge assistance from Zhang Hao from G.A.S. GmbH for GC-IMS analysis, and Song RN and Lv MH for helping with samples collection and analysis. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.envpol.2019.113130.
Abalos, M., Prieto, X., Bayona, J.M., 2002. Determination of volatile alkyl sulfides in wastewater by headspace solid-phase microextraction followed by gas chromatography-mass spectrometry. J. Chromatogr. A 963, 249e257. Agus, E., Lim, M.H., Zhang, L.F., Sedlak, D.L., 2011. Odorous compounds in municipal wastewater effluent and potable water reuse systems. Environ. Sci. Technol. 45, 9347e9355. Andersen, K.B., Feilberg, A., Beukes, J.A., 2012. Use of non-thermal plasma and UVlight for removal of odour from sludge treatment. Water Sci. Technol. 66, 1656e1662. Anonymous, 1965. Fermentation processes - antibiotics. Biotechnol. Bioeng. 7, 29e&. Azaria, S., van Rijn, J., 2018. Off-flavor compounds in recirculating aquaculture systems (RAS): production and removal processes. Aquacult. Eng. 83, 57e64. Bentley, R., Meganathan, R., 1981. Geosmin and methylisoborneol biosynthesis in streptomycetes. Evidence for an isoprenoid pathway and its absence in nondifferentiating isolates. FEBS Lett. 125, 220e222. Bingley, G.D., Verran, J., Munro, L.J., Banks, C.E., 2012. Identification of microbial volatile organic compounds (MVOCs) emitted from fungal isolates found on cinematographic film. Anal Methods-UK 4, 1265e1271. Cai, L.S., Koziel, J.A., Lo, Y.C., Hoff, S.J., 2006. Characterization of volatile organic compounds and odorants associated with swine barn particulate matter using solid-phase microextraction and gas chromatography-mass spectrometryolfactometry. J. Chromatogr. A 1102, 60e72. Citron, C.A., Barra, L., Wink, J., Dickschat, J.S., 2015. Volatiles from nineteen recently genome sequenced actinomycetes. Org. Biomol. Chem. 13, 2673e2683. D'Imporzano, G., Crivelli, F., Adani, F., 2008. Biological compost stability influences odor molecules production measured by electronic nose during food-waste high-rate composting. Sci. Total Environ. 402, 278e284. Denawaka, C.J., Fowlis, I.A., Dean, J.R., 2014. Evaluation and application of static headspace-multicapillary column-gas chromatography-ion mobility spectrometry for complex sample analysis. J. Chromatogr. A 1338, 136e148. Dionigi, C.P., Ahten, T.S., Wartelle, L.H., 1996. Effects of several metals on spore, biomass, and geosmin production by Streptomyces tendae and Penicillium expansum. J. Ind. Microbiol. 17, 84e88. Domeno, C., Rodriguez-Lafuente, A., Martos, J.M., Bilbao, R., Nerin, C., 2010. VOC removal and deodorization of effluent gases from an industrial plant by photooxidation, chemical oxidation, and ozonization. Environ. Sci. Technol. 44, 2585e2591. Estrada, J.M., Lebrero, R., Quijano, G., Kraakman, N.J.R., Munoz, R., 2013. Strategiesforodoour control. In: Belgiorno, V., Naddeo, V., Zarra, T. (Eds.), Odour Impact Assessment Handbook. John Wiley & sons, Ltd., UK, pp. 85e124. Franke, W., Frechen, F.B., Giebel, S., 2009. H2S, VOC, TOC, electronic noses and odour concentration: use and comparison of different parameters for emission measurement on air treatment systems. Water Sci. Technol. 59, 1721e1726. Gallegos, J., Arce, C., Jordano, R., Arce, L., Medina, L.M., 2017. Target identification of volatile metabolites to allow the differentiation of lactic acid bacteria by gas chromatography-ion mobility spectrometry. Food Chem. 220, 362e370. Garcia-Alcega, S., Nasir, Z.A., Ferguson, R., Whitby, C., Dumbrell, A.J., Colbeck, I., Gomes, D., Tyrrel, S., Coulon, F., 2017. Fingerprinting outdoor air environment using microbial volatile organic compounds (MVOCs) - a review. Trac. Trends Anal. Chem. 86, 75e83. Gerber, N.N., Lechevalier, H.A., 1965. Geosmin an earthy smelling substance isolated from actinomycetes. Appl. Microbiol. 13, 4. Gerhardt, N., Birkenmeier, M., Sanders, D., Rohn, S., Weller, P., 2017. Resolutionoptimized headspace gas chromatography-ion mobility spectrometry (HS-GCIMS) for non-targeted olive oil profiling. Anal. Bioanal. Chem. 409, 3933e3942. , E., 2011. Odour-causing Godayol, A., Alonso, M., Besalú, E., Sanchez, J.M., Antico organic compounds in wastewater treatment plants: evaluation of headspace solid-phase microextraction as a concentration technique. J. Chromatogr. A 1218, 4863e4868. Higuchi, T., 2013. Determination of odour concentration by the triangular odour bag method. In: Belgiorno, V., Naddeo, V., Zarra, T. (Eds.), Odour Impact Assessment Handbook. John Wiley & sons, Ltd., UK, pp. 43e50. Higuchi, T., Shigeoka, K., 2018. Improvement of interlaboratory evaluation method of olfactometry in Japan. Arch. Environ. Prot. 44, 94e99. Jachymova, J., Votruba, J., Viden, I., Rezanka, T., 2002. Identification of Streptomyces odor spectrum. Folia Microbiol. 47, 37e41. Juenger, M., Vautz, W., Kuhns, M., Hofmann, L., Ulbricht, S., Baumbach, J.I., Quintel, M., Perl, T., 2012. Ion mobility spectrometry for microbial volatile organic compounds: a new identification tool for human pathogenic bacteria. Appl. Microbiol. Biotechnol. 93, 2603e2614. Kamarulzaman, N.H., Le-Minh, N., Fisher, R.M., Stuetz, R.M., 2019. Quantification of VOCs and the development of odour wheels for rubber processing. Sci. Total Environ. 657, 154e168. Kim, K.-H., Szulejko, J.E., Kumar, P., Kwon, E.E., Adelodun, A.A., Reddy, P.A.K., 2017. Air ionization as a control technology for off-gas emissions of volatile organic compounds. Environ. Pollut. 225, 729e743. Kwong, C.W., Chao, C.Y.H., Hui, K.S., Wan, M.P., 2008. Removal of VOCs from indoor environment by ozonation over different porous materials. Atmos. Environ. 42, 2300e2311. Li, W.F., Yang, W.H., Li, J.Y., 2018. Characterization and prediction of odours from municipal sewage treatment plant. Water Sci. Technol. 77, 762e769.
