Measurement report: Hydrolyzed Amino acids in fine and coarse atmospheric aerosol in Nanchang, China: concentrations, compositions, sources and possible bacterial degradation state

s. Amino acids (AAs) are relevant for nitrogen cycles, climate change and public health. Their 15 size distribution may help to uncover the source, transformation and fate of protein in the atmosphere. This paper explores the use of compound-specific δ 15 N patterns of hydrolyzed amino acid (HAA), δ 15 N values of total hydrolyzed amino acid (δ 15 N THAA ), degradation index (DI), and the variance within trophic AAs (∑ V ) as markers to examine the sources and processing history of different sizes particle in the atmosphere. 2-weeks of daily aerosol samples from five sampling sites in the Nanchang area (Jiangxi 20 Province, China) and samples of main emission sources of AAs in aerosols (biomass burning, soil and plants) were collected (Zhu et al., 2020). Here, we measured the concentrations and δ 15 N values of each HAA in two size segregated aerosol particles (>2.5μm and PM2.5). Our results showed that the average concentrations of THAA in fine particles was nearly 6 times higher than that in coarse particles (p<0.01) and composition profiles of fine and coarse particles were quite different from each other. The δ 15 N 25 values of hydrolyzed glycine and THAA


Introduction
Recently, an increasing number of researchers highlight the importance of amino acids (AAs) in the atmosphere because AA is considered to be one of the most important organic nitrogen compounds in atmosphere (Zhang et al., 2002;Matos et al., 2016). Moreover, AAs are bioavailable and can be directly 40 utilized by plant and soil communities (Wedyan and Preston, 2008;Song et al., 2017). Its key role in atmosphere-biosphere nutrient cycling and global nitrogen cycle has aroused greatly concern (Samy et al., 2013;Zhang and Anastasio, 2003). Besides that, AAs and proteins are important constituents of allergenic bioaerosol (Miguel et al., 2009;Huffman et al., 2013). The distribution of AAs and proteins in different particle sizes will determine whether these compounds can reach the pulmonary alveoli and the 45 allergenicity of aerosols (Di Filippo et al., 2014). And the distribution of AAs associated with different particle sizes can help to trace the sources and transformation of atmospheric aerosols (Barbaro et al., 2019;Feltracco et al., 2019;Di Filippo et al., 2014).
The sources of atmospheric proteinaceous matter are very complex. Primary biological aerosol particles (e.g, plants, soil, pollen, bacteria, fungi, spores and deris of living things), biomass burning, and 50 agricultural activities are generally suggested to be the main contributing sources of atmospheric AAs (Matos et al., 2016;Mace et al., 2003). It is still unclear whether AAs fine and coarse particles influenced by different sources.
Compound-specific nitrogen isotope analysis of individual amino acids provide an opportunity to offer the key information on widely varied photochemical processes and origins of proteinaceous matter in the 55 atmosphere. Nitrogen sources information and any possible nitrogen isotopic fractionation caused by transformation processes could be hold by the δ 15 N-AA pattern (Mccarthy et al., 2007;Bol et al., 2002).
At the same time, the δ 15 N value of total hydrolysable AA (δ 15 NAvg-THAA), calculated as the average molarweighted δ 15 N value of individual AA, has been used as a proxy for total protein δ 15 N value . However, to our knowledge, no study has used the δ 15 N-AA pattern and δ 15 NAvg-THAA values 60 to identify the sources of AAs distributed in different particle sizes.
It is generally accepted that AAs in aerosols are mainly controlled by abiotic photochemical aging processes. On the contrary, the biological degradation of AAs in aerosols are neglected. This can be attributed to two factors. First, the sources and transformation pathways of protein matter and AAs in aerosols are highly complex (Wang et al., 2019;Zhu et al., 2020). Second, and the residence time of of atmospheric AAs is limited. For example, two studies on marine aerosols by Wedyan and Preston (2008) and Kuznetsova et al. (2005), and one study on precipitation by Yan et al. (2015). The degradation index (DI) proposed by Dauwe et al. (1998Dauwe et al. ( , 1999 has been wildly used to assess the degradation state of organic materials (OM) in terrestrial, aquatic, and marine environment (Dauwe and Middelburg, 1998;Wang et al., 2018;Dauwe et al., 1999). This value is based on the molar percentage (Mol%) of the amino acid pool and higher DI values denote a more "fresh" state of protein matter. However, DI values of AAs in aerosol particles and whether bacterial degradation plays a role in the levels and compositions of AAs in different particle sizes are still unknown.
A consensus has recently been reached on selective use of the 15 N depleted or enriched trophic AAs during bacterial heterotrophy processes can lead to large nitrogen isotopic fractionation in trophic AAs 80 (McCarthy et al., 2004). Thus, substantial δ 15 N pattern shifts of trophic AAs can index bacterial heterotrophy processes. ∑V, defined as the average deviation in the δ 15 N values of the Tr-AA, has therefore been established to track the degree of bacterial degradation of AAs in marine and terrestrial environment (Mccarthy et al., 2007;Philben et al., 2018;Yamaguchi et al., 2017).
In the present work, we sought to improve our understanding of AAs distributed in different particle sizes.

