Spatiotemporal dynamics of the microbial diversity on salt-preserved goatskins assessed by culturing and 16S rRNA gene amplicon sequencing

Wet-salted skin, as a special artificial high-salt environment, is rich in protein, fat, collagen and other nutrient substrates, and is a rich resource of halotolerant and halophilic microorganisms. However, knowledge gaps regarding the microbial community structure and inter taxa associations of wet-salted skin are large. In this study, the spatiotemporal dynamics and community structure of microorganisms present on wet-salted goatskins were investigated using 16S rRNA gene amplicon sequencing and culturable technique. Alpha diversity analysis based on Sobs, Chao, Ace and Shannon indices revealed that microbial diversity on the wet-salted goatskins exhibited a trend of ‘down → up → down → flat’ with time. During preservation, genera belonging to the bacteria domain such as Acinetobacter, Weissella and Streptococcus were slowly dying out, whereas those belonging to halophilic archaea such as Natrialba and Haloterrigena were gradually flourishing. Moreover, to resist high-salt stress, microorganisms on the wet-salted goatskin gradually migrated from the outside to the inside, eventually leading to the microbial diversity inside the skin being the same as or even higher than that on the skin surface. Venn diagram analysis revealed that the strains of some genera, including Psychrobacter, Salimicrobium, Salinicola, Ornithinibacillus, Halomonas, Bacillus and Chromohalobacter, were distributed throughout the interior and exterior of the wet-salted goatskin and existed during various periods. Accordingly, 45 protease-producing halophilic or halotolerant microorganisms were isolated and screened from the wet-salted goatskin using the gradient dilution plate method. Importantly, 16S rRNA genes of some bacteria exhibited less than 99.5% similarity to valid published species, indicating that they likely are novel species and have a good potential for application.


Introduction
The leather industry is an age-old industry producing various goods such as leather bags, leather shoes and leather garments, which have always been popular and preferred among consumers globally. The raw hides and skins, coming from the meat industry, are used as raw material for this industry. Animal skin, a fibrous proteinaceous material, contains hair, sweat glands, fat, blood vessels, water and collagen fibers [1]. It is more susceptible to be attacked by microorganisms, including bacteria, fungi and archaea, which derived from various sources such as air, water, soil, manure and extraneous filth [2]. Hence, the raw material starts degenerating if it is not properly treated within a few hours after being stripped from the animal [3]. The spoilt skins will severely affect leather quality, causing economic losses and environmental pollution. To date, various skin preservation methods have been discovered by researchers to increase the shelf lives of skin. These methods include sun drying, freezing, and use of sodium chloride, boric acid and aryl alcohol [4]. NaCl application is the most prevalent technique for hide/skin preservation because of its accessibility, low cost and ease of operation [5]. The generally accepted treatment is to add 30-50% (w/w) NaCl to the inside and outside of the skin according to skin weight, thus reducing the moisture content from 70 to 30% and making the skin unfavorable for the growth of ordinary bacteria [6].
Nevertheless, a special type of organisms, named extremophiles, can reproduce freely in extreme environments, such as high temperature, hypoxia, high pressure, low temperature and high salt [7]. Halophilic extremophiles, or simply halophiles, are microorganisms that can grow and often thrive in high salt (NaCl) concentration areas. Therefore, NaCl curing cannot inhibit the reproduction of halophiles but creates an improved breeding environment for them by inhibiting the growth of other nonhalophiles. Moreover, halophiles are a potential source of novel extremozymes such as amylase, protease, collagenase, cellulase, xylanase, esterase and lipase [8]. These extremozymes could dehair the hair, damage the collagen dermal fibers and reduce leather quality [9][10][11][12]. The extracellular halo-alkali-thermophilic protease from Halobacterium sp. strain HP25, an obligate halophile that grew optimally at 25% (w/v) NaCl, was purified by Elbanna et al. [13]. Ali et al. [14] also noted the highest protease yield at 7% (w/v) NaCl from Bacillus subtilis  [16]. However, unlike the increasing knowledge garnered from cultivable microorganisms isolated from wet-salted skins, considerably less is known about the spatiotemporal dynamics of microbial community distribution and their movement trail in NaCl-cured skins. However, this information is crucial to gain an improved understanding of microbial ecology and to reduce damage caused to the original material. Afterward, high throughput DNA sequencing, a culture-independent molecular screening technique, including 16S rRNA gene amplicon sequencing and metagenomics, has allowed the examination of a facet of microbial diversity that is not necessarily reflected by culturing study results [17]. Compared with other non-culture-based methods, the amplicon sequencing methodology with a higher sensitivity and detection ability can comprehensively identify microbial populations present in the sample. Moreover, it offers the advantage of detecting microbes that existing in a viable but nonculturable state [18]. We here first investigated the microbial diversity on wet-salted ovine skins from five different world regions, namely Australia, Spain, Inner Mongolia (China), Zhejiang (China) and Hebei (China). We then reported the spatiotemporal dynamics of the microbial diversity on NaCl-fixed fresh goatskin by using 16S rDNA amplicon sequencing. These results will be useful for further understanding the reasons for the spoilage of wet-salted skins and provide additional knowledge to obtain high-quality leather.