8
X. Yang et al. / Environmental Pollution 255 (2019) 113130
Mironov, V.A., Sergienko, O.V., Nastasyak, I.N., Danilenko, V.N., 2004. Biogenesis and regulation of biosynthesis of erythromycins in Saccharopolyspora erythraea. Appl. Biochem. Microþ 40, 531e541. Nagata, Y., Takeuchi, N., 1990. Measurement of odor threshold by triangle odor bag method. Bull. Jpn. Environ. Sanitation Center 17, 13. Ochiai, N., Takino, M., Daishima, S., Cardin, D.B., 2001. Analysis of volatile sulphur compounds in breath by gas chromatography-mass spectrometry using a threestage cryogenic trapping preconcentration system. J. Chromatogr. B 762, 67e75. Persson, P.E., 1980a. On the odor of 2-methylisoborneol. J. Agric. Food Chem. 28, 1344-1344. Persson, P.E., 1980b. Sensory properties and analysis of two muddy odour compounds, geosmin and 2-methylisoborneol, in water and fish. Water Res. 14, 1113e1118. Persson, P.E., 1983. Off-flavours in aquatic ecosystems eAn Introduction. Water Sci. Technol. 15, 1e11. Qi, Y.Q., Shen, L.M., Zhang, J.L., Yao, J., Lu, R., Miyakoshi, T., 2019. Species and release characteristics of VOCs in furniture coating process. Environ. Pollut. 245, 810e819. Rezanka, T., Prell, A., Sigler, K., 2008. Identification of odorous compounds from nine fermentor-cultivated Streptomyces strains. Folia Microbiol. 53, 315e318. Schiavon, M., Martini, L.M., Corra, C., Scapinello, M., Coller, G., Tosi, P., Ragazzi, M., 2017. Characterisation of volatile organic compounds (VOCs) released by the composting of different waste matrices. Environ. Pollut. 231, 845e853. Scholler, C.E.G., Gurtler, H., Pedersen, R., Molin, S., Wilkins, K., 2002. Volatile metabolites from actinomycetes. J. Agric. Food Chem. 50, 2615e2621. Shen, Y., Chen, T., Gao, D., Zheng, G., Liu, H., Yang, Q., 2012. Online monitoring of volatile organic compound production and emission during sewage sludge
composting. Bioresour. Technol. 123, 463e470. Smith, R.L., Pittenger, R.C., Bungay, H.R., 1962. Growth-biosynthesis relationships in erythromycin fermentation. Appl. Microbiol. 10, 4. Staunton, J., Wilkinson, B., 1997. Biosynthesis of erythromycin and rapamycin. Chem. Rev. 97, 2611e2630. Taylor, C., Lough, F., Stanforth, S.P., Schwalbe, E.C., Fowlis, I.A., Dean, J.R., 2017. Analysis of Listeria using exogenous volatile organic compound metabolites and their detection by static headspace-multi-capillary column-gas chromatography-ion mobility spectrometry (SHS-MCC-GC-IMS). Anal. Bioanal. Chem. 409, 4247e4256. Ueno, H., Amano, S., Merecka, B., Kosmider, J., 2009. Difference in the odor concentrations measured by the triangle odor bag method and dynamic olfactometry. Water Sci. Technol. 59, 1339e1342. Wang, X., Guan, J., Stuetz, R.M., 2012. Catalytic ozonation for odour removal of high temperature alumina refinery condensate. Water Sci. Technol. 66, 1781e1786. Wilkins, K., 1996. Volatile metabolites from actinomycetes. Chemosphere 32, 1427e1434. Wu, C., Liu, J., Zhao, P., Li, W., Yan, L., Piringer, M., Schauberger, G., 2017. Evaluation of the chemical composition and correlation between the calculated and measured odour concentration of odorous gases from a landfill in Beijing, China. Atmos. Environ. 164, 337e347. Young, W.F., Horth, H., Crane, R., Ogden, T., Arnott, M., 1996. Taste and odour threshold concentrations of potential potable water contaminants. Water Res. 30, 331e340. Zhou, X.Z., Zhang, Q., Huang, A.H., 2012. Evaluation of a plant material-based air purifier for removing H2S, NH3 and swine manure odour. Environ. Technol. 33, 2751e2756.
Contents lists available at ScienceDirect
Environmental Pollution journal homepage: www.elsevier.com/locate/envpol
Profiling and characterization of odorous volatile compounds from the industrial fermentation of erythromycin* Xiaofang Yang a, *, Ruyuan Jiao a, Xinmeng Zhu a, Shan Zhao a, b, Guiying Liao c, Jianwei Yu a, Dongsheng Wang a, c a State Key Laboratory of Environmental Aquatic Chemistry, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing, 100085, China b Research and Development Center, Beijing Drainage Group Co., Ltd, Beijing, 100124, China c Faculty of Materials Science and Chemistry, China University of Geosciences (Wuhan), Wuhan, Hubei, 430074, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 22 April 2019 Received in revised form 14 August 2019 Accepted 27 August 2019 Available online 12 September 2019
Complaints caused by odors from the fermentation production of pharmaceuticals are common in China. The elimination of odor remains a challenge for the pharmaceutical industry to meet the increasingly strict environment regulations. Erythromycin is a representative antibiotic produced by microbial fermentation. The fermentation exhaust gas of erythromycin fermentation has an unpleasant odor, but the composition of the key odorants has not been identified. The major odorants from the fermentation production of erythromycin API were analyzed by electronic nose, olfactory measurements, gas chromatography-coupled ion mobility spectrometry (GC-IMS) and gas chromatography-mass spectrometry (GC-MS) analysis. Two compounds, 2-methylisoborneol (2-MIB) and geosmin, were identified as the major odorants of erythromycin fermentation. These had not been detected before using only GCMS analysis of exhaust gas. Aldehydes, including hexanal, octanal, decanal, and benzaldehyde, also contribute to the odor. The composition analysis of odorants using the fermentation broth headspace was more efficient and reliable, considering the significant dilution effect of exhaust gas. The concentration of 2-MIB and geosmin in the fermentation broth greatly exceeded their odor thresholds. The production of major odorants started in the early fermentation stage and became significant in the middle stage (30 e70 h). Due to the extremely low odor thresholds of 2-MIB and geosmin, advanced purification may require deodorization of erythromycin fermentation exhaust gas. © 2019 Elsevier Ltd. All rights reserved.