85
We measured the concentrations and δ 15 N values of each hydrolyzed amino acid in two size segregated aerosol particles (>2.5 μm and PM2.5) in aerosols collected in the Nanchang area (southeastern China).
Furthermore, δ 15 N values of Gly and THAA in fine and coarse particle were compared with those in main emission sources (biomass burning, soil and plant sources) to identify the potential sources of fine and coarse particles. In addition, the DI, ∑V values and δ 15 N values pattern of hydrolyzed AA in fine and 90 coarse particles were analyzed to explore the possible bacterial degradation of HAAs in fine and coarse particles. Aerosols from straw burning were sampled by pumping into a high-volume air sampler (KC-1000, Qingdao Laoshan Electronic Instrument Company, China) from the funnel on the combustion furnace during July 2017. The combustion furnace is a domestic furnace widely used by local residents.

Analyses of the concentration and δ 15 N value of individual hydrolyzed amino acid (HAA)
For hydrolyzed AA analysis, samples were prepared using a modified version of Wang et al. (2019) and Ren et al. (2018). One-sixteenth of each fine aerosol filter (~80 m 3 of air) or Two-seventh of each coarse aerosol filter (~366 m 3 of air) was broken into small pieces and placed in a glass hydrolysis tube. Prior to the hydrolysis, 25 μL of ascorbic acid at a concentration of 20 μg μL -1 (500 μg absolute) was added to each filter sample. Then, 10mL and 6M Hydrochloric acid (HCl) was used to convert all of the combined 125 AAs to free AAs. To avoid oxidation of AAs, the hydrolysis tube was flushed with nitrogen and tightly sealed before hydrolysis. The mixture was later placed in an oven at 110 °C for 24 h.
For plant and soil samples, approximately 30-40mg of plant or 500-600mg of soil were ground separately in liquid nitrogen into fine powders using a mortar and pestle. Then, well ground and homogenized soil and plant power were hydrolyzed in the same way as the aerosol samples.

130
After cooling to room temperature, the hydrolyzed solution was dried with a stream of nitrogen and HCl was removed. The dried solution was then redissolved in 0.1 M HCl and purified by a cation exchange column (Dowex 50W X 8H + , 200-400 mesh; Sigma-Aldrich, St Louis, MO, USA). Later, tert-Butyldimethylsilyl (tBDMS) derivatives of HAAs were prepared following the method described by our previous study Zhu et al. (2018).

135
The concentrations of HAAs were analyzed using a gas chromatograph-mass spectrometer (GC-MS).
The GC-MS instrument was composed of a Thermo Trace GC (Thermo Scientific, Bremen, Germany) connected into a Thermo ISQ QD single quadrupole MS. The single quadrupole MS was operated in electron impact ionization (70 eV electron energy) and full scan mode. The temperatures of the transfer line and ion source were 250°C and 200°C, respectively. More details on quality assurance and control 140 (recoveries, linearity, detection limits, quantitation limits, and corresponding effective limits in the aerosol samples of AAs), are provided in Zhu et al. (2020) δ 15 N values of AA-tert-butyl dimethylsilyl (tBDMS) derivatives were analyzed using a Thermo Trace GC (Thermo Scientific, Bremen, Germany) and a conflo IV interface (Thermo Scientific, Bremen, Germany) interfaced with a Thermo Delta V IRMS (Thermo Scientific, Bremen, Germany). The

145
analytical precision (SD, n=3) of δ 15 N was better than ±1.4‰. Moreover, AABA with known δ 15 N value (-8.17‰±0.03‰) was added in each sample to check the accuracy of the isotope measurements. The analytical run was accepted when the differences of δ 15 N values of AABA between GC-IRMS and EA-IRMS values were at most ±1.5‰. Each reported value is a mean of at least three δ 15 N determinations.
For more details of the analyses of HAA δ 15 N values refer to our previous publication (Zhu et al., 2018).

