1 Introduction

Dissolved organic matter (DOM) is the most active component of soil’s carbon pool. Its changes can directly affect carbon sequesterers within soil ecosystems and the total amount of carbon sequestered (Schmidt et al. 2011; Marin-Spiotta et al. 2014; Doetterl et al. 2015; van der Wal and De Boer 2017). Microbes derive energy sources from DOM, which directly influences microbial abundance and community composition, and thus affects biogeochemical cycles (Lehmann and Kleber 2015; Roth et al. 2019). Moreover, DOM can form complexes with heavy metals and adsorb organic pollutants and hence affect their bioavailability (Lv et al. 2016b; Chen et al. 2018; Roth et al. 2019). Therefore, the study of DOM in soil environments has significant environmental and ecological value.

Land use type, temperature, moisture, and local microbial communities are important factors affecting the occurrence and transport of DOM (Lapierre et al. 2015; Campbell et al. 2021; Johannsson et al. 2021; Palviainen et al. 2022). In general, elevated temperature accelerates the degradation of DOM, especially carbohydrates and polypeptides (Liu et al. 2017; Liáng et al. 2023). Therefore, temperature rises can decrease the relative content of carbohydrate-like and peptide-like molecules, while increasing the average molecular mass of DOM as well as the aromaticity (Chen et al. 2018). Warmer temperature also promotes plant growth and thus increases litter input. An increase in soil moisture can promote plant growth and thus increase the input of organic materials to soil, thereby affecting the content and composition of soil DOM. Consequently, the decomposition ability of microorganisms will be enhanced by increased water and organic carbon prevalence. Soil properties such as pH, C/N ratio and iron oxides are additional key factors affecting the molecular diversity of DOM (Li et al. 2018, 2019b).

Large-scale investigations of soil DOM with respect to climatic gradients are important to reveal the influence of climatic factors on soil DOM. Li et al. (2018) studied the DOM content and composition in soil of 16 rice-fields in eastern coastal China, however the relationship between soil DOM composition and climatic factors was not analyzed in this study. Ding et al. (2020) analyzed soil samples from 33 different climate zones to study the composition and influencing factors of soil DOM but did not consider the interference of land use types. Gao et al. (2017) analyzed soil samples from 19 farmland types in mainland China, but it was unclear whether the crop types and farming methods were consistent. The different types of land use and the diversity of vegetation on the surface can cause some disturbances to the composition of soil DOM. Insights into the influence of climatic factors on large-scale soil DOM characteristics therefore require large-scale soil DOM surveys that decouple land use disturbance effects. With the transition of population from rural to urban areas, greenspace in cities is increasing rapidly, which can have a huge impact on land use and lead to huge changes in ecosystem function. Urban greenspace is an important part of urban ecosystem and an important place for urban carbon cycle and nutrient cycle (Sun et al. 2019; Mabon and Shih 2021). In addition, changes in urban green space area can also lead to changes in soil microorganisms (Delgado-Baquerizo et al. 2021).

Climate zones affect the physiological state of soil microbial community to a large extent, and then affect the content of soil DOM. Climate zones can affect plant growth and thus alter the input of organic carbon to the soil, which can cause changes to DOM composition (Nielsen and Ball 2015; Luo et al. 2017). Climate zones can also affect soil microorganisms. Bacteria are the most abundant, diverse, and functionally diverse group of soil microorganisms and are sensitive to climate change (Chen et al. 2016). Wang et al. (2018) studied the spatial distribution pattern of soil microorganisms in semi-arid regions of northern China and found that the microbial community structure was mainly influenced by the average annual rainfall at a large regional scale. This was because precipitation could directly affect soil microbial communities by altering soil water availability and increasing the efficiency of microbial decomposition of surface litter (Manzoni et al. 2014; Sierra et al. 2017). A positive correlation between bacterial abundance and environmental humidity was noted in 24 arid and semi-arid steppes of the Mongolian Plateau (Chen et al. 2015) and in a 3500 km alpine sample zone of the Tibetan Plateau (Chen et al. 2016). After analyzing 16 rice-soil samples in mainland China, Li et al. (2018) found that DOM distribution on the continental scale was closely related to the classification characteristics and metabolic potential of rice microbiota. The revealed microbial communities and the relationship between diversity and DOM are an important basis for understanding the influence of climatic factors on DOM components.

