Introduction
Lipid biomarkers from terrigenous plants, algae, fungi and soil
microorganisms have been reported extensively in aerosols (Conte and Weber,
2002; Gagosian et al., 1987, 1981; Kawamura, 1995; Kawamura et al., 2003;
Simoneit, 1977; Simoneit et al., 2004), sediments (Kawamura, 1995; Kawamura
and Ishiwatari, 1984; Kawamura et al., 1987; Zhang et al., 2014), ice cores
(Sankelo et al., 2013) and rain/snow (Kawamura and Kaplan, 1986;
Satsumabayashi et al., 2001; Yamamoto et al., 2011). These studies have
utilized fatty acids as a proxy to assess the terrigenous contribution of
higher plant waxes to various environmental samples owing to their abundant
presence in biopolymers of plants and microorganisms. Similarly, certain
hydroxy fatty acids (e.g., C10–C18 β-hydroxy FAs) have been
proposed as a tracer to understand the airborne bacterial transport (Tyagi et
al., 2015).
Among the airborne soil microbes, the Gram-negative bacterium (GNB) is one of
most extensively studied bacteria and is documented in aerosols, snow and
rain samples (Morris et al., 2011). Owing to considerable ground-based
emissions of GNB and its ability to act as cloud condensation nuclei (CCN),
these bacteria, which are plant pathogens, can influence the regional as well
as global climate through cloud–aerosol interactions (Morris et al., 2011,
and references therein). In particular, GNB contains β-hydroxy FAs
(C10–C18) in lipopolysaccharides (LPS) as constituents of the
outer cell membrane (Westphal, 1975). Moreover, the environmental toxic
effects of GNB are, in part, due to the presence of β-hydroxy FAs
present in LPS (endotoxin) (Larsson, 1994; Saraf et al., 1997; Spaan et al.,
2008).
Apart from β-hydroxy FAs, other positional isomers such as α-,
ω- and (ω-1)-hydroxy FAs have also been documented in various
environmental archives viz. aerosols (Kawamura, 1995; Tyagi et al., 2015) and
sediments (Kawamura, 1995; Wakeham et al., 2003; Zhang et al., 2014). Short
chain α-hydroxy FAs (C12–C18) are the constituent
biopolymers of fungi (Zelles, 1997), soil bacteria (Steinberger et al., 1999;
Zelles and Bai, 1994) and protozoa (Ratledge and Wilkinson, 1988). In
contrast, long chain α-hydroxy FAs (C16–C26) are abundant
in plants, microalgae and cyanobacteria (Matsumoto and Nagashima, 1984).
Likewise, ω- and (ω-1)-hydroxy FAs are highly cross-linked
constituents of the cell walls of algae (Blokker et al., 1999) and plant
seeds, suberin and cutin in terrestrial higher plants (Molina et al., 2006).
In addition, ω- and (ω-1)-hydroxy FAs are the intermediates
in the oxidation of monocarboxylic acids to dicarboxylic acids in sediments
and marine aerosols (Kawamura, 1995; Kawamura and Gagosian, 1990).
Furthermore, specificity of hydroxylation in FAs depends on the types of
microorganisms involved (Wakeham, 1999).
These tracer compounds in snow samples may be important to better understand
the contribution of plant and pathogenic bacteria to regional
vs. long-range
atmospheric transport (Hines et al., 2003; Lee et al., 2004, 2007; Tyagi et
al., 2015) as their presence in the atmosphere can affect the CCN and ice
nuclei activity (Morris et al., 2011). To the best of our knowledge, our
study is the first to report α-, β- and ω-hydroxy FAs
in snow samples. Snow efficiently scavenges airborne particles including soil
microbes and higher plant metabolites in the free boundary layer of
troposphere. Since hydroxy FAs from GNB and plants are inert in nature, they
do not undergo chemical modification during snow accumulation. Therefore,
hydroxy FAs in fresh snow can be used as a tracer to assess the sources and
transport pathways of microorganisms and plant metabolites.