150
The concentrations and δ 15 N value of Cys, Trp, Asn and Gln in HAAs could not be determined using this method because, under strong acidic condition, Cys and Trp is destroyed, and Asn and Gln are converted to Asp and Glu, respectively. The concentration and δ 15 N value of hydrolysable Asp represents the sum of Asp and Asn; the concentration and δ 15 N value of hydrolysable Glu represents the sum of Glu and Gln.

DI index
Degradation process could significantly modify the mole composition of protein amino acids (Dauwe et al., 1999). Accordingly, a quantitative degradation index (DI) has been developed based on the mole composition of hydrolyzed amino acids pool. The degradation index (DI) was calculated using the formula Eq.
(1) originally proposed by Dauwe et al. (1999): where DI is the degradation index, Var is the mole% of the each individual HAA, Avgi, and SDi are the average mole% and standard deviation of each HAA in our data set, respectively, and PC1 is the loading of the amino acid i obtained from principal component analysis (Table S2).

δ 15 N values
The natural abundance of 15 N was calculated as δ 15 N values in per mil (‰), using atmospheric N2 as the 165 international standard: where R is the ratio of mass 29/mass 28.

170
Gly4, Phe, USGS40, USGS41a, and Val) with known δ 15 N values (−26.35 to +47.55‰) was prepared to assess the isotope measurement reproducibility and normalize the δ 15 N values of the amino acids in the samples (Zhu et al., 2018).

∑V parameter
The ∑V parameter is defined as the average absolute deviation in the δ 15 N values of the Trophic AA

175
(including: Ala, Asp, Glu, Ile, Leu, and Pro) (Mccarthy et al., 2007). This parameter has been used as a proxy for the degree of heterotrophic resynthesis and calculated by Eq. (3): where ꭓAA is defined as the deviation of the δ 15 N of each trophic amino acid from the δ 15 N of the mean of trophic amino acids (δ 15 N AA-average δ 15 N of Ala, Asp, Glu, Ile, Leu, and Pro), and n is the total 180 number of trophic amino acids used in the calculation.

δ 15 NTHAA values
The δ 15 N values of total hydrolysable amino acids (δ 15 NTHAA) is calculated as the mole percent weighted sum of the δ 15 N values of each individual HAA, following Eq. (4): Where mol%HAA is the mole contribution of each HAA and δ 15 NTHAA is the δ 15 N value of individual HAA.

190
We performed a Two-way ANOVA for the concentration of THAA, the DI index, δ 15 NTHAA values and ∑V values, testing the effect of aerosol sizes, location, and their interaction. Tukey's Honestly Significant Differences (Tukey-HSD) test was used to evaluate which combinations of location and aerosol size were significantly different. Two-way ANOVA was also conducted for DI values, examining the effect of aerosol sizes, coefficients (obtained by using first principal component score or previous reported 195 coefficients) and their interaction. The differences in δ 15 NGly values for fine particles between 5 sampling locations were examined using the one-way analysis of variance (ANOVA) procedure, and compared using the Tukey-HSD test.
The exponential regression was analyzed to evaluate changes in DI index as a function of the concentration of THAA.

200
To test for changes in the concentration of THAA, DI index and ∑V values following the rain events, a two-way ANOVA was performed, testing for effects of precipitation, aerosol sizes and their interactions.
Tukey-HSD test was conducted to compare the significant difference. Changes in mol% of each HAAs concentrations following precipitation were tested for significance by using ANOVA procedure followed by a Tukey-HSD test to compare significant differences. For all tests, statistically significant differences 205 were considered at p<0.05.

Concentrations and mol% composition profile of HAA in size-segregated aerosol
Fourteen hydrolyzed amino acids (Ala, Val, Leu, Ile, Pro, Gly, Ser, Thr, Phe, Asp, Glu, Lys, His and Tyr) were found in fine and coarse aerosol samples collected in Nanchang areas during spring 2019 ( Fig. 1).

210
The average concentrations of THAA in fine and coarse particles were 2542±1820 pmol m -3 and 434±722 pmol m -3 , respectively. The mean concentration of THAA for fine particles was nearly 6 times higher than that for coarse particles (p<0.01) (Fig. S1).
For fine particles, the average concentration of THAA in 5 sampling sites were significantly different However, for coarse particles, the difference in THAA concentrations between 5 sampling sites were not significant (p>0.05) (Fig. S1). The mean concentration of THAA in agricultural area, urban, forest, town 220 and suburban location was 540±821 pmol m -3 , 230±300 pmol m -3 , 654±1152 pmol m -3 , 437±583 pmol m -3 and 291±426 pmol m -3 , respectively. The highest concentration of atmospheric AAs at the agricultural area would be ascribed to the enhanced agricultural activities and natural source emission (e.g., pollen grain) in spring .
The composition profiles of HAA in fine and coarse particles during the whole campaign are shown in for an average of 25 ± 12%, 17 ± 8%, 12± 3% and 11 ± 6%, respectively, of the THAA pool.
For coarse particles, Pro were the most abundant THAA specie, with an average contribution of 63± 31% to the THAA pool. Leu, Ala and Val were the next most abundant species, each accounting for 7-9% of 230 the THAA pool, while other individual HAA was only minor component in coarse particles (Fig. 2). The HAA distribution among the different sampling locations for both fine and coarse particles appeared similar (Fig. 2).