The objectives of this study were to (1) investigate the distribution pattern of DOM at a large scale, (2) investigate the effects of climate zones on DOM components and microbial communities, and (3) investigate the relationship between microbial communities and DOM components. This study involved the collection of 54 urban greenspaces samples from five climate zones in mainland China and utilization of FT-ICR-MS and 16S rRNA sequencing techniques to reveal the relationship between microbial communities and DOM, hence providing a deeper understanding of the impact of climatic factors on DOM components.

2 Materials and methods

2.1 Soil sample collection and geographic and meteorological data

During September 2019, soil samples were collected from urban residential green spaces in 54 cities in mainland China (Fig. 1), with sampling areas spanning five typical climatic zones (18.26◦–47.85◦N, 86.09◦–126.54◦E). The sampling sites were laid out in accordance with the spatial distribution characteristics of China’s urban population – dense in the southeast and sparse in the northwest. To reduce the interference of human factors, the communities where the soil sample collection points were located belonged to the same developer and property management company, so the management method of each sampling point was relatively consistent. Three plots of 4 m2 were randomly selected in each sampling site area. Each plot was selected at a depth of 0–20 cm and collected using a ring knife of size 50 × 50 mm, and six samples were taken from each plot. The collected samples were scraped with a knife to remove 2 cm from the surface layer of the soil, and then all the samples obtained from the three plots were mixed, and then grounded to form a mixed sample after manually removing stones and impurities to form a mixed sample. The samples were collected and placed in marked polyethylene plastic boxes and plastic bottles and transported frozen to the laboratory to air-dry for subsequent determination of soil physical and chemical properties.

Fig. 1
figure 1

Distribution of 54 sampling sites. TroM: tropical monsoon climate; SubM: subtropical monsoon climate; TemM: temperate monsoon climate; TemC: temperate continental climate; AlpP: alpine plateau climate

Full size image

Mean annual temperature (MAT) and mean annual precipitation (MAP) of the sampling sites in the past 10 years were obtained from the Chinese Meteorological Network (http://data.cma.cn/). The physicochemical properties of all sampling sites are shown in Tables S1 and S2.

2.2 Analysis of soil quality, DOC and FT-ICR-MS

The soil samples were extracted with deionized water at solid/liquid ratios of 1:10 (w/v) by a high-speed refrigerated centrifuge (Himac CR-21G; Hitachi, Tokyo, Japan) shaking at 180 rpm for 1 h. The extracted samples were then centrifuged (ZONKIA, HC-3018, China) for 6 min and filtered through a 0.45 μm cellulose acetate membrane filter. All soil extracts were stored at 4 °C prior to analysis. Each treatment was performed in triplicate. The extraction solution was divided into two parts. One part was used for analysis of soil pH, dissolved organic matter (DOM), total nitrogen (TN), total phosphorus (TP), ammonia nitrogen (NH4+-N), nitrate nitrogen (NO3−-N), available phosphorus (AP), available potassium (AK), and DOC concentration. The above indicators were analyzed according to the Water and Wastewater Detection Method. The DOC concentration of soil extracts was measured with a total organic carbon analyzer (Shimadzu, TOC-L, Japan). The other part was used for subsequent mass spectrometry.

DOM was extracted and purified by solid phase extraction using solid phase extraction (SPE) procedures (Bond Elut, PPL, Varian, 1 g per 6 mL). The final eluent concentration (DOC concentration) in each sample was adjusted to 10 mg L-1 by controlling the volume of water filtered through a given cartridge based on the initial DOC concentration. Program blanks were prepared by treating deionized water in the same manner of DOM samples (Kellerman et al. 2014, 2015).

