Foliar particulate matter retention and toxic trace element accumulation of six roadside plant species in a subtropical city

As a major source of air pollution, particulate matter (PM) and associated toxic trace elements pose potentially serious threats to human health and environmental safety. As is known that plants can reduce air PM pollution. However, the relationship between PM of different sizes and toxic trace elements in foliar PM is still unclear. This study was performed to explore the association between PM of different sizes (PM2.5, PM10, PM>10) and toxic trace elements (As, Al, Cu, Zn, Cd, Fe, Pb) as well as the correlation among toxic trace elements of six roadside plant species (Cinnamomum camphora, Osmanthus fragrans, Magnolia grandiflora, Podocarpus macrophyllus, Loropetalum chinense var. rubrum and Pittosporum tobira) in Changsha, Hunan Province, China. Results showed that P. macrophyllus had the highest ability to retain PM, and C. camphora excelled in retaining PM2.5. The combination of P. macrophyllus and C. camphora was highly recommended to be planted in the subtropical city to effectively reduce PM. The toxic trace elements accumulated in foliar PM varied with plant species and PM size. Two-way ANOVA showed that most of the toxic trace elements were significantly influenced by plant species, PM size, and their interactions (P < 0.05). Additionally, linear regression and correlation analyses further demonstrated the homology of most toxic trace elements in foliar PM, i.e., confirming plants as predictors of PM sources as well as environmental monitoring. These findings contribute to urban air pollution control and landscape configuration optimization.

elements in the environment 23 .Nonetheless, there is still a lot of confusion about foliar PM and associated toxic trace elements.What is the composition of toxic trace elements in PM of different sizes retained by plants?Is it possible that plant species affect PM of different sizes to retain toxic trace elements?Who determines the concentration of toxic trace elements in foliar PM?
In this study, six roadside plants were investigated in Changsha, Hunan Province, a typical subtropical city in China.We measured the content of PM in different size fractions on their leaves and associated toxic trace elements.We also analyzed the association between PM of different sizes and toxic trace elements as well as the correlation among toxic trace elements, which is rarely seen in previous studies.The objectives of this study were to (1) evaluate the ability of different plants to retain particulate matter of different size fractions as well as toxic trace elements, (2) analyze driving factors of toxic trace element content in particulate matter, and (3) reveal potential associations between toxic trace elements from particulate matter.Our results can contribute to the knowledge of the ability of common roadside plants to retain particulate matter, toxic trace elements, and their potential associations in the subtropical area.Simultaneously, our study can provide a theoretical basis for the plant configuration in urban green belts and the application of plants in environmental monitoring.In addition, our work can provide innovative perspectives for the study of foliar PM.