Similar contribution sources of fine and coarse particles
The detailed size-resolved investigation for the sources of atmospheric AAs is limited. Filippo et al. production and biomass burning. These results could not provide conclusive evidence to define the origin of atmospheric AAs in the different particle sizes.
With the development of stable N isotope technology, δ 15 N values and δ 15 N pattern has become effective tools to trace the sources of nitrogen compounds. Our previous study found that the δ 15 N value of Gly in PM2.5 can be used to trace the potential emission sources for aerosol AAs because the N isotope 245 fractionation associated with Gly transformation in aerosol is relatively small (Zhu et al., 2020). To trace the sources of fine and coarse particles, we measured the nitrogen isotopic compositions of hydrolyzed Gly and THAA sampled from main emission sources in the study areas, including biomass burning, soil and local plants (Fig. 3). The average δ 15 N value for hydrolyzed Gly from the biomass burning, soil, and plant sources was +15.6 ± 4.3‰, +3.0 ± 4.4‰, and −11.9±1.4‰, respectively, and the mean δ 15 NTHAA 250 value was +15.8 ± 4.5‰, +5.5 ± 2.2‰, and −0.0 ± 1.8‰, respectively.
In this study, the δ 15 N values of hydrolyzed Gly in fine and coarse particles exhibited wide ranges: −1.0‰ to +20.3‰ and −0.8‰ to +15.7‰, which fall within the ranges of biomass burning, soil, and plants sources ( of THAA in potential emission sources of atmospheric protein AA, both fine (+0.7‰ to +13.3‰) and coarse particles (−2.3‰ to +10.0‰) had the δ 15 NTHAA value also typically in the range of these three main emission sources (Fig. 3). Therefore, it is likely that the main sources of atmospheric AAs for both fine and coarse particles were mainly biomass burning, soil, and plants.
However, there is no significant difference in the δ 15 NTHAA value between fine and coarse particles in each sampling sites (p>0.05) (Fig. 4c) and the average offset of δ 15 NTHAA value between fine and coarse particles was lower than 1.5 ± 1.7‰ at 5 sampling sites (Fig. 4a).Thus, it is suggested that the main sources of AAs in fine and coarse particles might be similar, all of which were influenced by biomass burning, soil, and plant sources.
In addition, as one of the main components of primary biological aerosol particles (PBAP), AAs are 275 proved to be ejected from ocean water by bursting bubbles Bigg, 2005a, 2005b;Bigg, 2007;Bigg and Leck, 2008). Marine source may also contribute to atmospheric AAs for both fine and coarse particles observed here. However, the sampling sites are located in an inland city.  Chikaraishi et al., 2009;Calleja et al., 2013), which was close to the range of the natural source including plant (range: -13.2‰ to -9.7‰) and soil (range: -1.6‰ to +7.4‰) sources.
Conclusively, the contribution from soil and plant sources mentioned in this study may include a very small amount of marine contribution.

290
The δ 15 NGly values of fine particles was significantly different at 5 sampling sites (p<0.05). The average δ 15 NGly value of fine particles in urban (average=14.3 ± 8.5‰) and town (average=9.4 ± 4.2‰) were more positive than that in suburban (average=6.7 ± 4.3‰), agricultural area (average= 6.9 ± 5.3‰) and forest site (average=6.5 ± 5.0‰) (Fig. 4b). The significantly higher δ 15 NGly values observed in the urban and town locations suggested an increased contribution from biomass burning sources to Gly in fine 295 particles at these two locations.