In this study, we determined hydroxy FAs in fresh snow samples collected
from Sapporo, Japan, to evaluate the qualitative contribution from GNB and
higher plant metabolites. Our results support the hypothesis that these
hydroxy FAs are important tracers to better understand the contribution of
microorganisms to the organic matter in snow. More importantly, we also
discuss the possible transformations of these chemical markers during
long-range atmospheric transport.
Experimental methods
Site description and sample collection
Sapporo (43.07∘ N, 141.36∘ E) is the capital of Hokkaido,
whose population is 1.9 million (June 2013). Sapporo receives cold and dry
air masses with heavy snowfall during the Asian winter monsoon. The average
temperature of Sapporo in winter goes up to ∼ 2 ∘C (Yamamoto
et al., 2011). Snow cover over the ground and fallen leaves of deciduous
plants suppress the suspension of soil particles during winter, whereas the
emissions of plant biomarkers from local vegetation are minimal. During the
winter season, the Asian monsoon affects the regional climate, air quality
and human health in Japan, delivering anthropogenic aerosols and dust from
China and Siberia (Yamamoto et al., 2011). Several studies have examined the
chemical and isotopic composition of ambient aerosols in various types of air
masses in Sapporo (Aggarwal and Kawamura, 2008; Pavuluri et al., 2013;
Yamamoto et al., 2011) to better understand the impacts of anthropogenic and
biogenic contributions from Siberia, northern China and the surrounding
oceans. However, no study that focuses on the transport of microorganisms
using organic markers is available from Sapporo.
In this study, eleven fresh snow samples were collected from the rooftop of
the Institute of Low Temperature Science (ILTS) building, Hokkaido University
in Sapporo, during intensive snowfall periods (January–March) in 2010 and
2011. The detailed description about snow collection and analytical protocol
of lipid fraction analyses is similar to that described in Yamamoto et
al. (2011). To avoid the contribution of any possible impurities from the dry
deposition of aerosols, 1–2 cm of surface snow cover were removed prior to
sample collection. Thereafter, snow samples were collected in a cleaned glass
jar (8 L) by using a stainless steel shovel. In each glass jar, mercuric
chloride (HgCl2) was added before sampling to prevent microbial
activity. Soon after the collection, glass jars were tightened with a
Teflon-lined screw cap and stored at -20 ∘C until analysis.
Identification and quantification of hydroxy FAs
The analytical protocol used for assessing the atmospheric abundances of
hydroxy FAs is described in Yamamoto et al. (2011). In brief, melted snow
samples (0.5–1 L) were saponified with 1.0 M KOH in methanol at
80 ∘C for 2 h. After saponification, the neutral fraction was
separated and the remaining solution was acidified with 6 M HCl to form free
carboxylic acids. Furthermore, these acids were derivatized with
BF3 / methanol to form their methyl esters. The hydroxy acid methyl
esters were isolated on a silica gel column by eluting with methylene
chloride / methanol (95 : 5). The hydroxy FA methyl esters were, then,
derivatized to their trimethylsilyl (TMS) ethers with
N,O-bis-(trimethylsilyl) trifluoroacetamide (BSTFA)
(SUPELCO™ Analytical) at 70 ∘C for
1 h. After the reaction, 50 µL of n-hexane solution containing
1.43 ng µL-1 of internal standard (C13
n-alkane/tridecane, Wako) were added to dilute the derivatives prior to GC/MS
injection (Hewlett-Packard Model 6890 GC coupled to the Hewlett-Packard Model
5973 mass-selective detector, MSD). The GC was installed with a
split/splitless injector and DB-5MS fused silica capillary column.
For the quantification of hydroxy FAs, the GC oven temperature was programmed
from 50 ∘C (2 min) to 305 ∘C (15 min) at
5 ∘C min-1. Data were acquired and processed with the
Chemstation software. Structural identification and comparison of retention
time of hydroxy FAs were performed using authentic TMS derivatives of
n-C12 and n-C16 α-hydroxy FAs, n-C12, n-C14,
n-C15, and n-C16 β-hydroxy FAs and n-C16, n-C20
and n-C22 ω-hydroxy FAs. The recoveries of authentic fatty acid
standards were better than 92 ± 4 % with analytical error (average
4.1 %) for acidic compounds (Yamamoto et al., 2011). Laboratory blanks
showed no contamination of any target compounds. The results of n-alkanes,
n-alkanols and n-alkanoic acids (terrestrial biomarkers) in snow samples are
reported in Yamamoto et al. (2011), which revealed a long-range atmospheric
transport of terrestrial organic materials from northeastern Asia to northern
Japan by the Asian winter monsoon.