Keywords: GC-IMS HS-SPME-GCMS Fermentation odor Antibiotics MVOCs
1. Introduction Complaints about the fermentation odor from pharmaceutical manufacturers increased in the last decade in China. A large quantity of active pharmaceutical ingredients (APIs) of antibiotics are produced using fermentation process (Anonymous, 1965), for instance, erythromycin thiocyanate. Erythromycin and its derivatives are the most widely used macrolide antibiotics and antiinfective agents (Mironov et al., 2004). With the growing production of erythromycin in China, the consequent volatile organic compounds (VOCs) and odor pollution have been noticed.
* This paper has been recommended for acceptance by Charles Wong. * Corresponding author. No.18 Shuangqing Road, Haidian District, Beijing, 100085, China. E-mail addresses: [email protected] (X. Yang), [email protected] (D. Wang).
https://doi.org/10.1016/j.envpol.2019.113130 0269-7491/© 2019 Elsevier Ltd. All rights reserved.
Erythromycin API is biosynthesized as a secondary metabolite of microorganism growth (Mironov et al., 2004; Staunton and Wilkinson, 1997). The industrial strain of erythromycin Saccharopolyspora erythraea, formerly Streptomyces erythraeus, is an actinomycete (Smith et al., 1962). The cultivation of S. erythraea strains produces odor (Rezanka et al., 2008; Scholler et al., 2002; Wilkins, 1996), however, in the case of industry fermentation production of erythromycin API, the composition and contribution of major odorous compounds has not been clarified. Formation of off-odor compounds by actinomycetes is not uncommon and occurs in the food processing industry, drinking water supplies, aquaculture, and air-quality control (Azaria and van Rijn, 2018; Garcia-Alcega et al., 2017; Persson, 1983). More than 100 microbial volatile organic compounds (MVOCs) in a variety of chemical classes have been detected during actinomycete culture. These include alkanes, alkenes, alcohols, ketones, esters, aromatic hydrocarbons, terpenoids, and sulfur compounds (Citron et al., 2015; Jachymova et al., 2002;
2
X. Yang et al. / Environmental Pollution 255 (2019) 113130
Rezanka et al., 2008; Scholler et al., 2002). Many MVOCs are potent odorants. Geosmin and 1-octen-3-ol have strong musty or earth odors and sulfur-containing compounds often exhibit a rotten egg odor (Bingley et al., 2012; Citron et al., 2015; Scholler et al., 2002). The production of odorous compounds is determined by the physiological characteristics of the microbial strains and influenced by the nutritional and environmental factors of cultivation (Citron et al., 2015). Thus, although the composition of secondary metabolites is complex, MVOCs had been used for detecting pathogenic bacteria in clinical medicine and microbial activity (Gallegos et al., 2017; Garcia-Alcega et al., 2017; Juenger et al., 2012; Taylor et al., 2017). By far, very little information is available on odor MVOC emission during fermentation for antibiotic production. Erythromycin fermentation required extensive aeration. Odorous compounds are volatilized and diffused into the atmosphere with the emission of exhaust gas causing air pollution problems in nearby communities. The fermentation pharmaceutical industry must reduce or eliminate the odor problem to address the complaints from residents. The most often used deodorization techniques include wet scrubbers, photo-oxidation, ozone oxidation, plasma decomposition and adsorption (Andersen et al., 2012; Domeno et al., 2010; Kim et al., 2017; Kwong et al., 2008; Wang et al., 2012; Zhou et al., 2012). However, due to the composition and chemical properties of the odor pollutants were poorly understood, the performance of the deodorization techniques is unsatisfactory and unwarranted. Thus, characterization of the odorous compounds from antibiotic fermentation production is a key step for the development of improved odor abatement technologies. The odorous compounds volatilized from the industrial production of erythromycin were analyzed. The major odorants in the VOC matrix were identified using analytical techniques developed for studying air pollution and odor in the water and food industries. These techniques were applied to fermentation exhaust gas, exhaust gas condensation water, and fermentation broth. Odorous volatile profiles were obtained by electronic nose and gas chromatography-coupled ion mobility spectrometry (GC-IMS) analysis (Gallegos et al., 2017; Gerhardt et al., 2017) and odor concentration measurements. Qualitative identification of volatiles and quantitative analysis of target potent odorants were accomplished by gas chromatography-mass spectrometry (GC-MS) with different pre-concentration methods (Abalos et al., 2002; Cai et al., 2006; Domeno et al., 2010; Godayol et al., 2011; Ochiai et al., 2001). 2. Materials and methods 2.1. Sample collection Fermentation exhaust gas, exhaust gas condensation water and fermentation broth samples were collected from a commercial fermentation antibiotic factory in Northwest China. The size of a single fermenter for erythromycin production was 250 m3, and the ventilation conditions varied during the fermentation process. The average ventilation amount was about 40 m3 h1 of air per 1 m3 of fermentation broth. The exhaust gas samples were taken at the fermentation gas outlet and collected in 8-L polyester bags using a negatively evacuated sampling instrument (Sinodour Co., Ltd, Tianjin, China). The sampling of exhaust gas for olfactory measurement was performed on three consecutive days during a normal fermentation production cycle. It worth mentioning that the exhaust gas of 5 fermenters at different fermentation stages was mixed and discharged at the gas outlet. The water effluent from the off-gas outlet was collected as exhaust gas condensation water samples. The broth samples were taken at time intervals during the fermentation process. All the liquid samples, including the condensation water and the fermentation broth, were collected in