Different degradation state of AAs between fine and coarse aerosol particles
In this study, a huge difference was observed in the concentrations and mol% compositions of THAAs between fine and coarse particles ( Fig. 1 and 3). As we discussed above, the sources of AAs in fine and coarse particles are similar, therefore this lager difference may be attributed to protein matter in fine and coarse undergoing different degrees of oxidation, nitration and oligomerization in the atmosphere (Liu et 310 al., 2017;Wang et al., 2019;Song et al., 2017;Haan et al., 2009). Another possibility is that, biologically relevant degradation of AAs may contribute to this variation observed between fine and coarse particles.
To investigate whether AAs in fine and coarse particles may be degraded by bacteria to different degrees, degradation marker (DI) and bacterial heterotrophy indicators (δ 15 N-AA distribution and ∑V) were used.
Protein as major components in all source organisms are sensitive to all stages of degradation (Cowie 315 and Hedges 1992). Moreover, compared to the alteration of the degradation, the dissimilarity in amino acid composition of protein in the source organisms are minor (Dauwe and Middelburg 1998). Therefore, the degradation index (DI) is developed, which are based on protein amino acid composition and factor coefficients based on the first axis of the PCA analysis (equation 1). Since AAs concentrated in cell walls are preferential accumulated during decomposition, whereas amino acids that are concentrated in cell 320 plasma tend to be depleted during degradation (Dauwe et al., 1999), the compositional changes of amino acids associated with degradation can be traced by the DI value. The higher DI values indicate the protein is relatively "fresh" (Yan et al., 2015) and changes tracked by DI are proposed to be driven in large part by enrichment of AAs concentrated in cell wall (Mccarthy et al., 2007).
For calculation of DI values for fine and coarse particles, the first principal component score from 325 principal component analysis (PCA) was applied to our own data (including Ala, Gly, Val, Leu, Ile, Pro, Ser, Thr, Phe, Asp, Glu, Lys, His and Tyr), following the method described by Dauwe et al. (1999). The first principal component explained 38% of the variability, and the second principal component explained 21% (Table S2). Fig. 5a shows plots of the scores of the first and second principal components of fine and coarse particles in 5 sites. Components of fine and coarse particles could be roughly separated. The This is the first report of the DI values for aerosol particles. We compared DI values obtained by our calculating method with those calculated by using the coefficients given in previous references (Dauwe et al., 1999;Yamashita and Tanoue, 2003). There is no significant difference between the DI values calculated using the first principal component score and the DI values calculated using the coefficients given in the previous reference (Dauwe et al., 1999;Yamashita and Tanoue, 2003) (p>0.05) (Fig. S3), confirming our calculation method is reliable.
A plot of factor coefficients of each individual amino acid in the first and second principal components was examined to clarify the reasons for variation of the scores of fine and coarse particles (Fig. 5b).
Based on this cross plot, 14 HAA species were divided into four groups. In Fig. 5b to percentage of HAA species in Group 2 (e.g., Ala, Val, Leu and Ile) (Fig. S4), indicating the difference in composition profiles of HAA between fine and coarse particles may affected by the degradation process. Plots of DI as a function of THAA concentration in both fine and coarse particles showed an exponential relationship (y=1067.4e -1.0x ; r=0.6, p<0.01); that was, that at higher values of DI, concentrations of THAA were higher, and vice versa (Fig. S5). The coarse particles had significantly 360 lower THAA concentrations compared to fine particles (Fig. S1). Clearly, both composition profiles of HAA and concentrations of THAAs in aerosols may be related to degradation processes.
DI values from literature data, where possible and DI values for fine and coarse aerosol particles are shown in Fig. 6a and Fig. 7. Fine particles had significantly higher DI values than that of coarse particles (p<0.05) (Fig. 6a). The DI values for fine and coarse particles ranged from -0.3 to 1.4 (average=0.6±0.4) 365 and -1.8 to 1.4 (average=-0.6±1.0), respectively (Fig. 7). The DI values of fine particles were close to those of "fresh" material. For instance, source materials (e.g., plankton, bacteria and sediment trap material). On the contrary, the DI values of coarse particles were comparable to those of surface soil, POM in coastal sediments and DOM in coastal area, which were proved to be more degraded materials (Fig. 7). In marine environment, high DI values (>0.5) indicate the better preservation of more fresh 370 organic matter from marine primary production (Jiang et al., 2014). On the contrary, low DI values (<0.5) indicate the presence of relatively degraded organic matter (Burdige, 2007;Wang et al., 2018). In this study, the lower DI values observed in coarse particles, implying that AAs in coarse particles may undergo more degradation than fine particles. Our result is also comparable to that observed in precipitation at Uljin and Seoul (Yan et al., 2015). The DI values measured in coarse particles are closer 375 to those observed in Seoul, where is believed to have more advanced degradation than Uljin, further supporting the degradation degree of amino acids in coarse particles is higher than that in fine particles.
However, the differences in DI values were not significant among 5 sampling sites for both fine and coarse particles (p>0.05) (Fig. S6). For fine particles, the average DI values in agricultural area, urban, forest, town and suburban location was 0.6±0.4, 0.5±0.5, 0.7±0.3, 0.6±0.3 and 0.7±0.2, respectively. For 380 coarse particles, the mean DI values in agricultural area, urban, forest, town and suburban location was -0.5±0.9, -1.0±1.1, -0.8±1.1, -0.3±1.1 and -0.5±1.1, respectively. As we discussed above, the sources of atmospheric HAA were different among 5 sampling sites. This result suggested that the degradation process of amino acids in the atmosphere is less affected by their emission sources.