The purified DOM solid was dissolved in chromatographic pure methanol and analyzed by FT-ICR-MS in negative ion mode. Instrument parameters: Magnetic field intensity was 15.0 T, ionization source was ESI. The operating conditions of the instrument were as follows: injection velocity was 120 μL h-1, capillary inlet voltage was 4500 V, capillary outlet voltage was -320 V, iontophoresis analysis cell flight time was 1.0 ms, source accumulation time was 0.001 s, and storage time was 0.1 s. The sampling point was 4 M, and the collection quality range was 100–750 Da. (Lv et al. 2016a, b).

2.3 Analysis of FT-ICR-MS data

After the acquisition of FT-ICR mass spectrometry measurements, the original data set was internally recalibrated using data analysis. In FT-ICR-MS atlas, the molecular formula of C5-50H10-100O0-40N0-2S0-1 was used to identify and calculate the peaks with signal-to-noise ratio (S/N) using Bruker Data Analysis software.

The molecular formula was given according to the requirement that the difference between the measured mass and the calculated mass is less than or equal to 0.5 PPM. No phosphorus—containing compounds were observed in all samples. The final molecular identification results were based on the results of two replicates. Further molecular calculations were performed in the concentration regions with H/C < 2.0 and O/C < 1.2 (Ding et al. 2020). The Van Krevelen diagram uses O/C and H/C as horizontal and vertical coordinates (Leyva et al. 2020) to visualize the chemical composition and structure of molecules. Accordingly, there are seven types of natural DOM compounds: (A) lipids, (B) proteins, (C) carbohydrates, (D) unsaturated hydrocarbons, (E) lignins, (F) tannins, and (G) condensed hydrocarbons (McKenna et al. 2021). The equivalent double bond number (DBE) was used to evaluate the number of double bonds and rings in the molecular formula (McKenna et al. 2021). According to Koch and Dittmar's (2006) method, aromaticity index (AI) was used to evaluate the aromaticity or proportion of aromatic structure of a compound. More detailed information was available in previous studies (Lv et al. 2016a; Liu et al. 2020).

2.4 The determination of bacteria

All DNA was extracted from 0.25 g soil samples using the DNeasy PowerSoil kit according to the instructions provided by the manufacturer. DNA quality was checked by 1% agarose gel electrophoresis and stored at -80 °C until analysis. Variable regions of 16S rRNA V3-V4 were amplified using its universal primer sequence 520 F: AYTGGGYDTAAAGNG; 802R: TACNVGGGTATCTAATCC.

Specific primers were synthesized according to the specified sequencing region 16S V3-V4 regions and PCR was performed to amplify 16S rRNA in soil samples. PCR products were detected by 2% agarose gel electrophoresis, purified by Agencourt AMPure XP system and sequenced based on Illumina Hiseq 2500 for 16S rRNA. The obtained raw data were optimized and counted using QIIME software. The pair-end reads were spliced into one sequence tag, while the sequence quality and splicing effect were quality controlled and filtered. Operational Taxonomic Unit (OTU) clustering and species annotation were performed on non-repeat sequences according to 97% similarity, and intra-sample diversity analysis (α-diversity) and inter-sample diversity analysis (β-diversity) were performed on OTU based on the results of OTU clustering analysis.