Results
Foliar PM retention.Figure 1 shows the differences among plant species in the retention of PM in different size fractions.The total PM retention capacity of P. macrophyllus was the highest (3.8464 g/m 2 ), which was significantly higher than that of the other five plants (P < 0.05).Podocarpus macrophyllus also had the greatest retention of PM 10 (0.1426 g/m 2 ) and PM >10 (3.5865 g/m 2 ).PM 2.5 accumulated on the leaf surface of C. camphora was the highest (0.4907 g/m 2 ), which accounted for 91.85% of the total PM retention of C. camphora, while C. camphora had the lowest PM >10 retention (0.0160 g/m 2 ), and it indicated that C. camphora was much more effective in accumulating fine particulate matter.Although O. fragrans had the lowest retention of total PM (0.3958 g/m 2 ), PM 2.5 retention of O. fragrans (0.2986 g/m 2 ) was only lower than that of C. camphora, and significantly higher than that of P. macrophyllus (0.1174 g/m 2 ), L. chinense var.rubrum (0.0254 g/m 2 ), M. grandiflora (0.0101 g/m 2 ) and P. tobira (0.0004 g/m 2 ).PM 2.5 , PM 10, and PM >10 on the leaf surface accounted for 75.43%, 17.60%, and 6.97% of the total PM retention of O. fragrans, which indicated that O. fragrans also was more effective in accumulating fine particulate matter.
Toxic trace elements in foliar PM. Figure 2 shows the toxic trace element concentrations in foliar particulate matter of different plants.The distribution of toxic trace elements in PM of the same particle size was different among the six roadside plants.Cinnamomum camphora carried the most amount of Al, Cu, Zn, Fe, and Pb in foliar PM >10 among the six roadside plants, and there were significant differences in the Al, Cu, Zn, Fe, and Pb concentrations in foliar PM >10 between C. camphora and the other five plant species (P < 0.05).The highest Al, Zn, Cd, Fe, and As concentrations were found in foliar PM 10 of P. tobira, which was significantly higher than the other five plant species (P < 0.05).For Al, Zn, Cd, and As, P. tobira had the highest concentration in foliar PM 2.5 among the six roadside plants.The highest Fe concentration in foliar PM 2.5 was observed in P. macrophyllus, which was significantly higher than C. camphora (P < 0.05).
The distribution of the same element varied with plant species and particle size.(20,169 mg/kg)] had the highest content of Al in PM 2.5 ; and there were significant differences among the three particle sizes in the same plant species (P < 0.05), while the Al content of L. chinense var.rubrum had no significant difference among the three particle sizes (P > 0.05).The distribution of Cd and As in the particulate matter of different plant leaves showed almost complete consistency, as evidenced by the highest levels of PM >10 in C. camphora foliage, PM 2.5 in M. grandiflora foliage, PM 10 in P. tobira and L. chinense var.rubrum foliage, and the Cd and As concentrations of O. fragrans and P. macrophyllus showed no significant difference among the three particle sizes (P > 0.05) (Table S1).
The results of two-way ANOVA (Table 1) showed that most of the toxic trace elements were significantly influenced by plant species, PM size, and their interactions.Al (F = 36.19,P < 0.001), Cd (F = 6.64,P < 0.001), and As (F = 7.63, P < 0.001) were more driven by the interaction of plant species and PM size, while some elements were also significantly influenced by these factors but were dominated by PM size, such as Cu (F = 8.70, P < 0.001), Pb (F = 11.63,P < 0.001), and Fe (F = 6.53,P < 0.01).More interestingly, there was no significant effect of PM size on Zn (P > 0.05).There was a highly significant effect of plant species on Zn in the particulate matter (P < 0.01).
Correlation analysis of toxic trace elements.Linear regression analysis of toxic trace element concentrations in foliar PM and the ability of foliar PM retention showed that there was a significant correlation between Al, As, Cd, and Zn in PM 2.5 and foliar PM 2.5 retention (P < 0.05).However, there was no significant correlation between the seven elements and foliar PM 10 , PM >10 retention (Fig. 3).

Discussion
Particulate matter retention among plant species.Urbanization has accelerated and increased the pressure on transportation, which has also led to a rapid escalation of particulate matter from roads [24][25][26] .Numerous studies have demonstrated that the roadside plant species can effectively adsorb and remove particulate matter of different sizes, which to a certain extent alleviates the environmental pollution caused by urban transportation [27][28][29] .In our study, P. macrophyllus had the highest retention of total PM, especially much more effective in accumulating large PM particles, which was consistent with previous research [30][31][32] .It has been proved that coniferous species can retain particulate matter more effectively than broadleaf species due to their unique leaf structure and microscopic characteristics.The microstructure of coniferous species is neat "ridged" stripes and well-arranged stomatal strips, with a large number of wrinkles and fluffy structures in the stomatal area, and the size of the fluff is nanoscale, which increases the surface area and surface roughness and is more conducive to the deposition of intercepted particles 30 .Narrow needles may be more easily hit by particles in the air than large    flat leaves, compared to flattened leaves, narrow conifer needles have a much larger Stokes number and therefore have higher capture efficiency 31 .Podocarpus macrophyllus was not considered outstanding in its ability to retain PM 2.5 as there were physiological differences between different conifer species 32 , and an appropriate selection of conifer species can lead to better air quality improvement 12 .Besides, we found C. camphora was preferable to P. macrophyllus and other plants in terms of their ability to retain PM 2.5 .Current studies on foliar PM 2.5 capture capacity for evergreen broad-leaved trees (C.camphora) are inconclusive.Some researchers found that C. camphora did not effectively capture small-size particulate matter 33,34 .But others proved that evergreen broadleaved trees had greater adsorption capacity of PM 2.5 than coniferous trees 31 , and this is consistent with our results.The reason was that the leaf stomata can adsorb PM 2.5