Estimation of endotoxin levels and mass loading of GNB
Since the endotoxins from GNB contain β-hydroxy FAs from C10 to
C18, previous studies attempted to quantify atmospheric abundances of
endotoxins using the concentrations of ambient hydroxy FAs measured (Lee et
al., 2004; Rietschel et al., 1984; Wilkinson, 1988). According to these
studies, concentrations of endotoxins in snow samples can be estimated from
the mathematical expression as below.
Endotoxins(LPS,ngkg-1of meltwater)=[(Σβ-hydroxy FAs from C10toC18;nmol kg-1)×8000]/4
In the above formula, the average molecular weight of endotoxin corresponds
to 8000 as reported by Mielniczuk et al. (1993). β-Hydroxy FAs in the
mathematical expression are the total (LPS-bound + free) hydroxy FAs for
the carbon numbers from C10 to C18. We also estimated the mass
loading of airborne GNB using the approach initially suggested by Balkwill et
al. (1988) and later on by Lee et al. (2004), in which they used the chemical
marker to bacterial mass conversion factor of 15 nmol of β-hydroxy
FAs (C10–C18) per mg dry cell weight. Therefore, we have converted
the sum of mass concentrations of β-hydroxy FAs from C10 to
C18 (in nmol kg-1) into the equivalent dry cell weight of GNB
(i.e., in mg kg-1 of meltwater) by normalizing with 15.
Results and discussion
Air mass backward-trajectory analysis
The air mass back trajectories (AMBTs) provide a means to qualitatively
assess the source regions of airborne pollutants over a receptor site. For
this study, we have computed 7-day isentropic AMBTs using the Hybrid Single
Particle Lagrangian Integrated Trajectory (HYSPLIT) model (Draxler and Rolph,
2013, and references therein). The meteorological parameters (GDAS data sets)
from the NOAA Air Resources Laboratory were used as an input for the HYSPLIT
model. Figure 1 shows the AMBT cluster at an arrival height of 500 m over
Sapporo during sampling days of the winter of 2010 and 2011. In almost all
snow-sampling periods in Sapporo, the AMBTs show plausible influence of air
masses from Russia and Siberia via the long-range atmospheric transport.
Air mass back-trajectory cluster at an arrival height of
500 m a.g.l. (above ground level) for the sampling days in
(a) winter 2010 and (b) winter 2011.
Mass concentrations (in ng kg-1) of α-, β- and
ω-hydroxy fatty acids (FAs) measured in snow samples (N=5)
collected from Sapporo during the winter of 2010.
2010
C-number
α-Hydroxy FAs
β-Hydroxy FAs
ω-Hydroxy FAs
Range
Mean ± S.E.
Median
Range
Mean ± S.E.
Median
Range
Mean ± S.E.