2.5 Statistical analysis

Graphs were drawn using Origin v2018 (OriginLab,vNorthampton, MA, USA) and R package ggplot2. Statistical analysis was performed using IBM SPSS Statistics v20.0 (IBM Corp., Armonk, NY, USA). One-way analysis of variance (ANOVA) tests (p < 0.05) were used to test the differences in DOC content among different samplings. Pearson’s correlation coefficients (r) were used to evaluate the relationship between DOM indexes and parameters (p < 0.05). The Shannon index and richness index of bacterial α-diversity were calculated using R package vegan. Bray–Curtis dissimilarity was used to compute the sparse matrices of DOM molecules and bacterial community. PERMANOVA test was conducted to determine whether DOM molecular or bacterial community Bray–Curtis dissimilarity was significantly different between regions. Constrain principal coordinates analysis (PCoA) was used to calculate the gradient in compositional changes of microbial community (based on Bray–Curtis) and different DOM categories (based on Bray–Curtis). Interaction network analysis was plotted by Cytoscape (version 3.0). Each node represents a given OTU or DOM molecule and each edge represents a significant correlation between two nodes. Structural equation modeling (SEM) was used to evaluate the direct and indirect effects of important environmental factors on soil bacteria communities and soil multifunctionality. The environmental factors included geographical factors (latitude and longitude), climatic factors (MAT and MAP) and edaphic factors (soil pH and DOC content). The data needed to be tested for normality prior to modeling, and all environmental variables were log transformed. All SEMs were analyzed using AMOS 20.0 (AMOS IBM, USA). The SEMs were optimized stepwise by adding correlations and excluding variables that were nonsignificant until the satisfactory model was obtained. R package lavaan and maximum likelihood estimation method were used to fit the structural equation model. Considering the small number of DOM samples used in this study, SEMs was available in modified with Satorra-Bentler correlation, and then Bollen-Stine guided test was used for further confirmation. A reasonable SEM needs to satisfy the X2 test of independence (P > 0.05), relative fitting index (CFI > 0.95), and approximate root mean square error (RMSEA < 0.05). The strength of the relationship between the two variables can be evaluated by the path coefficient of the structural equation model and p value.

3 Results

3.1 The differences of DOM distribution

A total of 14,978 molecular formulas of DOM compounds were determined in all soil samples. From the molecular structure, it could be seen that there were obvious differences in the peaks of DOM mass spectrometry for the samples of the five climatic regions (Fig. 2a), among which the SubM had the largest number of DOM molecules and a high response intensity. The types of DOM molecules with the highest proportion in each region were counted. In terms of elemental composition, sulfur element was found in the most abundant DOM molecules in each climate zone (Fig. 2b). Although the single-molecule strength of CHOS type organic compounds was significantly higher than that of CHO and CHON type organic compounds, the overall content of CHOS type organic compounds was small, accounting for 10%-20%. The overall content of CHO compounds accounted for the highest proportion in each region with the largest number of CHO compounds, while the intensity of the single molecule was generally not high (Fig. 2c).

Fig. 2
figure 2

The results of DOM molecular mass spectrometry. a The mass spectrograms of samples of the five climatic regions. b Distribution of 20 DOM molecular formulas with the highest relative abundance for the five climatic regions. c The proportion of CHO, CHON and CHOS type molecules for the five climatic regions (n = 54). *p < 0.05; **p < 0.01; ***p < 0.001

Full size image

Figure 3 showed the distribution of the seven components in five typical climate zones. Figure 4c showed the CPCoA results of DOM component differences in different climate zones. It can be seen that the direct differences in different climate zones were highly significant. Figure 4a showed that the DOM in soil is dominated by lignins, which accounted for more than 30% in each region, followed by proteins, unsaturated hydrocarbons and lipids. The results showed a distribution trend that the content of lignins gradually decreased from south to north, whereas the unsaturated hydrocarbons, lipids and proteins gradually increased.

