The 20-cm depth soil samples were collected from the GZAAS Huadu and Nansha bases, respectively, according to a nine-point sampling method, well-mixed and air dried in our laboratory. The dried soil then went through a 0.15 micrometer, 300-mesh sieve and was stored at −20˚C, for the uses of soil microbe investigation and analysis of physical and chemical soil properties.

The DNA of soil samples were extracted using a soil DNA extraction kit after they went through a sieve. The obtained DNA was subject to quality verification, PCR amplification (16S rDNA V3-V4 region for bacteria and ITS2 region for fungi) and then high through-put sequencing of the PCR products by Guangzhou Gene Denovo Bio-Tech Ltd. Co. (Guangzhou, China). The obtained data were assembled into sequence tags and subject to BLAST in GenBank for classification of the microbes, followed by OTU, diversity and inter-sample comparative analyses.

Nine environmental factors―pH, available phosphorous, basic-group nitrogen, available potassium, hydrogen peroxide, sucrose, urease, phosphatase, and organic matters―were measured in the Huadu and Nansha soil samples using standard protocols, with an original uncultivated soil sample (maintained in our laboratory) as the control. A canonical correspondence analysis (CCA) was conducted to reveal the association between the dominating microbial populations (on the genus level) in the soil samples and these factors.

32.9 and 33.0 Mbp nucleotides from fungal ITS2 regions were obtained from the Huadu and Nansha samples respectively, and then the clean data were assembled into 117,012 and 118,378 sequence tags, respectively. The Huadu sequences ranged 201 - 349 bp, and the Nansha ones 201 - 349 bp in length. For the bacterial 16S rDNA regions, 61.6 and 68.2 Mbp nucleotides containing 204,507 and 194,103 unique tags, respectively, were obtained from the Huadu and Nansha samples.

A GenBank BLAST analysis of the assembled sequences indicated that in the Huadu but not Nansha sample, abundant sequences of Fusarium (e.g. F. oxysporum), known as a plant pathogenic fungal genus, were identified, whereas fungi from genera Clostridium, Devosia, Bradyrhizobium, Bdellovibrio, Phenylobacterium, Penicillium, and Emericella dominated in the Nansha soil sample, and the Emericella sp., Penicillium menonorum and P. pimiteouiense were found in the Nansha but not in the Huadu sample (Figure 1). A terpenoid, Emericella-sp-d01, has been found as a seaborne fungus-derived secondary metabolite with antibiotic activity. P. menonorum may promote plant growth and nutrient management, while P. pimiteouiense play the roles as calmodulin and maintenance factor for plants. It is suggested that these three particular microorganisms are associated with the special soil environment in the Nansha base under constant influence of saline tide, and that they may be conducive to the growth of crops’ above-ground parts.

Both the Huadu and Nansha samples contained diverse bacterial groups. The dominating genera in the two samples were Sphingomonas and Acinetobacter respectively, and genera Actinomycetes, Bacillus and Marinobacter were present with different abundances in both samples (Figure 2). The Huadu sample

generated 2260 Bacillus tags, significantly more than the Nansha sample (18 tags). In contrary, the Nansha sample contained much abundant Actinomycetes (52828 tags) and Marinobacter (21978 tags)bacteria than the Huadu sample (17976 and 8628 tags respectively). It has been reported that higher Bacillus titeris harmful while Actinomycetes is beneficial, to plant growth [1] [2], suggesting that the soil environment in the Nansha base is more suitable for vegetable plantation than the Huadu base.

Compared to the control, both the Huadu and Nansha soil samples had higher pH and contents of eight types of indexing substances: available phosphorous, basic-group nitrogen, available potassium, catalase, sucrose, urease, phosphatase, and organic matters (Table 1). Both the samples were weekly acid (pH 6.16 and 6.22). The Nansha sample contained much lower amounts of available

phosphorous and basic-group nitrogen but higher available potassium, catalase, sucrose, urease, phosphatase, and organic matters (Table 1). The result suggest that the saline and alkaline soil in the Nansha base adjacent to the sea can be improved and developed into high-quality plough land via appropriate cultivation measures and use of fertilizers.

Our results suggested that the ten microbial genera dominating in the Huadu sample, the Nansha sample, and the control (known with Gemmatimonas, Pseudomonas, and Paenibacillus dominating) respectively were influenced by the nine environment factors. The canonical correspondence analysis (CCA) indicated that the microbes from genera Clostridium, Devosia, Bradyrhizobium, Bdellovibrio, Phenylobacterium, and Paenibacillus were mainly impacted by pH in a positive correlation; that from genera Sphingomonas and Acinetobacter mainly by available phosphorous content in a positive correlation; and that from genera Gemmatimonas and Pseudomonas by both pH and available phosphorous content in a negative correlation (Figure 3).