duplicated in amber glass vessels with minimal headspace and stored under 4 C cold conditions. Field blanks and matrix spike samples were included for analysis. 2.2. On-site analysis of exhaust gas An electronic nose analyzer with 10 different metal oxide semiconductor gas sensors was used for on-site recognition of the exhaust gas odor characteristics (AIRSENSE PEN 3, Airsense Analytics GmbH, Germany). The 10 sensors are sensitive to different classes of compounds, as labeled in Fig. 1 (D'Imporzano et al., 2008; Franke et al., 2009). The sensors of the e-nose are positioned in a measurement chamber with 1.8 mL of volume. Each measurement was performed for 100 s and repeated three times for each sampling. The response values of the sensor array recorded in the 80e90 s were averaged to represent the odor characteristics of the sample. The concentrations of H2S, NH3, and CO2 of exhaust gas were measured at the sampling site using a portable multi-gas analyzer with photo-ionization detector (PID) (Eranntex, China). 2.3. Olfactory measurement of exhaust gas Olfactory measurements were performed using the national standard triangular odor bag method for exhaust odor (GB/T 14675-1993) (Ueno et al., 2009; Higuchi, 2013; Higuchi and Shigeoka, 2018). Briefly, the measurement was carried out with a panel consisted of six qualified odor panelists. For a positive measurement, the odor panelist needs to judge the presence of odor in one of three odor bags. The odor bag was 3-L and filled with odor free air. Odorous exhaust gas was injected into one of the three bags. If the panelist chose the correct odor bag that contained diluted sample, the sample was diluted further for the next round of testing until the panelist gave an incorrect judgement. The individual threshold of the panelist to the sample was determined from the dilution factors of the last correct judgement and the first incorrect judgement. The odor concentration was calculated from the individual thresholds of the panel. Thus, the odor concentration was expressed as a dimensionless dilution factor of the gas sample, by which the odor of the sample could not likely be sensed. All the samples were measured as soon as possible, normally within 36 h after sampling. 2.4. GC-IMS analysis Gas chromatography combined with ion mobility spectrometer (GC-IMS) (G.A.S, Germany) was used for rapid measurement of volatile compounds in the gas samples. The liquid samples were
Fig. 1. The e-nose response outputs (G/G0) of erythromycin fermentation exhaust gas.
X. Yang et al. / Environmental Pollution 255 (2019) 113130
measured using a static headspace sampling method with an automatic sampler unit (CTC, Switzerland) followed by GC-IMS analysis (FlavourSpec, G.A.S, Germany) (Denawaka et al., 2014). The GC system was equipped with a FS-SE-54-CB capillary column (15 m 0.53 mm 1 mm) and a heated splitless injector operated at 150 C. A 1 mL sample of the gas was injected into the heated injector for analysis. The liquid sample was incubated at 65 C for 15 min with speed agitation of 500 rpm and 500 mL of head space gas was sampled with a heated syringe. Chromatographic separation was performed under isothermal conditions at 40 C for gas and condensation water samples and 60 C for fermentation broth samples. High purity nitrogen was used as the carrier gas with a varied flow rate (5e100 mL/min for 40 C column measurements, 2e130 mL/min for 60 C column measurements). The drift tube was operated at a constant temperature of 45 C with a nitrogen flow of 150 mL/min. The IMS cell was operated in positive ion mode. Decanal, 2-nonanone, 2-octanone, 2-heptanone, 2-hexanone, 2pentanone, 2-butanone purchased from J&K and Supelco were used as external reference compounds for composition identification analysis. The detection of 2-MIB and geosmin was checked with standard substances purchased from Sigma-Aldrich. The GC-IMS spectra data were recorded as a 3D array including retention time (seconds), drift time (ms) and ion current signal intensity (mV) and displayed as a 2D topographic map. The spectra were normalized relative to the reaction ion peak (RIP) position to compensate for instrumental variability. The measurements determined under comparable analytical conditions were used to perform fingerprint profile alignment and multiple data analysis. The qualitative analysis was carried out using the GC-IMS library provided by G.A.S. The spectral data were extracted from the data set and applied to principal component analysis (PCA) for visualization of the differentiation of spectra features between samples. 2.5. GC-MS analysis The chemical components of the exhaust gas samples were preconcentrated using a three-stage cryogenic trapping concentrator (Entech7100, USA) and analyzed with GC separation (7890, Agilent, USA) and MS detection (5975C, Agilent, USA). A 400 mL gas sample was concentrated and injected into the GC-MS system with a DB-5 MS capillary column (60 m 0.32 mm 1.0 mm) (Agilent, USA). Headspace solid-phase micro-extraction (SPME) concentration followed by GC-MS analysis was used for liquid samples. DVB/Car/ PDMS Coated SPME fibers (55/30 mm, Supelco, USA) were selected (Godayol et al., 2011). Before use, the SPME fiber was conditioned in the GC injection port according to Supelco instructions. Headspace SPME extractions were carried out at 60 C for 20 min. After extraction, the SPME fiber was thermally desorbed at 250 C for 3 min in a splitless injection port of the GC for analysis. After each run, the SPME fiber was cleaned by reheating in the injection port. A splitless injection technique was used. GC conditions were as follows: carrier gas He 1.5 mL min1, injection temperature 100 C for gas samples and 250 C for SPME, EI ionization 70 eV and the ion source temperature 230 C. The GC temperature was programmed from 35 C (holding for 2 min) to 150 C at 5 C/min, then to 220 C at a rate of 15 C/min and held at 220 C for 7 min. The full-scan mode between 15 and 350 amu was conducted for qualitative analysis and the SIM mode was used for quantitative analysis.