Fig. 3
figure 3

Van Krevelen diagrams of unique DOM components identified by Venn diagram analysis of the 54 sample sites

Full size image
Fig. 4
figure 4

Distribution differences of DOM molecules and bacterial communities in five climatic zones. a The proportion of 7 components in five climatic zones. *p < 0.05; **p < 0.01; ***p < 0.001. b Predominant phyla and bacterial community at sampling sites, and distribution of bacterial community in each region. The gold bar represents the number of OTUs contained in all the gold dots, and the black bars represent the number of OTUs contained in the black dots. The pie chart shows the phylum of OTU, and the bar chart on the top right shows the top 10 phylums in abundance. c CPCoA analysis of seven DOM components. d CPCoA analysis of bacterial communities

Full size image

3.2 Distribution patterns of bacteria

Figure 4d showed the differences in bacterial community composition in different climatic zones. There were differences in bacterial community composition but no differences were observed (p > 0.05). The results of CPCoA after removing the sampling points in alpine plateau climate (AlpP) were shown in Fig. S2. There were clear differences but no statistical significance in the distribution of bacteria community composition at the sampling points in the four climate zones. Figure 4b showed the dominant bacterial phyla and bacterial communities at the sampling sites, as well as the distribution of bacterial communities in each region, among which the DOM inant bacteria in the soil were Proteobacteria, followed by Acidobacteria and Actinobacteria. The bacterial diversity in the SubM was the highest, and the bacterial diversity in the TroM was the lowest.

3.3 Response of DOM to MAT and MAP

Figure 5a showed the correlations of the seven DOM components with the spatial location of the sampling sites, MAT and MAP, and soil pH and DOM content. Figure S3 showed the specific value of correlation between each component and the factor (longitude, latitude, MAT, MAP, pH and DOM). Among them, unsaturated hydrocarbons, carbohydrates, lipids and proteins all showed a decreasing trend with the increase of MAT and MAP while condensed hydrocarbons, tannins and lignins showed opposite trends. SEM showed that the response patterns of unsaturated hydrocarbons and lignins to longitude were negatively correlated, while lignins were negatively correlated to latitude. While the proportion of condensed hydrocarbons increased with increasing longitude, the proportion of lipids increased with increasing latitude. At the same time, the DOM components had clear spatial differences with the climate zones. The model showed that the DOM components responded mainly to MAP. The response patterns of lignins and proteins to MAP were negatively correlated, while lipids and tannins were related to MAP with a positive correlation. Figure 5b showed that the β-diversity of DOM increased with the increase of spatial distance and climate difference. The β-diversity of bacteria did not show any significant changes.

Fig. 5
figure 5

Correlation between seven DOM components and environmental factors. a Spearman correlation analysis of DOM components and α-diversity with environmental factors. *p < 0.05; **p < 0.01. b Bray–Curtis distances of DOM (blue), microorganisms (grey) and geographic and climatic distances. The depth of color represents the distance of edaphic factors. c DOM components SEM with geographic and climatic factors. Positive correlations are indicated using red lines, negative correlations are indicated using black lines, and non-significant correlations (p > 0.05) are indicated using dotted lines. R2 indicates the variance of DOM composition explained by the model. Specific index(X.2, df, P, CFI, RMSEA) of each model are shown in Table S3

Full size image

3.4 The relationship between DOM and microorganisms

Figure 6a showed the interaction network analysis which selected the top 100 dominant bacterias and dominant DOMs. It was found that acid bacteria were strongly affected by the dynamic changes of DOM molecules, and in this study CHO-type molecules were mainly metabolized. Figure 6b showed that the diversity and richness of DOM were mainly affected by the diversity and richness of bacteria. It appears that the diversity of DOM has the same trend with the bacterial diversity. While higher bacterial diversities reduce the richness of DOM, higher bacterial richness lowers the diversity of DOM. The correlation analysis between the dominant genera and DOM components and diversity in bacteria showed that Actinobacteria, Bacteroidete and Verrucomicrobia can increase the content of proteins in soil, and Chloroflexi can increase lignin and protein content (Fig. 6c). In addition, the top 100 genes by relative abundance and the top 100 DOM molecules by relative abundance were selected to construct a coexistence network (Fig. S4), which showed that glycoside hydrolase (GH) was the one most closely related to DOM molecules. It was followed by glycosyltransferases (GT), the carbohydrate-binding modules, whereas certain GH and CHO-type organic compounds were more closely related.