Median
C9
b.d.–7.1
2.4 ± 1.3
1.7
0.5–2.7
1.8 ± 0.47
2
b.d.–1.7
0.97 ± 0.4
1.4
C10
b.d.–37.3
14.6 ± 7.6
10.9
1.7–6.5
4.6 ± 1.2
5.1
b.d.–5.1
1.7 ± 1.1
0
C11
b.d.–35.1
21 ± 6.5
21.1
3.4–7.9
6.1 ± 0.8
6.2
b.d.–6.4
2.2 ± 1.4
0
C12
b.d.–46.7
25.3 ± 7.8
22.6
8–10.1
9.2 ± 0.4
9.8
b.d.–95.6
47.2 ± 17.8
32.7
C13
b.d.–45.2
20 ± 7.3
18
3.5–11.9
7.1 ± 1.8
6
b.d.–5.1
3.7 ± 0.9
4.4
C14
b.d.–53.4
27.1 ± 8.5
27.6
16.6–40.9
23.5 ± 4.4
19.6
b.d.–196.7
101 ± 34.7
79.8
C15
b.d.–44
18.6 ± 7.2
16.4
2.9–10.8
6.8 ± 1.4
6.7
b.d.–17
9.6 ± 3.1
12.8
C16
b.d.–139
89.2 ± 23.6
97.8
21.7–79.4
45.1 ± 9.4
4.4
2.3–754.1
296 ± 129
256.3
C17
b.d.–26.5
12.4 ± 4.4
10
3.1–10.7
7.5 ± 1.3
8.4
b.d.–12.6
7.1 ± 2
8.1
C18
b.d.–44.7
26.2 ± 8.1
26.3
23.4–52.3
33.5 ± 6.6
29.1
b.d.–43.9
21.2 ± 6.9
21
C19
b.d.–20.1
11.5 ± 3.4
11.5
5.3–21.7
10.4 ± 3.8
7.3
b.d.–12.2
5.5 ± 2
5.7
C20
b.d.–46.6
25 ± 7.8
21.5
14.4–120
48.3 ± 25
29.2
0.2–45.6
17.2 ± 7.6
13.5
C21
b.d.–21.1
12.1 ± 3.7
11.2
5.6–28.8
14.8 ± 5.4
13
b.d.–8.7
3.6 ± 1.4
3
C22
b.d.–73.7
40.8 ± 13.1
37.7
11.2–30.4
19.5 ± 4.1
18.2
b.d.–318
96.4 ± 56.5
50.7
C23
b.d.–32.8
18.5 ± 5.8
18.3
2.8–33.9
13.2 ± 7.1
8.1
b.d.–9.2
3.8 ± 1.6
3.6
C24
b.d.–145
64 ± 25
56.8
6.2–29
15 ± 5.1
12.3
b.d.–72.4
24.1 ± 12.7
13
C25
b.d.–39.1
18.4 ± 6.7
15.4
1.4–17.4
7.7 ± 3.4
5.9
b.d.–2.6
1.02 ± 0.5
1.2
C26
b.d.–49.3
18.6 ± 9
15.8
b.d.–18
7.5 ± 3.8
6
b.d.–3.2
0.6 ± 0.6
0
C27
b.d.–14.4
4.4 ± 2.8
1.1
b.d.–2.7
0.7 ± 0.7
0
b.d.–0.2
0.03 ± 0.03
0
C28
b.d.–10.9
4 ± 2.5
0
b.d.–1.6
0.3 ± 0.3
0
C29
b.d.–0.54
0.1 ± 0.1
0
C30
b.d.–0.32
0.06 ± 0.06
0
Total
432–774
593 ± 88
582
70–379
247 ± 52
252
2–1411
643 ± 228
530
Note: b.d. = below detection limit ≤ 0.02 ng kg-1; S.E. (standard error) = σ/N1/2, where
σ refers to the standard deviation of total samples (N).
Mass concentrations (in ng kg-1) of α-, β- and
ω-hydroxy fatty acids (FAs) measured in snow samples (N=6)
collected from Sapporo during the winter of 2011.
2011
C-number
α-Hydroxy FAs
β-Hydroxy FAs
ω-Hydroxy FAs
Range
Mean ± S.E.
Median
Range
Mean ± S.E.
Median
Range
Mean ± S.E.