Site description and sample collection
Wet-salted ovine skins without no apparent damage were randomly collected from Australia, Spain, Inner Mongolia (China), Zhejiang Province (China) and Hebei Province (China). Each sample was placed in a sterile plastic bag and transported to the lab, and processed within 24 h. First, the surface of the skins was cleaned several times with sterile water and the filtrate was collected. After several long high-speed (12,000 rpm in 50 mL EP tubes) centrifugation steps, the microorganism containing sediment was obtained. Second, the surface-treated samples were shredded, beaten and washed with sterile water. The microbial deposits were collected through filtration and centrifugation at 10,000 × g for 10 min. The same methods were used to dispose the five different sample sources, and accordingly, 10 microbial sediments were obtained.
In addition, a fresh goatskin was obtained from a farm on the outskirts of Chengdu (30°67′ N, 104°07′ E), Sichuan Province, south-west China. After sampling (goatskins without salt curing treatment were marked as sample 0), the remaining skin was smeared with 40% (w/w) NaCl, and then placed at room temperature. Then, the microbial communities were harvested at 7 time points: 3, 6, 10, 15, 30, 60 and 90 days by using the aforementioned operation.

Total genomic DNA extraction
Each sediment sample was ground with liquid nitrogen, and subjected to genomic DNA extraction by using a specific DNA extraction kit (MOBIO Laboratories, Inc., USA) according to the manufacturer's protocols. The genomic DNA quality was determined through 1% (w/v) agarose gel electrophoresis in 1 × Tris-acetate-EDTA buffer, and the A260/A280 absorbance ratio was calculated to estimate DNA concentration and purity by using a Nanodrop Life Spectrophotometer (Thermo Fisher Scientific, USA).

16S rDNA V4 region amplification and sequencing
Qualified DNA samples were used as templates for polymerase chain reaction (PCR). The V4 hypervariable region of the microbial (for both bacteria and archaea) 16S ribosomal RNA gene was amplified using the specific primers 515FmodF (5'-GTG YCA GCMGCC GCG GTAA-3') [19] and 806RmodR (5'-GGA CTA CNVGGG TWT CTAAT-3') [20]. The PCR mixture (20 μL) contained FastPfu buffer, dNTPs, primers, TransStart FastPfu DNA polymerase (TransGen Biotech, China), bovine serum albumin, template DNA and ddH 2 O. All PCR reactions were conducted in an ABI GeneAmp ® 9700 thermocycler by using the following procedure: initial denaturation at 95 °C for 3 min; 28 cycles of denaturation 95 °C for 30 s, annealing at 53 °C for 30 s and extension at 72 °C for 45 s; and a final extension at 72 °C for 10 min. The PCR products were detected through 2% (w/v) agarose gel electrophoresis and purified using a QIAquick Gel Extraction Kit (QIAGEN, Germany). The collected DNA fragments were analyzed using high throughput sequencing on an Illumina Miseq platform at Majorbio Bio-Pharm Technology Co., Ltd (Shanghai, China).

Sequence denoising
The original image data file generated through highthroughput sequencing was converted into the raw sequence after base recognition by using CASAVA software, which was called raw data or raw read and stored in Fastq format. The raw reads were trimmed for barcodes and primers by using Trimmomatic [21], and chimeric sequences were removed using the Uchime algorithm  [22]. Using PEAR [23], the effective reads were merged and grouped into operational taxonomic units (OTUs) in UPARSE [24] at a 97% similarity. The raw reads have been deposited into the NCBI Sequence Read Archive database under the accession numbers PRJNA755827 and PRJNA755859.