385
The existence of microorganisms in aerosol particles has been documented. However, whether bacterial degradation processes play a role in atmospheric protein degradation is not well understood. The negative correlation of the DI with the concentration of free γ-aminobutyric acid (GABA) and its mole percentage are depicted in Figure S7. Since bacteria are known to produce free GABA from their protein precursors Moreover, it is interesting to note that a substantial δ 15 N-AA shifts in trophic AA group was observed between fine and coarse particles among 5 sampling sites. Ala, Leu, Ile and Asp was 15 N-enriched in coarse particles compared to fine particles, whereas Pro in coarse particles was 15 N-depleted than those in fine particles (Fig. 8). Clearly, there is no uni-directional 15 N depletion or enrichment of Trophic-AA was observed between fine and coarse particle samples. The δ 15 N -AA distribution in the Trophic-AA group is more ''scattered'' in coarse particles than that in fine particles (Fig. 8). However, the difference

415
-AA shifts in only selected AA indicates the N of an assimilated AA has been replaced through a de novo heterotrophic AA resynthesis pathway with N isotope fractionation. Therefore, the substantial δ 15 N-AA shifts in trophic AA group could be observed when bacterial heterotrophy has occurred and those new resynthesized protein has become an important part of protein material measured (Mccarthy et al., 2007). Fogel and Tuross (1999) first observed that δ 15 N-AA patterns of degraded material was highly "scattered" and the N isotope fracionation between degraded material and fresh protein were up to 15‰. Moreover, obviously changes for the δ 15 N values of several AA were founded in high molecular weight dissolved organic carbon after bacterial reworking (Calleja et al., 2013). Similarly, the ''scattered'' characteristic of δ 15 N-AA distribution in Tr-AA group of coarse particles may be due to the nitrogen fractionation occurred in microbial consumers selectively using Trophic-AA.

425
∑V is defined as the average deviation in six Trophic-AA and has been proposed to reflect the extent of protein resynthesis during microbial degradation processes (Mccarthy et al., 2007). Fig. 9 shows the ∑V values measured in fine particles, coarse particles, and local natural sources, as well as ∑V values reported in previous references. ∑V values for main natural sources collected around the sampling sites were calculated. ∑V values for local plants (needles of Pinus massoniana (Lamb.) and leaves of Camphora 430 officinarum) ranged from 1.0‰ to 2.1‰, with a mean of 1.7±0.4‰ (Fig. 9). ∑V values in local soil (paddy soil, road soil and forest soil) ranged from 1.4‰ to 2.1‰, with a mean of 1.7±0.3‰. Overall, coarse particles had higher ∑V value (average = 3.6±1.5‰) than that of fine particles (p<0.05) (Fig. 9).
The mean ∑V value of fine particles in 5 sampling sites (average=2.4±1.1‰) was similar to or slightly higher than that of plants and soil collected around sampling sites, phytoplankton (1.0‰) and ∑V could reflect the increasing trend of "scatting" δ 15 N-Trophic AA pattern related to more intensive bacterial resynthesis (Batista et al., 2014;Calleja et al., 2013;Yamaguchi et al., 2017). In this study, the significant higher values of ∑V were measured in coarse particles than those in fine particles (p<0.05) (Fig. 6). Moreover, the mean ∑V value of fine particles was similar to or slightly higher than that 445 measured in "fresh" materials (Mccarthy et al., 2007;Philben et al., 2018;Batista et al., 2014), while ∑V values of coarse particles were equal to or even higher than those of more degraded materials (Fig. 9).
These corroborate that more bacterial heterotrophic resynthesis occurred in coarse particles compared to fine particles.
Despite the uncertainties surrounding oxidation, nitration and oligomerization of AAs in the atmosphere, 450 main observations remain that the difference in δ 15 N values of Source-AA (Gly, Ser, Phe and Lys) and total hydrolysable amino acids (δ 15 NTHAA) between coarse particles and fine particles was relatively small (Fig. 3). The average offset of δ 15 NTHAA value between fine and coarse particles was lower than 1.5‰ (Fig. 4a). These results appear to contrast with what one might expect for AAs in either sizes particles undergo particularly more photochemical transformation than the other. Therefore, significantly lower 455 DI values, ''scattered'' characteristic of δ 15 N distribution in Tr-AA and higher ∑V values observed in coarse particles in this study provide evidence that the difference in the THAA concentration and mol% composition distribution between fine and coarse particles may be related to AAs in coarse particles have stronger bacterial degradation state than those in fine particles.