median
C9
b.d.–27.2
14.2 ± 5.7
13.8
1–8.5
5.1 ± 1.3
6
b.d.–16.4
11.0 ± 2.6
12.9
C10
b.d.–65.4
30.9 ± 11.2
33.3
1.7–12.7
8.1 ± 1.8
8.8
b.d.–4.7
0.8 ± 0.8
0
C11
19.8–66.6
34.2 ± 8.5
28.5
1.7–13.3
9.2 ± 1.9
10.1
b.d.–4.7
0.8 ± 0.8
0
C12
20.7–60.4
36.5 ± 6.6
32.9
1.3–15.3
8.7 ± 2.2
8.8
b.d.–13.4
4.3 ± 2.7
0
C13
b.d.–49.2
21.5 ± 8.2
21.8
4.5–15.8
9.1 ± 2.1
8.6
b.d.–7.3
2.1 ± 1.2
1
C14
7.5–55.3
28.6 ± 7.7
28.4
4.5–25.5
13.7 ± 4
16.6
b.d.–61.5
17.7 ± 9.3
9.1
C15
b.d.–77.6
29.2 ± 13.1
23.3
1.9–11.1
6.3 ± 1.8
7.7
b.d.–12.1
4.0 ± 2.2
3.9
C16
14.3–186
94.0 ± 29.3
92.5
2.8–55.8
30.5 ± 10.2
32.8
b.d.–159
42.9 ± 24.7
19.4
C17
2.8–29.3
15.3 ± 4.3
14.5
1.6–12.2
7.7 ± 2.2
9
b.d.–8.2
1.9 ± 1.3
0.3
C18
8.0–55.8
31.3 ± 8.2
29.9
0.6–31.4
14.4 ± 5.3
13.6
b.d.–18.2
5.8 ± 2.8
3.9
C19
b.d.–22.4
6.2 ± 4.4
0
1.9–10.9
6.5 ± 1.5
7.1
b.d.–6.5
1.5 ± 1.0
0.5
C20
11.5–97.9
53.5 ± 18.6
47.3
1.2–43.4
23 ± 8.6
27.4
b.d.–10.5
3.3 ± 1.5
2.3
C21
b.d.–95.2
29.1 ± 17.2
13
1.0–16.6
8.8 ± 3.2
8.8
b.d.–3.4
1.0 ± 0.5
0.6
C22
13.4–109
60.8 ± 19.9
56.1
1.6–27.2
19.8 ± 6.1
25.2
b.d.–48.1
13.7 ± 7.4
8.2
C23
8.1–58.1
32.2 ± 10.1
26.3
5.7–11.6
9.1 ± 1.7
10
b.d.–6.8
1.2 ± 1.1
0
C24
12.3–92.2
74.9 ± 34
34
19.1–24.3
22.2 ± 1.6
23.1
b.d.–38
9.1 ± 6.0
3.2
C25
2.6–51.3
18.4 ± 8.9
9.8
3.3–11.1
8.5 ± 2.6
11.1
b.d.–3.7
1.0 ± 0.6
0
C26
2.6–52.0
24.2 ± 9
23.5
b.d.–15.9
6.4 ± 3.1
4
b.d.–10
2.2 ± 1.6
0.1
C27
b.d.–5.6
2 ± 1.3
0
b.d.–9.2
3.3 ± 1.6
2.1
C28
b.d.–4.8
1.4 ± 0.9
0
b.d.–10.6
4.3 ± 2.1
2.3
b.d.–1.4
0.2 ± 0.2
0
C29
b.d.–3.35
0.7 ± 0.67
0
C30
b.d.–0.60
0.12 ± 0.12
0
Total
169–1279
639 ± 187
651
6–354
179 ± 64
170
27–422
149 ± 73
102
Note: b.d. = below detection limit ≤ 0.06 ng kg-1. S.E. (standard error) = σ/N1/2, where
σ refers to the standard deviation of total samples (N).
Concentrations of hydroxy fatty acids
Homologous series of α-,
β- and ω-hydroxy FAs were detected in fresh snow samples
collected from Sapporo. Their mass concentrations are summarized in Tables 1
and 2 for the winters of 2010 and 2011, respectively. Based on 2-year
seasonal data on hydroxy FAs, we found that concentrations of α-hydroxy FAs are significantly higher than β- and ω-hydroxy
FAs. The predominance of α-hydroxy FAs can be explained by the
α-oxidation pathway of FAs, which generally occurs in plants, animals
and bacteria (Cranwell, 1981, and references therein) whereas β- and
ω-oxidation is specific to bacteria (Lehninger, 1975). α-Hydroxy FAs, in particular high molecular weight ones, come from the
epicuticular waxes of higher plants as well as from algae. However, we also
found higher abundance of α-hydroxy FAs in the biomass burning
aerosols collected over Mt. Tai, China (Tyagi et al., 2015), possibly due to
photochemical oxidation of higher molecular weight fatty acids. Such a
possibility of in situ formation of α-hydroxy FAs has also been
reported in the hydrolysis products of leaf waxes and wood, and in microalgae
and sea grasses (Feng et al., 2015). Furthermore, microbial oxidation could
also be a possible source of α-hydroxy FAs (Eglinton et al., 1968) in
the snow samples studied. Hence, we suggest that α-hydroxy FAs cannot
be employed as the tracers of plant waxes only, as they can come from
microbial/photochemical oxidation of higher molecular weight fatty acids
during long-range atmospheric transport.