31
, and the broad-leaved trees (C.camphora) have more stomatal density than coniferous trees (P.macrophyllus) [35][36][37] , so C. camphora showed very high adsorption capacity of PM 2.5 .In addition, the subtle changes in ambient water vapor brought by the evapotranspiration of plant leaves would also be effective in retaining PM 2.5 15 .In our study, P. macrophyllus had the highest ability to retain PM, and C. camphora excelled in retaining PM 2.5 .Therefore, our results suggest that the combination of P. macrophyllus and C. camphora was highly recommended to be planted in the subtropical city to effectively reduce PM from the air.
Toxic trace elements in foliar particulate matter.Previous studies mainly focused on the adsorption of particulate matter by plants and the factors influencing the reduction of PM 31 , but the relationship between PM of different sizes and toxic trace elements in foliar PM remains unclear.Our results demonstrated that there was a "particle size effect" for most toxic trace elements in foliar PM.All of the trace element was influenced by particle size except Zn.However, the mechanism of the effects of particle size on trace elements is still unclear, which can be a direction of future research.Additionally, the different content of the same elements in particulate matter of the same particle size indicates that the particulate matter may be from different sources, such as www.nature.com/scientificreports/ the elemental composition of particulate matter originating from industrial areas and non-industrial areas and their content was completely different 34 .To avoid the influence of environmental factors, all the trees studied were in the same road adjacent area.Therefore, the most likely reason is that different plant leaves absorb toxic trace elements from the same source of particulate matter, leading to variations in their concentrations.There was further evidence from the two-way ANOVA results that plant species and the interaction of plant species and particle size had a highly significant effect on the seven toxic trace elements, which has rarely been proved by such methods in previous studies 38,39 .Different plants generally have different foliar microstructures, i.e., some plants are more likely to absorb trace elements from large particulate matter on the leaf surface 40,41 , while others are more inclined to fine particulate matter, such as PM 2.5

42
. All toxic trace elements in PM >10 on C. camphora foliage in our study were higher than those in other particle-size particulate matter.Some plants are selective in the adsorption of particulate matter due to their leaf physicochemical factors, thus explaining why C. camphora has the worst PM >10 adsorption capacity but its ability to retain toxic trace elements is better 43 .
Association between toxic trace elements and their association with foliar retention capacity of particulate matter.In theory, the more particulate matter retained on the leaf surface, the more toxic trace elements should be accumulated accordingly 39,44 .However, the linear regression results showed that some toxic trace elements showed a significant negative correlation with the particulate matter content when the particle size was less than 2.5 µm.This result indicated the toxic trace element may be absorbed by plant leaves.Furthermore, it has been proved that plant species can affect the content of toxic trace elements in foliar particulate matter through their microstructures as well as particulate matter selection 33 .There is a finding that plant health also affects their ability to retain particulate matter.At the same time, particles can also affect plant growth, photosynthesis, respiration, and transpiration when the particles on the leaf surface reach a certain amount 45 .Therefore, the differences in foliar PM retention capacity and toxic trace element accumulation in PM can also be attributed to uncertainties associated physical and chemical processes involved in plant-particles interactions.
In our study, the phenomenon of a closely association between toxic trace elements in leaf PM was comparable to the results of most studies [46][47][48][49] .Even though the plant leaf samples were collected in the same time-space, there were some discrepancies in the sources of PM even in the same area, for instance, roadside PM may originate from car exhaust emissions or road PM 36 .There are records suggesting that Cd, Cu, Pb, and Zn generally originate from road PM 50,51 , while vehicle exhaust emissions bring As, Cd, and Pb 52,53 .The highly significant correlation between As and Cd in all particle sizes in this study further confirms the above statement and also illustrates that PM retention on plant foliage can reflect fairly well the sources as well as the distribution of toxic trace elements in the environment of roadsides 54,55 .