OTU clustering, taxonomic diversity and bioinformatics analysis
To validate the microbial diversity on wet-salted skins, α-diversity indices, including Sobs, Chao, Simpson and Shannon indices were quantified in terms of OTU richness. Moreover, rarefaction curves of the observed numbers of OTUs were plotted to evaluate sample adequacy.
To remove the bias triggered by different sequencing depths, the OTU table was rarefied and an even sampling depth was established by randomly subsampling the same number of sequences from each sample. OTU taxonomy was annotated using the SILVA database [25] with the Ribosomal Database Project classifier at a 70% confidence threshold (version 2.11) [26]. Meanwhile, Veen analysis was conducted to address the commonalities and differences between samples, and to identify the species with the greatest impact on the wet-salted skin. These analyses were conducted using the 'vegan' package in R [27].

Culturable bacteria isolation and identification
Based on the results of bioinformatics analysis and the fact that wet-salted skins lose hair during preservation, the sediments collected from wet-salted goatskins were selected and spread on Luria-Bertani (LB) agar medium supplemented with 5% (w/v) NaCl and 2% (w/v) milk by using the standard dilution plating technique. This lead to the isolation and screening of protease-producing bacteria. The plates were incubated at 30 °C and examined daily, and a single colony was transferred to fresh high salinity LB plates. The bacteria were identified on the basis of morphological characteristics and results of 16S rRNA gene analysis, and the sequences obtained were uploaded to the GenBank database and their similarity with published sequences in the EzBioCloud database was determined using BLAST. The protease activity was measured using the Li protocol [10] with casein as a substrate. Some of the sequences were randomly selected from the samples, and the corresponding number of OTUs was counted. The results of multiple random samplings were used to plot dilution curves, presenting the number of sequences drawn and the corresponding number of OTUs. Under the Shannon and Sobs algorithms, the dilution curves of the 10 samples tended to be flat, thereby indicating that the sample size and sequencing data were reasonable and that the resulting sequence could reflect the actual bacterial community structure (Additional file 1: Fig. S1). Furthermore, alpha diversity analysis under the Sobs, Chao, Ace and Shannon indices showed that the microbial diversity inside the wet-salted ovine skins was usually higher than that on the skin surface (Additional file 1: Fig. S2). Because the interior of the wet-salted skin is rich in fat, protein and other nutrients, and is less affected by high NaCl concentrations, microorganisms tend to grow on the inside under salt stress, displaying signs of migration from outside to inside the skin. Therefore, microbial diversity inside the wet-salted skin is equal to or often higher than that outside the skin.

Sequence processing and assigning taxonomic identity
To profile the microbial community composition of the wet-salted goatskin, we performed 454 pyrotag sequencing of the V4 region of the 16S rRNA gene on wet-salted hide samples from 8 different times. In total, 3,381,689 high-quality sequences (~ 253 bp on average) were obtained from the 51 samples in the pyrosequencing run, with coverage ranging from 32,381 to 74,992 sequences per sample (Table 1). All effective sequences were clustered on the basis of 97% sequence identity cut-off, and  (Table 1). In addition, as shown in Additional file 1: Fig. S3, the rarefaction curves of all samples under Chao and Shannon indices tended to approach the saturation plateau. Thus, the sequencing depth basically covered all species in the samples, and the existing data could truly reflect the microbial community structure of the wet-salted skin, meeting the requirements of subsequent bioinformatics analysis.