Release of coarse "Fresh" bioparticles during the rainfall
A tight relationship between atmospheric bioaerosols and precipitation has been found by previous studies (Huffman et al., 2013;Yue et al., 2016). Since biological sources contain a large abundance of AAs (Ren et al., 2018), HAAs in aerosols can be used as tracer compounds to indicate the release of biological sources during precipitation. However, detailed size-resolved and time-resolved observation for the release of bioparticles initiated by precipitation are spare and the degradation state of different 465 sizes bioparticles has never been examined.
In this study, precipitation was observed to exert different impacts on the concentrations of the THAA in fine and coarse particles. The average concentration of THAA in fine particles on rainfall days (1948±1546 pmol m -3 ) was significantly lower than that measured on dry days (3137±1898 pmol m -3 ) (p < 0.05), whereas the average concentrations of THAA in coarse particles displayed no significant changes 470 between rainy and dry days (p>0.05) ( Fig. 1 and Fig. S1). For coarse particles, the average concentrations of THAA on rainy and dry days was 660±947 pmol m -3 and 212± 266 pmol m -3 , respectively. It is expected that the concentrations of individual AAs in aerosol were assumed to decrease on days which precipitation fell because of the high scavenging ratio of AAs in aerosol (Gorzelska and Galloway, 1990).
In this study, from rainy to dry days, the concentrations of THAA for fine particles decreased (p<0.05)

475
( Fig. S1), but the concentration of THAAs for coarse particles displayed not significant change (p>0.05) (Fig. S1). Similar variation trends of different size particles following the precipitation were also observed by Huffman et al.(2013). They also found the steep increase of coarse particles while low concentrations of fluorescent bioparticles and total aerosol particles were found in fine particles during the precipitation, suggesting the new released AAs during the precipitation are mainly distributed in 480 coarse particles.
It is worth noting that the influence of precipitation on the mole composition profile of HAA is different for the coarse and fine particles (Fig. 2). For fine particles, only the percentage of Pro significantly increased from 14±6% on dry days to 20±9% on rainfall days (p<0.05). There was no apparent trend in the percentage of other individual HAAs for fine particles following the precipitation.
For coarse aerosol, the percentage composition of HAA on dry days is quite different from that on rainy days for coarse particles (Fig. 2). From dry days to rainfall days, the percentage of Pro in coarse particles significantly decreased from 74±25% to 53±34% (p<0.05), meanwhile the percentage of Ala, Val, Leu, Ile and Glu in coarse particles significantly increased (p<0.05). These HAA species together accounted for 39% of the total THAA pool on dry days, while on rainfall days, this proportion was only 20%.