3
3. Results and discussion 3.1. Odor characteristics of fermentation exhaust gas The fermentation odor is difficult to describe due to the complex composition of the odorous compounds and the combination effects between odorants on olfaction (Li et al., 2018; Scholler et al., 2002). However, the fermentation exhaust gas of erythromycin had a strong musty or earthy odor, which could be easily recognized. The olfactory intensity of the exhaust gas was high, and the odor concentrations (dimensionless dilution factor) varied from 4121 to 231739 (Table 1). The high odor concentration values indicated that the odorants contained in the exhaust gas could still be sensed at low concentrations after high dilution. The conductivity responses of the e-nose sensor array were used directly for odor-quality profiling. Sensors Nos. 2, 6, 7, 8, 9 gave high responses to the exhaust gas (Fig. 1). Sensors 2 and 6 were highly sensitive to a broad range of compounds. Sensors 7, 8, and 9 were selective for sulfur-containing compounds, oxygen-containing organics and terpenes, which usually have low odor threshold concentrations (OTCs) in the range of 1e10 ppb (Nagata and Takeuchi, 1990). Thus, the response intensity of sensor 2 represented the odor intensity, and the other sensors were used to recognize the major groups of odorous compounds. The e-nose responses indicated that the odor intensity was high, and the exhaust gas may contain a certain content of oxygen-containing organics and/or sulfur compounds. 3.2. Chemical profile of the odorous volatiles of fermentation using GC-IMS and GC-MS The topographic plots of GC-IMS spectra in Fig. 2 indicate a complex composition of the volatiles from gas, water and broth samples. The analytes were separated depending on their retention index (RI) (y-axis) and ion mobility (x-axis) (Fig. 2). About 20 compounds were identified among the complex matrix of volatiles in all the samples by analyzing the GC-IMS spectra (Table 2). The most frequently detected compound classes were alcohols, aldehydes and ketones. Several common VOCs were identified in all the gas, water and broth samples. These were 1-propanol, 1-butanol, acetone, 3-pentanone, hexanal, benzaldehyde, and 2methylisoborneol (2-MIB). Compounds with low molecular weight or high vapor pressure, including ethanol, acetone, 3pentanone, appeared in the low RI region (RI < 700 under the conditions in this study) and showed strong signal intensity. Compounds with medium RI and high RI (RI > 950), including hexanal, benzaldehyde, and 2-MIB, appeared in the upper part of the plots, which could be considered a fingerprint of the samples. Geosmin and 2-MIB were detected in the headspace of liquid samples, and these compounds had not been previously reported as fermentation odorants. The 2-MIB is a secondary metabolite of actinomycetes and has a strong musty odor (Persson, 1980b; Scholler et al., 2002). It exhibits a camphoraceous odor at high aqueous concentrations (>10 mg,L1) (Persson, 1980a). Geosmin is also a known microbial metabolite with an earthy odor and a very low OTC (ca. 50 ng,m3 in air) (Agus et al., 2011; Gerber and Lechevalier, 1965). The presence of 2-MIB and geosmin explained the strong musty or earthy odor character of erythromycin exhaust
Table 1 Odor concentration and inorganic components detected in fermentation exhaust gas. Sample quantities
OC Min. value
OC Max. value
OC Geometric average
Relative humidity (%)
CO2 (mg/m3)
H2S (mg/m3)
NH3 (mg/m3)
26
4121
231739
24023
>85
>1300
0.52 ± 0.36
1.00 ± 0.43
4
X. Yang et al. / Environmental Pollution 255 (2019) 113130
gas. Geosmin and 2-MIB were produced and accumulated in broth and volatilized during ventilation. They were diluted to a low concentration in the exhaust gas, and then partly re-condensed into a water phase in the pipelines. The differences in the signal intensity supported the deduction that the concentration of 2-MIB and geosmin in the off-gas was much lower than in the headspace of liquid samples, and the concentration of geosmin was probably lower than that of 2-MIB (Fig. 2). The GC-MS qualitative analysis results are shown in Table 2. Various volatile substances of different chemical groups were detected, including alcohols, ketones, and hydrocarbons. The GCMS results confirmed the presence of 2-MIB and geosmin in the headspace of fermentation broth, however, the terpenoids were not detected in the exhaust gas. This was probably because the concentrations of 2-MIB and geosmin in the exhaust gas were lower than the method detection limits. In addition, various hydrocarbons, mainly alkanes and alkylated benzenes, were only detected in exhaust gas using GC-MS but not in the headspace of liquid samples. These hydrocarbons may come from the ambient air that was used for broth aeration. No hydrocarbon was detected by GC-IMS due to the positive ion measurement mode that was used (Denawaka et al., 2014). The chemical concentrations of volatiles detected in exhaust gas are plotted in Fig. 3. The total VOCs concentration of fermentation exhaust gas was rather low compared with other industrial waste gases, such as paint industry and petrochemical industry (Domeno et al., 2010; Kamarulzaman et al., 2019; Qi et al., 2019; Schiavon et al., 2017). However, the odor concentration of fermentation offgas was very high (Table 1), suggesting the presence of odorants with extremely low OTCs. The air odor thresholds of 2-MIB and geosmin are at the ng$m3 level (106 ppm), which means that 50 ng m3 of 2-MIB and geosmin in air could still be smelled by humans. Thus, the presence of 2-MIB and geosmin even at very low chemical concentrations could cause odor problem. Combining the information obtained from the different analytical methods, the major odorants identified from erythromycin fermentation are 2MIB and geosmin. Aldehydes, such as hexanal, octanal, decanal, and benzaldehyde also contributed to the odor. Although ketones and esters were frequently detected, their contribution to odor are generally less significant than aldehydes. The composition analysis of fermentation odorous volatiles using broth headspace instead of exhaust gas is more efficient and reliable, considering the subsequent dilution effect of air ventilation. The low odor threshold characters of the major odorants pose a challenge for deodorization of the fermentation off-gas. The odor activity value of a compound is determined by dividing the chemical concentration of the compound by its OTC (Shen et al., 2012; Wu et al., 2017). It means the concentration of 2-MIB and geosmin has to be reduced to extremely low level to significantly deodorize the off-gas. However, most of the currently used technologies, such as scrubbing combined O3 oxidation or photooxidation, zeolite or carbon materials adsorption, and biological technologies could only partly remove the odor or otherwise operated at-all-cost (Estrada et al., 2013). New techniques or advanced integrated techniques are needed for better purification performance. 3.3. Variation of major odorants during fermentation A rapid profile analysis of the organic volatiles during the fermentation period was performed by displaying the GC-IMS spectra fingerprint information in a gallery plot (Fig. 4). The composition of volatiles detected in the headspace of broth varied with fermentation time. Some compounds that were detected in the broth at 16th hours were not detected in the broth samples
Fig. 2. Topographic plots of GC-IMS spectra of (a) exhaust gas, (b) exhaust gas condensation water, and (c) fermentation broth of erythromycin. RI is the retention index of the compounds.