Analysis of microbial diversity of complex bacterial communities
As observed in Fig. 1, the microbial community diversity inside and on the surface of the wet-salted goatskins changed significantly with an increase in salting time after industrial NaCl treatment, and the alpha diversity at Sobs, Chao, Ace and Shannon indexes showed an overall trend of 'down → up → down → flat' . Initially, animal skins, as an ideal habitat, gathered numerous microorganisms, such as Halomonas, Staphylococcus, Vibrio and Weissella species (Fig. 2). In the absence of environmental stress, microbes more easily accumulated on animal skin surface, resulting in a considerably higher microbial diversity outside the skins than on the inside. The high osmotic pressure formed by uniform application of large amounts of NaCl on goatskins creates a special high-salt environment, which often causes water loss from the cells of common microorganisms. This leads to cell wall separation and affects the metabolic reproduction of these microorganisms, but has no effect on halotolerant and halophilic bacteria. This results in the gradual extinction of common microorganisms such as Acinetobacter, Weissella, Streptococcus, Kocuria and Macrococcus species (Fig. 3a), and the reduction in microbial communities and diversity on the surface and inside of the skin. With the removal of salt-intolerant microorganisms, Staphylococcus, Halomonas, Salinivibrio, Salimicrobium, Chromohalobacter and other salt-loving and salt-tolerant microorganisms received a wider space for growth and more nutrient substrates (Fig. 3b). They thus proliferated and the diversity of microbial communities on the wet-salted skins rapidly increased from the 3rd day of storage and peaked on the 10th day, with the detection of over 900 species. From the 10th day, after the the salt curing time was extended, water loss on the wet-salted skin was considerable, and the salt concentration further increased beyond the tolerance range of slight halophile, moderate halophile and halotolerant bacteria. This thus inhibited the reproduction of all microorganisms except extreme halophile, and gradually decreased the microbial community diversity both inside and outside the wet-salted skin. Ultimately, the wet-salted skin with high salt, little water and many nutrient substrates became an excellent habitat for extremophilic prokaryotic microorganisms such as Alkalibacillus, Halovibrio, Natrialba and Haloterrigena strains (Fig. 3c). On the other hand, they maintained a low microbial community diversity on the inside and outside. Furthermore, from the 10th day of salt curing, the microbial diversity (Fig. 1) and the content of Alkalibacillus, Halovibrio and Natrialba strains were higher inside the skin than on the skin surface (Fig. 3c). This indicated that some wet-salted skin microorganisms would migrate from the skin surface to the interior under salt stress. This is consistent with the results of research about microorganisms on wet-salted ovine skins from five different regions (Additional file 1: Fig. S2). More importantly, the surface and inside of the wet-salted skin was almost devoid of archaea in the early salting stage. However, when the salting time was extended, the percentage of archaea on the inside and outside of the skins increased, and gradually, they became the dominant prokaryotic microorganisms. The relative abundance of archaea on the inside and outside of the skins was as high as 28.58 and 48.50% respectively on the 90th day of salting (Fig. 2). This indicated the gradual transition from bacteria to archaea on the wet-salted skins with the extended salt curing time.

Correlation analysis of microorganisms in wet-salted skin
Many environmental factors affect the microbial community composition of wet-salted skins, but the most influential environmental factor is time (T). Therefore, in terms of relative abundance, the top 50 OTUs were selected at the genus taxonomic level and correlation network plots were drawn using the correlation coefficient type Spearman with the environmental factor T. As shown in Fig. 4a, extremophilic prokaryotic microorganisms, represented by species of the genera Alkalibacillus, Natrialba and Haloterrigena, showed a better positive correlation with time, the longer the salt curing time, the more ideal conditions for their survival and the higher their abundance. By contrast, non-halophile represented by members of the genera Lysinibacillus, Acinetobacter and Streptococcus displayed a negative correlation with time; the longer the salting time, the more severely they were inhibited and the lower their abundance on the wetsalted skin.
To reveal the affinities of microorganisms on wet-salted skins at the genus level from the perspective of molecular evolution, the OTU data corresponding to the top 40 genera in terms of relative abundance were subjected to random sampling and multiple sequence comparisons. The phylogenetic evolutionary tree was constructed using the maximum parsimony (MP) method [22]. As seen in Fig. 4b, 40 genera distributed in 26 families formed a good topology in the MP evolutionary tree. The non-halophile genera Acinetobacter and Psychrobacter are grouped together in the family Moraxellaceae, the moderate halophile genera Halomonas, Chromohalobacter and Halovibrio are grouped together in the family Halomonadaceae, and the archaeal genera Halococcus, Halorubrum, Natrialba and Haloterrigena form a large independent branch. These facts suggest that certain affinities exist among microorganisms on wet-salted skins and the species belonging to the same family taxonomic level may have similar salt tolerance and halophilicity.
To visualize the similarity and overlap of microbial species composition of wet-salted skins at different time, a petal map of all samples was constructed using R language based on the species classification at the genus level. Figure 4c reveals that the wet-salted goatskin contained the most species on the 10th day of storage, with 587 and 571 genera on the inside and outside of the hide, respectively. In addition, 17 groups of microbial samples shared 11 genera, including Salinicola, Staphylococcus, Luteimonas, Psychrobacter, Salimicrobium, Ornithinibacillus, Halomonas, Bacillus, Brevundimonas, Chromohalobacter, and unclassified genera of the family Burkholderiaceae. This indicated that strains belonging to these genera were present throughout wet-salted skin preservation and distributed throughout the interior and exterior of the skin. Some species of these genera have been reported to produce pigments, lipases, proteases and collagenases [10,[29][30][31][32]. Thus, their growth and metabolism have been hypothesized to lead to erythema, desquamation and rotting of wet-salted skins.