490
Besides that, compared to fine particles, the large variation in mole composition of THAA for coarse particles was observed on rainfall days (Fig. 2). From dry days to rainfall days, the percentage change of Pro for coarse particles (21%) was roughly 4 times greater than that for fine particles (6%). Similarly, from dry days to rainy days, the increase in the percentage of Ala, Val, Leu, Ile and Glu in coarse particles was significantly greater than that in fine particles. For example, following the precipitation, Val in coarse 495 particles increased by 4%, whereas Val in fine particles only increased by 0.3%. These large variations in the percentage of some HAA species (e.g., Pro, Ala, Val, Leu, Ile and Glu) were observed in coarse particles on rainy days, which imply the states of coarse particles measured on rainfall days were different from the ones measured on dry days (Fig. 2).
This conclusion also supported by the variation of DI and ∑V values for coarse particles on days which aerosol particles were also significantly affected by precipitation. From dry to rainy days, ∑V values of coarse aerosol particles decreased from 4.5±1.5‰ to 3.0±1.3‰ (p<0.05). In contrast, the average ∑V value of fine particles on dry and rainy days was identical (2.4±1.1‰). From dry to rainy days, DI values in coarse aerosol particles were significant increased (p<0.05) but the ∑V value was significantly decreased (p<0.05), suggesting more fresh AAs in coarse particles were released on days which 510 precipitation fell, whereas, on dry days AAs in coarse particles were more degraded.
Furthermore, we observed an obviously temporal variations of the concentration and mol% composition of HAA for coarse particles during the precipitation. The higher concentration of THAAs in coarse particles occurred on April 30, May 5, May 6 and May 13 when the daily precipitation amount was above 1mm and the hourly rainfall amount was above 0.2mm ( Fig. 1 and Table S4). Previous studies 515 demonstrated that droplets splashing on porous medium can deliver fresh biological aerosols in porous medium to the aerosol and this mechanism is closely related to the amounts and intensity of the rainfall events (Joung and Buie, 2015;Huffman et al., 2013;Yue et al., 2016). Thus, the temporal variation trend of HAA concentration for coarse particles in this study can attributed to the active release of biological aerosols caused by droplets and it highly depends on the amounts and intensity of the rainfall. Moreover, 520 the mol% composition of HAA in coarse particles measured on days with higher daily precipitation amount and hourly rainfall amount was significantly different from that observed on days with lower precipitation amount and intensity. Specifically, a steep decrease in the percentage of Pro and increase of other HAAs in coarse particles mainly occurred on days with daily precipitation amount above 1mm and hourly rainfall amount above 0.2mm, whereas the mol% composition of HAA on days with lower daily 525 and hourly precipitation amount were similar to those observed on dry days (Fig. 2). As we discussed above, AAs in coarse particles on dry days were more degraded. Therefore, we conclude that those "fresh" protein matters in coarse particles are likely prone to be released by droplets and amounts and intensity of the rainfall are the key factors controlling this mechanism.

530
This size distribution of AAs can help understand its transformation and fate in the atmosphere. Therefore, verification of the different types, concentrations, origin and atmospheric processes of AAs distribution along the different air particle sizes is important and meaningful.
This study presents the first isotopic evidence that the sources of AAs for fine and coarse aerosol particles may be similar, all of which were influenced by biomass burning, soil, and plant sources. It is therefore that the huge difference in the concentrations and mol% compositions of THAAs between fine and coarse particles observed in this study is closely relevant to the degradation processes of AAs in aerosols.
Although the oxidation, nitrification and oligomerization processes of protein substances in the atmosphere have been widely reported, these abiotic photochemical aging processes that occur between fine particles and coarse particles have not been compared. In this study, the difference in δ 15 N values of 540 Source-AA (Gly, Ser, Phe and Lys) and total hydrolysable amino acids (δ 15 NTHAA) between coarse particles and fine particles was relatively small. The average offset of δ 15 NTHAA value between fine and coarse particles was lower than 1.5‰. These results appear to contrast with what one might expect for AAs in either sizes particles undergo particularly more photochemical transformation than the other.
On the contrary, the degradation of atmospheric AAs in aerosols is rarely investigated. This is the first 545 report of using degradation marker (DI) to investigate the degradation state of aerosol particles. Both composition profiles of HAA and concentrations of THAAs in aerosols are showed to be closely related to DI. And fine particles had significantly higher DI values than that of coarse particles (p<0.05), suggesting the degradation degree of amino acids in coarse particles is higher than that in fine particles.
Combining new compound-specific nitrogen isotope tool (δ 15 N-HAA) and effective bacterial 550 heterotrophy indicator (∑V) , ''scattered'' characteristic of δ 15 N distribution in Tr-AA and higher ∑V values were observed in coarse particles in this study, which firstly provide evidence that the stronger degradation state the found in coarse particles are coupled with more bacterial heterotrophic resynthesis occurred in coarse particles.
This study suggests the potentially significant role of bacterial degradation processes in concentration 555 and composition of protein distribution in size-segregated aerosol particles. Since the degradation state of airborne protein distribution along size-segregated particles is closely linked to its biological availability, ecological processes and plant nutrition after deposition, further studies of quantitative assessment of this biological related process in aerosols should be conducted.

560
Author contributions. Ren-Guo, Zhu., Zequn Wen and Yuwen Zhu designed the experiments, performed analyses, and analyzed the data. Hua-Yun Xiao were the principal investigators of the project that supported this work. All the authors have helped in the discussion of the results and collaborated in writing this article.    Figure 6. DI values (a) and (b) ∑V for fine (red box) and coarse (blue box) particles. The box encloses 50% of the data, the whisker is standard deviation of the data, the horizontal bar is the median, solid circles are outliers. The differences in means were statistically significant (two-way ANOVA, p < 0.05). Different uppercase letters denote means found to be statistically different (Tukey-HSD test) between fine and coarse particles. Different lower case letters denote means found to be statistically different (Tukey-HSD test) between rainy and dry days.