Conclusions
In this study, P. macrophyllus had the strongest ability to retain PM, and C. camphora excelled in retaining PM 2.5 .The combination of P. macrophyllus and C. camphora is highly recommended to be planted in subtropical cities to effectively reduce PM.PM size and plant species and their interactions played an important role in the toxic trace element concentration of foliar particulate matter.Additionally, the toxic trace element concentration of foliar particulate matter is co-dependent on the physiological characteristics of the plant itself as well as environmental sources.The plants can predict the source of PM by toxic trace elements.This study provides a new reference for city greening and air pollution control.

Materials and methods
Study area.Changsha is located in the eastern part of Hunan Province (111° 53′-114° 15′ E, 27° 51′-28° 40′ N), which is in the low latitude zone with a subtropical monsoon climate.The annual average temperature is 17.2 °C, and the annual precipitation is 1361.6 mm in the city.The annual average PM 2.5 concentration in 2017 was higher than the limit value of the Chinese Secondary Standard of Environmental Air Quality Standards (GB 3095-2012) (35 μg/m 3 ), and the annual average PM 10 concentration is equal to the limit value of the Chinese Secondary Standard of Environmental Air Quality Standards (GB 3095-2012) (70 μg/m 3 ).

Sample collection.
By surveying greening tree species in Changsha, we selected six typical evergreen broad-leaved and coniferous greening tree species in Changsha to monitor and analyze their PM retention capacity, including four arbor species: Cinnamomum camphora, Osmanthus fragrans, Magnolia grandiflora and Podocarpus macrophyllus, and two shrub species: Loropetalum chinense var.rubrum and Pittosporum tobira (Table 2).The plants we selected were not listed as national and provincial key protected wild plants in China nor threatened species on the IUCN Red List.Therefore, no specific permissions or licenses were needed for the sampling of plants for research purposes according to the regulations of the People's Republic of China on the protection of wild plants.During the sampling process, we followed the local sampling guideline to ensure no substantial harm to the collecting individual.The plant leaf samples were collected on both sides of the main roads (Shaoshan Road and Furong Road) in Changsha City (Fig. 5) on December 25, 2017, when there was no previous rainfall for more than 7 days.The width of the green belt was about 3 m, and the plants in the green belt included arbors, shrubs, and herbs with consistent habitat conditions.To avoid the impacts caused by the differences in location and distance, the samples were selected within 3 m from the road and all samples were collected within 1 day.The height of sampling was 2-6 m for arbors and 0.5-3 m for shrubs.For each species, ten trees were selected as samples, and the leaves were collected randomly in four different directions (east, west, south, and north) at the upper, middle, and lower positions of the tree canopy.The leaves collected are required to be healthy and free of pests and diseases.After sample collection, the samples were carefully stored in sealed plastic bags and immediately transferred to a refrigerator at 4 °C for subsequent experiments and analysis.Meanwhile, the diameter at breast height (DBH) and basal diameter (BD) of plants were measured with a diameter at breast height ruler, the height of each tree was measured with a height measuring device, and the width of the crowns was measured with a ruler in two directions: north-south and east-west.
Extraction and processing of foliar particulate matter.The retention of particulate matter on the surface of plant leaves was determined by using the microporous membrane weighing method 53 .The number of test leaves was determined according to leaf size and type to ensure that the experimental leaf samples were adequate and randomized, and three replicates of each plant were required.In general, the larger leaves in broadleaf needed 30-50 pieces, and the smaller leaves needed 150-300 pieces.Put the selected plant leaf samples into a beaker with deionized water for half an hour, and then carefully clean all the PM on the leaves with a small brush.Next, pinch out the leaves with pointed tweezers (be careful not to damage the leaves) and rinse them three times with an appropriate amount of deionized water, then put them on newspaper to dry.Finally, the dried leaves were scanned with a scanner (Epson scanner 11000G) to obtain the leaf projection, and export all the leaf scanned images, then opened Image J software, the edges of the leaves were circled and the leaf area was calculated by the software.To reduce the error, the area of each leaf needed to be calculated three times and get the average value.
The solution was filtered through the dried and weighed microporous membrane of 10 μm pore size, and then the filtrate was filtered through the dried and weighed microporous membrane of 2.5 μm and 0.1 μm pore size to obtain three different particle size levels of PM >10 , PM 10 and PM 2.5 using the same procedure as above.www.nature.com/scientificreports/Before and after each filtration, the microporous filter membrane was put in an oven at 60 °C31 and dried to a constant mass (two measurements ≤ 0.0002 g), and weighed on a balance with an accuracy of one ten-thousandth.The plant leaf PM retention was determined by the mass difference method.The calculation method was as follows: PM retention per unit leaf area (g/m 2 ): In the formula, M is the retention of particulate matter per unit leaf area (g/m 2 ); M 1 is the dried mass of filter membrane before filtration (g); M 2 is the dried mass of filter membrane and particulate matter after filtration (g); S is the total area of test leaf samples (m 2 ).
Toxic trace element accumulation of foliar particulate matter.The content of toxic trace elements in PM of plant leaf surface was determined by aqueous solubilization digestion.Approximately 0.2 g of filter membranes were digested in a glass tube with HNO 3 and HCl mixture (1:3 v/v ratio) using a graphite heater furnace (Polytech PT60, Polytech 3 Instrument, Beijing, China) and kept the mixture in a slightly boiling state for 2 h.Then the content of toxic trace elements in the digestion solution was determined by the Inductively coupled plasma optical emission spectrometer (ICP-OES, Optima 8300, USA), and the detection limit of this instrument is 0.001-0.1 mg/L.The national standard GSS-5 (Hunan soil) was selected for quality control and the measured standard recoveries ranged from 85.1 to 114%.In this experiment, As, Al, Cu, Zn, Cd, Fe, and Pb, which are common toxic trace elements in PM, were measured.The standard solutions were selected from the national standard samples of single-element standard solutions: As (GSB 04-1714-2004), Al (GSB 04-1713-2004), Cu (GSB 04-1725-2004), Zn (GSB 04-1761-2004), Cd (GSB 04-1721-2004), Fe (GSB 04-1726-2004) and Pb (GSB 04-1742-2004), with the concentration of 1000 μg/ml.