Screening and preliminary identification of protease-producing bacteria
The aforementioned results of bioinformatics analysis revealed that numerous microorganisms were accumulated on the wet-salted goatskin, especially the halophilic bacteria belonging to genera such as Salinicola, Ornithinibacillus, Alkalibacillus, Chromohalobacter, Halomonas and Natrialba. They resulted in a very high diversity of wet-salted skin microorganisms. In high-salt environments, halophiles have evolved unique physiological structures and metabolic pathways to synthesize active substances such as tetrahydropyrimidines, polysaccharides and proteases. Of them, proteases are the main substances responsible for localized hair loss or hair loosening in wet-salted goatskins (Additional file 1: Fig. S4a). Accordingly, 45 strains of protease-producing bacteria were screened and isolated from the microbial samples of the wet-salted goatskin (Additional file 1: Fig. S4b), some of which could also produce white, yellow, orange and red pigments on modified LB plates (Additional file 1: Fig. S4c). The 16S rRNA genes of protease-producing bacteria were amplified, sequenced, registered in the NCBI database and searched in the EzBioCloud database. As listed in Additional file 1:

Conclusions
16S amplicon sequencing and bioinformatics analysis revealed that microbial diversity on the wet-salted goatskin showed a trend of 'down → up → down → flat' over time. To resist high salt stress, the microbial community changed from bacteria to archaea, and migrated from outside to inside the skin. Moreover, traces of species belonging to the genera Staphylococcus, Luteimonas, Psychrobacter, Salimicrobium, Salinicola, Halomonas, Bacillus, and Ornithinibacillus were found at all times during wet-salted goatskin preservation. Based on the results of bioinformatics analysis of wet-salted goatskin microorganisms and the fact that the goatskin showed local hair loss during storage, 45 strains of halophilic or halotolerant microorganisms secreting extracellular active proteases were isolated and screened, among which strains L3, L5 and L9 were suspected to be new species. These results fill the gap in the study of the microbial community structure of wet-salted skins and provide a theoretical basis for the further improvement and development of methods for high-quality preservation of animal skins.
Additional file 1. Table S1. Comparison results of 45 protease-producing strains in EzBioCloud database. Fig. S1. Rarefaction curves of 16S rRNA gene sequencing of the wet-salted ovine skins from five areas. a: the rarefaction curve was assembled using Shannon index; b: the rarefaction curve was executed using Sobs algorithm. A, S, I, Z and H represent Australia, Spain, Inner Mongolia (China), Zhejiang Province (China) and Hebei Province (China), respectively. SB and IB indicate the surface and the inside of the wet-salted ovine skins, respectively. Fig. S2. Alpha diversity analysis of microorganisms present in the interior and exterior of wetsalted ovine skins from five different areas at indices Sobs, Shannon, Chao and Simpson. A, S, I, Z and H represent Australia, Spain, Inner Mongolia (China), Zhejiang Province (China) and Hebei Province (China), respectively. SB and IB indicate the surface and the inside of the wet-salted ovine skins, respectively. Fig. S3. Dilution curves of Chao (a) and Shannon (b) indices for microbial communities in homemade salt-preserved goatskin. Fig. S4. Screening and characterization of protease-producing bacteria: a, localized dehairing of wet-salted goatskin; b, milk plate hydrolysis circle screening bacteria; c, characterization of protease-producing bacteria on the modified LB plates.