A characteristic feature of our data is the predominance of C16 hydroxy
FAs in all the types of hydroxy FAs measured. However, significant shifts
were observed in the carbon numbers of the second-most abundant β-hydroxy FAs (mostly C-number > 16) and ω-hydroxy FAs (i.e.,
C-number < 16; see Tables 1 and 2). A likely explanation for this
observation is that β-hydroxy FAs above C16 were formed by β-oxidation of long chain FAs, which are more common in microorganisms as
discussed previously. In contrast, ω-hydroxy FAs below C16 are
present in plants and microbes (Cardoso and Eglinton, 1983), in which ω-oxidation of fatty acids is the secondary choice for microbial oxidation.
Molecular distributions
Figure 2 presents molecular distributions of α- (C9 to
C30), β- and ω-hydroxy FAs (C9 to C28) in snow
samples from Sapporo during the winter of 2010 and 2011. Even carbon number
predominance is noteworthy for α-, β- and ω-hydroxy
FAs. α-Hydroxy FAs show molecular distributions with the order
C16 > C24 > C22 in both years (see Fig. 2a). Likewise,
β-hydroxy FAs show the predominance of C16 followed by C18
or C20 and then by C14 in both winters. However, we found the
predominance of C20 β-hydroxy FAs over C16 in one snow
sample during 2010. Similarly, ω-hydroxy FAs showed dominance of
C16 followed by the others as
C14 > C12 ∼ C22 ∼ C24 during snowfall
in both the years.
Molecular distributions of (a) α-hydroxy fatty
acids (FAs) (C9–C30), (b) β-hydroxy FAs
(C9–C28) and (c) ω-hydroxy FAs
(C9–C28) in the snow samples collected from Sapporo during the
winter of 2010 and 2011.
Table S1 in the Supplement describes the statistically significant
differences in the ratios of even to odd carbon numbers for α-,
β-, and ω-hydroxy FAs in snow samples based on a two-tailed
unpaired t test. No significant differences were observed between 2010 and
2011 for the ratios of even to odd carbon number α-hydroxy FAs. In
contrast, the difference is statistically significant between 2010 and 2011
for β- and ω-hydroxy FAs. In fact, the difference is much
larger for ω-hydroxy FAs than that for β isomers. In the 2010
winter, AMBTs show atmospheric transport from the continents at 500, 1000 and
1500 m above ground; however, at the same heights in the 2011 winter, the
air masses came from the oceans during one sample collection. Higher plants
in the continents contribute to higher abundances of hydroxy FAs than the
oceans, and thus explain higher abundances of β- and ω-hydroxy
FAs in 2010 than in 2011. On average, even carbon numbered α-, β- and ω-hydroxy FAs in their total mass concentrations account for
∼ 69, 68 and 84 %, respectively. The even carbon number
predominance is also found in recent marine and lacustrine sediments (Cardoso
and Eglinton, 1983; Goossens et al., 1986; Kawamura, 1995; Zhang et al.,
2014).
Similar to our study, Volkman et al. (1980) documented the bimodal
distribution of α-hydroxy FAs with peaks at C16 and C24 in
the intertidal sediments from Victoria, Australia, and attributed their
contribution from sea grass (i.e., Zostera muelleri) detritus owing
to similar distribution patterns. However, it is noteworthy that our AMBTs
show a continental origin rather than the oceanic origin. Therefore, it is
possible that waxes emitted from continental grasses via wind abrasion can be
transported to Sapporo through the atmosphere. We speculate that α-hydroxy FAs (C16–C28) in Sapporo snow can be used as a tracer of
plant waxes. Likewise, higher plant-derived cutin and suberin have been
suggested as a significant source of C16 to C22 α-, β- and ω-hydroxy FAs (Cardoso and Eglinton, 1983). In a similar way,
it has been proposed that hydroxy FAs (C20–C30) are principally
derived from terrestrial higher plants (Kawamura and Ishiwatari, 1984).