Statistical analysis.
One-way ANOVA was used to compare the differences in the PM retention capacity of different plants.Two-way ANOVA was used to reveal the underlying drivers of toxic trace element concentrations in particulate matter.Linear regression was used to explore the relationship between toxic trace element concentrations of particulate matter and the ability of the foliage to retain particulate matter.Correlation analysis based on Pearson's method was used to examine the association among different toxic trace elements.GraphPad Prism 8 as well as the R-based packages "ggplot2" and "GGally" were used for data visualization.All statistical analyses in our work were done in R 4.1.1(R-Core-Team, 2013).
Ethical approval.All procedures were followed in compliance with institutional, national, and international rules and legislation.The plants we selected were not listed as national and provincial key protected wild plants in China nor threatened species on the IUCN Red List.Therefore, no specific permissions or licenses were needed for the sampling of plants for research purposes according to the regulations of the People's Republic of China on the protection of wild plants.

Figure 1 .
Figure 1.Particulate matter (PM) retention capacity of different plants.(a) Total PM; (b) PM 2.5 ; (c) PM 10 ; (d) PM >10 .Different lowercase letters in the graph represent significant differences in the PM retention ability of tree species (P < 0.05).

Figure 2 .
Figure 2. Toxic trace element concentration in foliar particulate matter of different plants.Different colors represent different plants and different shapes represent particulate matter of different sizes.

Figure 3 .
Figure 3. Regression analysis of toxic trace element concentrations with foliar PM retention.Different colors are used in the figure to distinguish the different toxic trace elements.Significance levels are indicated by P < 0.05, P < 0.01, and P < 0.001.

Table 1 .
Two-way ANOVA of toxic trace elements influenced by plant species, PM size, and their interactions.P values *< 0.05, **< 0.01, ***< 0.001, F p : plant species, F s : PM size, F P × F S : the interaction of plant species and PM size.

Table 2 .
Basic information of the investigated tree species.† Data are means ± standard deviation (SD).