Fig. 6
figure 6

Correlation of seven DOM components with bacteria and bacterial diversity. a Interaction network analysis of top 100 most abundant bacterial community and top 100 most abundant DOM molecules that were significantly correlated (p < 0.05, |r|> 0.4). Circles represent bacterial community; triangles represent DOM molecules; bacterial community and DOM molecules relative abundances are set proportional to node size. Labile DOM formula: H/C > 1.5, Recalcitrant DOM formula: H/C < 1.5. Positive correlations are indicated using red lines; negative correlations are indicated using black lines. The thickness of the line indicates the strength of the correlation. b SEM of DOM and microbial diversity and abundance. Positive correlations are indicated using red lines; negative correlations are indicated using black lines; non-significant correlations are indicated using dotted line. c Spearman correlation analysis of seven DOM components and dominant bacterial communities. The black box: p < 0.05

Full size image

4 Discussion

4.1 Differences of DOM components in different climatic regions

The results of this study showed that the components of DOM at molecular level were significantly different among the different climate zones. By selecting soil in urban greenspaces as the research object, the effect of land use types on DOM components could be largely excluded in this study. Therefore, the results of this study indicate that climatic factors can largely influence the composition of soil DOM.

This study showed that the content of CHO type organic compounds was the highest, accounting for more than 50%, followed by CHON and CHOS. Previous studies indicated that CHOS molecule is the most stable, while CHO molecule is more active and unstable (Schmidt et al. 2011; Ohno et al. 2014). Among the dissolved organic components in soil, heteroatomic compounds under different climate, latitude and soil physicochemical factors show the trend that CHO molecules are the most, followed by CHO. Li et al. (2019b) collected rice soil samples from 16 sites in China, and the detected DOM components of CHO and CHON accounted for 44.3% and 42.9%, respectively, and CHOS accounted for 12.4%, which was consistent with the results in this study. However, in this study, there was a sample site in SubM with a lower CHO and CHON type molecule and a higher CHOS type molecule content, but its overall pattern was consistent with other sample sites.

The soil dissolved organic compounds showed the distribution trend of climatic zone. DOM were mainly lignin content, and the lignin content decreased gradually from TroM to AlpP. This was followed by proteins, unsaturated hydrocarbons and lipids, which increased from TroM to AlpP. Therefore, there was a significant increase in organic carbon of plant origin from cold and dry climates to warm and humid climates. Tannins and lignins are however generally more difficult for microorganisms such as bacteria to break down, so soil DOM contained more tannins and lignins in the warm-humid climate zone. Conversely, unsaturated hydrocarbons, lipids, and proteins are more easily degraded by microorganisms such as bacteria. Roth et al. (2019) found that DOM content is higher when the climate is warm, and lower when the climate is dry and cold. In general, as MAT and MAP increases, plants and bryophytes grow faster which results in more organic carbon input into the soil. Although there may be more inputs of such compounds in warm and humid climates, it can be inferred that microbial decomposition may outpace plant inputs. Therefore, there were relatively few unsaturated hydrocarbons, lipids, and proteins in the warm and humid regions. It can be seen that the composition of DOM in soils in warm and humid climates is largely dependent on microbial decomposition rather than exogenous input.