X. Yang et al. / Environmental Pollution 255 (2019) 113130
5
Table 2 Number of identified VOCs and frequently detected compounds by GC-MS and GC-IMS. Samples
Exhaust gas
Analytical method
GC-MS
Alcohols Ethanol 1-propanol 1-butanol 1-pentanol 1-hexanol Aldehydes methyl-propanal Hexanal Octanal Decanal benzaldehyde Ketones Acetone 2-butanone 3-pentanone 4-methyl-2-pentanone Terpenoids 2-MIB Geosmin
2 ✓ ✓
Hydrocarbons Carboxylic acids and esters Heterocyclics S-compounds Total a b c
1
2 ✓ ✓
18 1 1 1 26
ConWatera
Broth
Exhaust gas
ConWatera
Broth
5
6 ✓ ✓
Odor threshold value (mg$m3)b
GC-IMS 5
3
✓ ✓ ✓ ✓ 5 ✓ ✓
✓
✓ ✓ 2 ✓ ✓
6 ✓ ✓ ✓ ✓ 2 ✓ ✓
2 ✓ ✓
2 ✓ ✓
3
1
4 ✓ ✓ ✓
✓ ✓ ✓ ✓ 3 ✓ ✓
✓ ✓ 4
✓ 5 ✓
✓ 5 ✓
✓ 5 ✓
✓ ✓ 1 ✓
✓ ✓ 1 ✓
✓ ✓ 2 ✓ ✓
✓ 6 ✓ ✓
4
17
14
20
14
✓ ✓
0.52 0.094 0.038 0.1 0.006 0.00035 0.00028 0.00001 0.00040 4.5c 42 0.44 0.028 0.17 0.006e0.015c 0.00005 0.001e0.01c
1 2 2 22
ConWater, condensation water of exhaust gas. Odor thresholds measured by the triangle odor bag method (Nagata and Takeuchi, 1990). Odor thresholds in water (mg$L1) (Persson, 1980b; Young et al., 1996).
Fig. 3. Concentration of volatiles identified in fermentation exhaust gas.
with longer fermentation time (Fig. 4a), and some compounds such as 1-pentonal and 2-MIB achieved their maximum concentrations during the middle stage of fermentation (Fig. 4b). Other compounds continuously increased as fermentation progressed. Without identification of all the volatiles, the gallery plot of GC-IMS spectra can provide fingerprint information on chemical composition of VOCs, especially oxygenated VOCs. To extract more information from the GC-IMS spectra, principal component analysis (PCA) was carried out using the relative peak heights of 53 selected signals for each sample as variables. Two components could explain 55% of the total variance. The observations, namely the samples taken at different fermentation stages, were grouped as shown in the score plot (Fig. 5a). The broth at the
initial fermentation stage (16 h) was clearly separated from other samples, and the samples with longer fermentation times were grouped more closely. These results show that the emission pattern of volatiles is related to the fermentation stage. The correlation of each principal component with the variables is shown in the loading plot in Fig. 5b. Clearly, 2-MIB, acetone and several unidentified variables (07, 56, 67, 70) are closely correlated to observations 32 and 45, which were broth after 32 and 45 h of fermentation. However, broth with a longer fermentation time was better correlated with the geosmin and benzaldehyde content. The variations of major odorant concentrations in broth with fermentation time were measured by HS-SPME-GC-MS. Fig. 6 shows that the concentration of 2-MIB was higher than 30 mg L1 for all the fermentation broths, which was 3000 times higher than its odor threshold in water. The maximum concentration of 2-MIB was 150 mg L1 at about 60 h of fermentation. The concentration of geosmin ranged from 25 to 55 mg L1 and significantly exceeded its OTC (0.001e0.01 mg L1). The concentrations of other aldehydes showed peak values during the middle fermentation stage. These results indicated that the production of major odorants began in the early stage of fermentation and became significant in the middle (30e70 h) stage of fermentation. The excretion of terpenoids is correlated to the differentiation of microbial strains (Bentley and Meganathan, 1981; Scholler et al., 2002). During differentiation, excretion of terpenoids occurred with the growth of aerial mycelium and sporulation. The compound 2-MIB is produced by the addition of a methyl group to a monoterpene precursor and geosmin is produced from the sidechain cleavage of a sesquiterpene precursor (Bentley and Meganathan, 1981). Therefore, the excretion of odorous compounds coinciding with desired secondary metabolites is vital and currently unavoidable for Saccharopolyspora erythraea cultivation. Due to the extremely low odor thresholds of 2-MIB and geosmin, new techniques or advanced purification processes are needed for
6
X. Yang et al. / Environmental Pollution 255 (2019) 113130
Fig. 4. Gallery plot of GC-IMS spectra for volatiles from erythromycin broth at different cultivation times. The original gallery plot was divided into two plots (a) and (b)to fit in the display.
Fig. 5. Principal component analysis of erythromycin broth headspace volatiles measured using GC-IMS. (a) Score plot showing the correlation between broths with different fermentation hours, and (b) loading plot showing the correlation of the compounds detected from each broth.