Therefore, α-, β- and ω-hydroxy FAs
(C16–C22) in snow samples can be related to their sources from
terrestrial higher plants through long-range atmospheric transport.
Previous studies documented ubiquitous occurrence of these hydroxy FAs in
soil microbes such as yeast and fungi (Van Dyk et al., 1994, and references
therein) and in the LPS of GNB (Lee et al., 2007). In this regard, prior
studies focussing on β-hydroxy FAs with the predominance of C16
and C18 suggested the contributions from yeast and fungi (Stodola, 1967;
Van Dyk et al., 1994, and references therein). Molecular distributions of
β-hydroxy FAs show a predominance of C16 followed by C18 or
C20 (see Fig. 2b), suggesting that they have been derived from soil
microbes. Likewise, FAs < C20 are derived from marine phytoplankton
(Kawamura, 1995, and references therein). β-Hydroxy FAs
(C10–C18) have been proposed as a biomarker for soil microbes as
they are the constituents of LPS of GNB (Lee et al., 2004; Szponar et al.,
2002). Hence, it is likely that β-hydroxy FAs in snow samples may have
been significantly influenced by GNB and terrestrial higher plant
metabolites.
Figure 3 depicts bar graphs showing the relative abundances of α-,
β- and ω-hydroxy FAs in the snow samples from Sapporo during
winter. We found that the proportions of two classified groups (low molecular
weight C9–C19 and high molecular weight C20–C30 or
C20–C28) of α-, β- and ω-hydroxy FAs are
very similar between 2010 and 2011 (Fig. 3). This observation is perhaps
related to their common sources/transport pathways of α-, β-
and ω-hydroxy FAs over Sapporo. This inference is further supported
by the AMBTs computed at arrival heights of 500, 1000 and 1500 m (see
Figs. 1 and S1 in the Supplement), indicating similar air mass transport
pathways from Russia and Siberia.
Endotoxin potency of GNB impact via aeolian transport
Endotoxin in GNB determines their viability and potentially causes
pathological effects on mammals (Lüderitz et al., 1981; Westphal, 1975).
In particular, GNB contain LPS in their outer membrane. When bacteria
multiply, die and lyse, LPS are released from the surface as a potential
bacterial toxin, and are therefore called an endotoxin (Westphal, 1975). In
addition to intact bacterial cells, this endotoxin can trigger and cause
allergies, respiratory problems and infections. Researchers have used LPS
concentrations as a measure of GNB, primarily by the limulus amebocyte lysate
(LAL) assay that has limited specificity (Saraf et al., 1997). The β-hydroxy FAs, markers for endotoxin/LPS, were assayed in various
environmental samples such as dust (Andersson et al., 1999; Hines et al.,
2000), aerosols (Lee et al., 2004, 2007; Walters et al., 1994), soils
(Keinänen et al., 2003), sewage (Spaan et al., 2008) and marine dissolved
organic matter (Wakeham, 1999).
Bar graph, showing the relative abundances of low molecular weight
(C9–C19) and high molecular weight fatty acids (C20–C30
for α-hydroxy; C20–C28 for β- and ω-hydroxy) in their total mass for the snow samples collected during the
winter of 2010 and 2011. The upper and lower horizontal bars for each type of
hydroxy fatty acid indicate the data for 2010 and 2011, respectively.
As mentioned in Sect. 2.3, we have estimated the abundances of endotoxin and
mass loading of GNB in fresh snow samples. This quantification is indeed
crucial for assessing a likely allergic impact of endotoxin globally via
long-range atmospheric transport. Here, we estimated the endotoxin
concentrations in snow as varying from 424 to 1080 ng kg-1 (ave.