4.2 Differences in bacterial communities in different climate zones

The biomass, activity and community composition of microorganisms jointly determine the degradation of organic carbon. In general, with the increase of temperature and precipitation, the biomass and activity of soil microorganisms increase. The response of microbial community composition to climate zone is more complex. Bacteria constitute the largest proportion of the soil microbiome. In this study, the bacteria in soil were mainly Proteobacteria, Acidobacteria, and Actinobacteria. The bacterial diversity from high to low was SubM, TemM, TemC, AlpP and TroM. The SEM showed that both MAT and MAP determined by the latitude and longitude could significantly affect the diversity index of bacterial communities. In general, elevated temperature and increased precipitation can promote the metabolic capacity of soil bacteria. Wang et al. (2018) studied the spatial distribution pattern of soil microorganisms in semi-arid regions of northern China and found that the microbial community was mainly affected by the MAP at a large regional scale, while it was mainly affected by local soil and vegetation cover at a microhabitat scale. Liu et al. (2014) analyzed 26 soil samples from a black soil region in the northeast of China and found that the effect of geographic distance differences was weaker than that of soil pH and DOM on the formation of bacterial communities. In another study, An et al. (2019) analyzed 42 wetland systems nationwide and found that the relative contribution rate of soil physicochemical properties to the microbial community structure was greater than the spatial pattern. The above studies showed that the climate zone could affect the bacterial community diversity, but the relative proportion of bacteria in a certain phylum or a certain genus did not show a significant promotion or inhibition, which was probably related to the selection of environmental factors. With the increase of spatial distance, the differences among bacterial communities did not increase significantly, which further indicated that the selection of local environment had a great influence on bacterial communities.

4.3 Relationship between DOM and climate zone

The variations of MAT and MAP in different climate zones selected in this study were consistent, and the order from large to small was TroM, SubM, temperate monsoon climate (TemM), AlpP and temperate continental climate (TemC). According to the element composition analysis, although there were differences in CHO-type organic carbon and CHON-type organic carbon between individual climatic regions, there was no obvious trend regarding the climatic gradient. At a molecular level, however, tannins and lignins showed a declining trend from the warm and humid climate zone to the dry and cold climate zone, while unsaturated hydrocarbons, lipids, and proteins showed an increasing trend.

It was also shown that the differences in DOM components increased significantly with increasing geographic distance (Fig. S1). The Spearman correlation analysis showed that MAT, MAP, latitude and longitude and pH all had significant correlations with each component of the DOM. However, it can be seen from the SEM that longitude and latitude determine both MAT and MAP, and MAP directly affects the composition of the DOM, while the MAT has no significant effect on the composition of the DOM. Among them, MAP had a significant positive effect on tannin and lignin, and a significant negative effect on lipids and proteins (p < 0.05). MAP can therefore be considered the primary factor affecting the differences in DOM components in different climate zones.

4.4 Relationship between DOM and bacterial community

Different components of DOM in soil had different bioavailability to different bacteria. Therefore, DOM from different sources or components can selectively promote the growth of some bacteria and inhibit the growth of others, thereby altering the bacterial community. On the contrary, microorganisms will preferentially decompose organic carbon with simple molecular structure and suitable carbon–nitrogen ratio, and also secrete some DOM during growth (Liang et al. 2017). These factoers allow bacteria to affect soil DOM composition.

The SEM showed that the Shannon index and Richness index of DOM could affect the Richness index of bacteria to a certain extent, while the Shannon index of bacteria had no obvious effect. In contrast, the Shannon index and Richness index of bacteria had stronger effects on the Shannon index and Richness index of DOM. This fully showed that in the soil ecosystem of this study, the composition of bacteria to DOM was much larger than the community composition of DOM to bacteria. The spatial distribution of DOM components was largely a climatic gradient-driven biological process mediated by microorganisms.

Spearman correlation analysis showed that the proportion of protein in DOM was negatively correlated with bacteria such as actinomyces and pedobacter in Actinomycetes. This indicated that the above-mentioned bacteria had a strong decomposing effect on protein organic carbon. Correspondingly, Ling et al. (2021) found that Proteobacteria, Bacteroidetes and Firmicutes decomposed DOM mainly with unsaturated aliphatic, lipid, protein and amino sugar components, while Acidobacteria, Verrucomicrobia, Chloroflexi and Cyanobacteria were positively correlated with refractory compounds, tannins, and lignin. Cyanobacteria and Chlorobi can carry out photosynthesis, not only without decomposing protein DOM, but may also secrete it, so they have a positive correlation with the proportion of protein.