X. Yang et al. / Environmental Pollution 255 (2019) 113130
7
References
Fig. 6. Variations of odorant concentrations of erythromycin broth with fermentation time measured using SPME-GC-MS.
the satisfactory deodorization of erythromycin fermentation exhaust gas. In addition, the formation of odorous compounds is also influenced by the culture conditions, such as medium composition, temperature, pH and oxygen content (Dionigi et al., 1996). The generation of MVOCs could vary significantly among different fermentation processes. Further investigation of odorants production during the fermentation process is required to increase understanding of odor emission patterns and the identification of major fermentation odorants. 4. Conclusions The fermentation exhaust gas of erythromycin had a strong musty or earthy odor. The olfactory intensity of the exhaust gas was high, and the odor concentrations varied from 4121 to 231739. Geosmin and 2-MIB were identified as the major odorants of erythromycin fermentation. Aldehydes, including hexanal, octanal, decanal, and benzaldehyde also contributed to the fermentation odor. A rapid profile analysis of the organic volatiles using GC-IMS spectra and principal component analysis (PCA) indicated that the emission pattern of volatiles is related to the fermentation stage. The production of major odorants was initiated in the very early stage of fermentation and became significant in the medium stage (30e70 h) of fermentation. The extremely low odor threshold values of the major odorants were responsible for the strong odor associated with erythromycin fermentation. More effective treatment processes are needed for the deodorization of erythromycin fermentation exhaust gas. Conflicts of interest No conflict of interest. Acknowledgments Financial support from the National Natural Science Foundation of China (51778604, 21577160) is gratefully acknowledged. The authors acknowledge assistance from Zhang Hao from G.A.S. GmbH for GC-IMS analysis, and Song RN and Lv MH for helping with samples collection and analysis. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.envpol.2019.113130.
Abalos, M., Prieto, X., Bayona, J.M., 2002. Determination of volatile alkyl sulfides in wastewater by headspace solid-phase microextraction followed by gas chromatography-mass spectrometry. J. Chromatogr. A 963, 249e257. Agus, E., Lim, M.H., Zhang, L.F., Sedlak, D.L., 2011. Odorous compounds in municipal wastewater effluent and potable water reuse systems. Environ. Sci. Technol. 45, 9347e9355. Andersen, K.B., Feilberg, A., Beukes, J.A., 2012. Use of non-thermal plasma and UVlight for removal of odour from sludge treatment. Water Sci. Technol. 66, 1656e1662. Anonymous, 1965. Fermentation processes - antibiotics. Biotechnol. Bioeng. 7, 29e&. Azaria, S., van Rijn, J., 2018. Off-flavor compounds in recirculating aquaculture systems (RAS): production and removal processes. Aquacult. Eng. 83, 57e64. Bentley, R., Meganathan, R., 1981. Geosmin and methylisoborneol biosynthesis in streptomycetes. Evidence for an isoprenoid pathway and its absence in nondifferentiating isolates. FEBS Lett. 125, 220e222. Bingley, G.D., Verran, J., Munro, L.J., Banks, C.E., 2012. Identification of microbial volatile organic compounds (MVOCs) emitted from fungal isolates found on cinematographic film. Anal Methods-UK 4, 1265e1271. Cai, L.S., Koziel, J.A., Lo, Y.C., Hoff, S.J., 2006. Characterization of volatile organic compounds and odorants associated with swine barn particulate matter using solid-phase microextraction and gas chromatography-mass spectrometryolfactometry. J. Chromatogr. A 1102, 60e72. Citron, C.A., Barra, L., Wink, J., Dickschat, J.S., 2015. Volatiles from nineteen recently genome sequenced actinomycetes. Org. Biomol. Chem. 13, 2673e2683. D'Imporzano, G., Crivelli, F., Adani, F., 2008. Biological compost stability influences odor molecules production measured by electronic nose during food-waste high-rate composting. Sci. Total Environ. 402, 278e284. Denawaka, C.J., Fowlis, I.A., Dean, J.R., 2014. Evaluation and application of static headspace-multicapillary column-gas chromatography-ion mobility spectrometry for complex sample analysis. J. Chromatogr. A 1338, 136e148. Dionigi, C.P., Ahten, T.S., Wartelle, L.H., 1996. Effects of several metals on spore, biomass, and geosmin production by Streptomyces tendae and Penicillium expansum. J. Ind. Microbiol. 17, 84e88. Domeno, C., Rodriguez-Lafuente, A., Martos, J.M., Bilbao, R., Nerin, C., 2010. VOC removal and deodorization of effluent gases from an industrial plant by photooxidation, chemical oxidation, and ozonization. Environ. Sci. Technol. 44, 2585e2591. Estrada, J.M., Lebrero, R., Quijano, G., Kraakman, N.J.R., Munoz, R., 2013. Strategiesforodoour control. In: Belgiorno, V., Naddeo, V., Zarra, T. (Eds.), Odour Impact Assessment Handbook. John Wiley & sons, Ltd., UK, pp. 85e124. Franke, W., Frechen, F.B., Giebel, S., 2009. H2S, VOC, TOC, electronic noses and odour concentration: use and comparison of different parameters for emission measurement on air treatment systems. Water Sci. Technol. 59, 1721e1726. Gallegos, J., Arce, C., Jordano, R., Arce, L., Medina, L.M., 2017. Target identification of volatile metabolites to allow the differentiation of lactic acid bacteria by gas chromatography-ion mobility spectrometry. Food Chem. 220, 362e370. Garcia-Alcega, S., Nasir, Z.A., Ferguson, R., Whitby, C., Dumbrell, A.J., Colbeck, I., Gomes, D., Tyrrel, S., Coulon, F., 2017. Fingerprinting outdoor air environment using microbial volatile organic compounds (MVOCs) - a review. Trac. Trends Anal. Chem. 86, 75e83. Gerber, N.N., Lechevalier, H.A., 1965. Geosmin an earthy smelling substance isolated from actinomycetes. Appl. Microbiol. 