789 ± 237 ng kg-1) in 2010 and from 36 to 1100 ng kg-1
(ave. 579 ± 435 ng kg-1) in 2011 samples. The estimated lower
limits of endotoxin in Tables 1 and 2 are calculated based on the minimum
concentration of β-hydroxy FAs (C10–C18), which are
specific to Gram-negative bacteria (GNB). β-Hydroxy FAs
(C10–C18) are the structural constituents of lipid A, which are
present in the outer cell membrane of GNB. Thus, the endotoxin concentrations
in snow samples were estimated based on the abundances of β-hydroxy
FAs having a carbon chain length from 10 to 18 (Sect. 2.3). Consistent with
this study, Lee et al. (2004) also reported endotoxin concentrations based on
β-hydroxy FAs (C10–C18). Although relative abundances of
endotoxin during the winter of 2010 (N = 5) are higher than those of
2011 samples (N = 6), the two-tailed t test revealed no significant
differences (t = 0.96; df = 9; P > 0.05) with regard
to mean concentrations of the 2 years.
Conceptual model to explain the scavenging of hydroxy fatty acids
(FAs) by fresh snow in the free troposphere. Snowfall in northern Japan acts
as a filter in reducing the hydroxy FAs (tracers of Gram-negative bacteria;
GNB), which in turn results in the removal of endotoxins from the atmosphere
and a reduction in their health effects during long-range aeolian dust
transport.
In this study, we estimated dry mass concentrations of GNB in snow samples to
be 26.3 ± 7.9 µg kg-1 in 2010 vs.
19.3 ± 1.4 µg kg-1 in 2011. Lee et al. (2007) reported
that airborne endotoxin is of crustal origin and thus can be transported long
distances to the outflow region. Since the AMBTs reveal the impact of
long-range transport from Russia and Siberia during the study period, we
infer that estimated endotoxin concentrations and dry cell weight of GNB over
Sapporo are derived from those source regions. Recently, Golokhvast (2014)
documented the airborne biogenic particles in snow from the Russian Far East
that cause allergy for the pedestrians. The airborne biogenic particles can
be scavenged efficiently by both wet precipitation and snowfall. Therefore,
we have looked for the literature describing the occurrence of GNB in
rainwater for comparison with our study on Sapporo snow. Towards this,
Gould (1999) and Lye (2002) have documented the presence of various GNB
(e.g., Salmonella, Shigella, Vibrio, Legionella and
Campylobacter
spp.) species in
rainwater. Likewise, Kawamura and Kaplan (1983) also reported the presence of
β-hydroxy FAs in rainwater samples collected from Los Angeles (USA)
and attributed their sources to bacterial membranes. So far, no literature
has been available on endotoxin and GNB concentrations in snow samples from
East Asia in order to make a comprehensive comparison with the present study.
Overall, the presence of endotoxin and GNB in snow affirms that biogenic
particles of soil microbes and their potential health impact should not be
overlooked. Routine and long-term measurements of airborne chemical markers
(hydroxy FAs in this study) could aid the monitoring of the microbial
content in long-range transported air masses. Further studies are required
to examine their distributions in the atmospheric environment and health
effects on human beings in the regional and global perspectives during
long-range atmospheric transport.
Conclusions
Although low temperature is considered to be a limiting factor for bacterial
activity in air/snow, some studies have shown that bacteria can be
metabolically active even at subzero temperatures (Polymenakou, 2012, and
references therein). Figure 4 summarizes the whole idea, which was addressed
in this study. We conclude that fresh snow in Japan acts as a filter, which
aids in reducing the burden of pathogenic microbes from the atmosphere via
wet scavenging of these particles.
Owing to prolonged winters and, thus, snowfall in Sapporo, it is likely that
ambient bacterial endotoxins (LPS) are largely scavenged from the atmosphere
by snow, which can decrease their effect on human health via inhalation
(Jacobs, 1989; Milton, 1996). However, without snow scavenging, ambient
bacterial endotoxin levels may stay high and can be transported further long
distances, which can cause severe impacts on human health over the North
Pacific and possibly in North America. Overall, bacteria and their debris
(biomass) can be evaluated in aerosols that are scavenged by snow in the free
troposphere without prior culture by the determination of hydroxy FAs for
both LPS and GNB.