In terms of the variation trend of molecular diversity of DOM, Proteobacteria was the main bacteria with positive effects, while Acidobacteria and Blomonas were the main bacteria with negative effects. Correspondingly, Takriti et al. (2018) found the presence of characteristic hyphae and microorganisms was capable of secreting enzymes to decompose complex biopolymers and promote elevated DOM diversity in soil (Zimmerman et al. 2011; Jeewani et al. 2020). Fierer’s study showed that Acidobacteria were most abundant in soils with very low C content and were less abundant in soils supplemented with high concentrations of organic carbon; β-Proteobacteria and Bacteroidetes were most abundant in soils supplemented with sucrose and could utilize readily degradable compounds, while there was no predictable response between abundances of α-Proteobacteria, Firmicutes, and Actinobacteria and C composition (Fierer et al. 2007; Ng et al. 2014). Li et al. (2019a) increased the aromatic components of organic carbon in soil by long-term application of organic fertilizers, resulting in an increase in the abundance of related Acidobacteria and Bacillus. Bonanomi et al. (2019) showed that alkyl compounds are positively correlated with actinomycetes, which demonstrates that bacterial diversity and DOM molecular diversity are determined by the chemical properties of DOM with dynamic changes.

It was also found that microorganisms were selective in the degradation of DOM molecules. Lignin, tannins and superpolyhydrocarbons are more difficult to degrade, while unsaturated hydrocarbons and lipids are more easily degraded. At the molecular level, the results of this study showed that the proportion of S-containing compounds was the lowest, and in the network analysis with bacteria, it was found that molecules without S and N had a strong correlation with bacterial genera. Similarly, some researchers found that small molecules of plant origin are preferentially consumed by microorganisms and biosynthesized into larger molecules (Roth et al. 2019).

Our results showed that climatic gradients did not selectively promote or limit the growth of certain bacteria. However, the abundance of carbon cycle-related genes in bacteria in different climatic regions is different. This indicated that climatic gradients selectively alter the genetic composition of the bacterial carbon cycle. According to the analysis results of elemental composition by FT-ICR-MS, CHO in DOM of AlpP and TemC with lower temperature and precipitation is higher than that of other climate zones, while the proportion of organic carbon containing nitrogen is the lowest. Figure S4 showed that GH can mainly promote the formation of CHO type and reduce CHOS type molecules, while GT can promote the formation of CHOS type. Cardenas et al. (2015) found that the increased abundance of GH can promote the decomposition of lignins and cellulose.

Therefore, from cold-dry regions to hot-humid regions, GT and other genes of carbon cycle in soil bacteria increase, which promotes the decomposition of CHO and other substances, thus causing the spatial distribution pattern of CHO type gradually decrease and CHOS molecules gradually increase in DOM components.

5 Conclusions

  1. (1)

    CHO had the highest proportion of DOM in soil urban greenspaces (40–80%), followed by CHON (20–50%) and CHOS (10–20%). At molecular level, the DOM in soil was dominated by lignins, followed by proteins, unsaturated hydrocarbons and lipids.

  2. (2)

    There were significant differences in the distribution of DOM components at large scale, and the SEM showed that DOM components mainly responded to MAP, among which lignins and proteins showed a negative response pattern to MAP, while lipids and tannins showed a positive correlation with MAP.

  3. (3)

    The differences in bacterial community composition among different climatic zones were not significant, and the dominant bacteria in the soil were Proteobacteria, followed by Acidobacteria and Actinobacteria. Bacterial diversity was highest in the SubM, lowest in the TroM.

  4. (4)

    The diversity and abundance of DOM were mainly influenced by the diversity and abundance of bacteria. Actinobacteria, Bacteroidete and Verrucomicrobia were negatively correlated with proteins, while Chloroflexi was positively correlated with lignins and proteins. In addition, the SEM indicated that MAT and MAP directly influenced the composition of DOM, while there was a significant correlation between the composition of DOM and the diversity and abundance of bacteria.