13, 4. Gerhardt, N., Birkenmeier, M., Sanders, D., Rohn, S., Weller, P., 2017. Resolutionoptimized headspace gas chromatography-ion mobility spectrometry (HS-GCIMS) for non-targeted olive oil profiling. Anal. Bioanal. Chem. 409, 3933e3942. , E., 2011. Odour-causing Godayol, A., Alonso, M., Besalú, E., Sanchez, J.M., Antico organic compounds in wastewater treatment plants: evaluation of headspace solid-phase microextraction as a concentration technique. J. Chromatogr. A 1218, 4863e4868. Higuchi, T., 2013. Determination of odour concentration by the triangular odour bag method. In: Belgiorno, V., Naddeo, V., Zarra, T. (Eds.), Odour Impact Assessment Handbook. John Wiley & sons, Ltd., UK, pp. 43e50. Higuchi, T., Shigeoka, K., 2018. Improvement of interlaboratory evaluation method of olfactometry in Japan. Arch. Environ. Prot. 44, 94e99. Jachymova, J., Votruba, J., Viden, I., Rezanka, T., 2002. Identification of Streptomyces odor spectrum. Folia Microbiol. 47, 37e41. Juenger, M., Vautz, W., Kuhns, M., Hofmann, L., Ulbricht, S., Baumbach, J.I., Quintel, M., Perl, T., 2012. Ion mobility spectrometry for microbial volatile organic compounds: a new identification tool for human pathogenic bacteria. Appl. Microbiol. Biotechnol. 93, 2603e2614. Kamarulzaman, N.H., Le-Minh, N., Fisher, R.M., Stuetz, R.M., 2019. Quantification of VOCs and the development of odour wheels for rubber processing. Sci. Total Environ. 657, 154e168. Kim, K.-H., Szulejko, J.E., Kumar, P., Kwon, E.E., Adelodun, A.A., Reddy, P.A.K., 2017. Air ionization as a control technology for off-gas emissions of volatile organic compounds. Environ. Pollut. 225, 729e743. Kwong, C.W., Chao, C.Y.H., Hui, K.S., Wan, M.P., 2008. Removal of VOCs from indoor environment by ozonation over different porous materials. Atmos. Environ. 42, 2300e2311. Li, W.F., Yang, W.H., Li, J.Y., 2018. Characterization and prediction of odours from municipal sewage treatment plant. Water Sci. Technol. 77, 762e769.
8
X. Yang et al. / Environmental Pollution 255 (2019) 113130
Mironov, V.A., Sergienko, O.V., Nastasyak, I.N., Danilenko, V.N., 2004. Biogenesis and regulation of biosynthesis of erythromycins in Saccharopolyspora erythraea. Appl. Biochem. Microþ 40, 531e541. Nagata, Y., Takeuchi, N., 1990. Measurement of odor threshold by triangle odor bag method. Bull. Jpn. Environ. Sanitation Center 17, 13. Ochiai, N., Takino, M., Daishima, S., Cardin, D.B., 2001. Analysis of volatile sulphur compounds in breath by gas chromatography-mass spectrometry using a threestage cryogenic trapping preconcentration system. J. Chromatogr. B 762, 67e75. Persson, P.E., 1980a. On the odor of 2-methylisoborneol. J. Agric. Food Chem. 28, 1344-1344. Persson, P.E., 1980b. Sensory properties and analysis of two muddy odour compounds, geosmin and 2-methylisoborneol, in water and fish. Water Res. 14, 1113e1118. Persson, P.E., 1983. Off-flavours in aquatic ecosystems eAn Introduction. Water Sci. Technol. 15, 1e11. Qi, Y.Q., Shen, L.M., Zhang, J.L., Yao, J., Lu, R., Miyakoshi, T., 2019. Species and release characteristics of VOCs in furniture coating process. Environ. Pollut. 245, 810e819. Rezanka, T., Prell, A., Sigler, K., 2008. Identification of odorous compounds from nine fermentor-cultivated Streptomyces strains. Folia Microbiol. 53, 315e318. Schiavon, M., Martini, L.M., Corra, C., Scapinello, M., Coller, G., Tosi, P., Ragazzi, M., 2017. Characterisation of volatile organic compounds (VOCs) released by the composting of different waste matrices. Environ. Pollut. 231, 845e853. Scholler, C.E.G., Gurtler, H., Pedersen, R., Molin, S., Wilkins, K., 2002. Volatile metabolites from actinomycetes. J. Agric. Food Chem. 50, 2615e2621. Shen, Y., Chen, T., Gao, D., Zheng, G., Liu, H., Yang, Q., 2012. Online monitoring of volatile organic compound production and emission during sewage sludge
composting. Bioresour. Technol. 123, 463e470. Smith, R.L., Pittenger, R.C., Bungay, H.R., 1962. Growth-biosynthesis relationships in erythromycin fermentation. Appl. Microbiol. 10, 4. Staunton, J., Wilkinson, B., 1997. Biosynthesis of erythromycin and rapamycin. Chem. Rev. 97, 2611e2630. Taylor, C., Lough, F., Stanforth, S.P., Schwalbe, E.C., Fowlis, I.A., Dean, J.R., 2017. Analysis of Listeria using exogenous volatile organic compound metabolites and their detection by static headspace-multi-capillary column-gas chromatography-ion mobility spectrometry (SHS-MCC-GC-IMS). Anal. Bioanal. Chem. 409, 4247e4256. Ueno, H., Amano, S., Merecka, B., Kosmider, J., 2009. Difference in the odor concentrations measured by the triangle odor bag method and dynamic olfactometry. Water Sci. Technol. 59, 1339e1342. Wang, X., Guan, J., Stuetz, R.M., 2012. Catalytic ozonation for odour removal of high temperature alumina refinery condensate. Water Sci. Technol. 66, 1781e1786. Wilkins, K., 1996. Volatile metabolites from actinomycetes. Chemosphere 32, 1427e1434. Wu, C., Liu, J., Zhao, P., Li, W., Yan, L., Piringer, M., Schauberger, G., 2017. Evaluation of the chemical composition and correlation between the calculated and measured odour concentration of odorous gases from a landfill in Beijing, China. Atmos. Environ. 164, 337e347. Young, W.F., Horth, H., Crane, R., Ogden, T., Arnott, M., 1996. Taste and odour threshold concentrations of potential potable water contaminants. Water Res. 30, 331e340. Zhou, X.Z., Zhang, Q., Huang, A.H., 2012. Evaluation of a plant material-based air purifier for removing H2S, NH3 and swine manure odour. Environ. Technol. 33, 2751e2756.