Sample Collection, DNA Extraction, and Library Construction Protocols of the Human Microbiome Studies in the International Human Phenome Project

Abstract

The human microbiome plays a crucial role in human health. In the past decade, advances in high-throughput sequencing technologies and analytical software have significantly improved our knowledge of the human microbiome. However, most studies concerning the human microbiome did not provide repeatable protocols to guide the sample collection, handling, and processing procedures, which impedes obtaining valid and timely microbial taxonomic and functional results. This protocol provides detailed operation methods of human microbial sample collection, DNA extraction, and library construction for both the amplicon sequencing-based measurements of the microbial samples from the human nasal cavity, oral cavity, and skin, as well as the shotgun metagenomic sequencing-based measurements of the human stool samples among adult participants. This study intends to develop practical procedure standards to improve the reproducibility of microbiota profiling of human samples.

Supplementary Information

The online version contains supplementary material available at 10.1007/s43657-023-00097-y.

Introduction

The human microbiome, defined as the sum of all microorganisms on and inside the human body (Docter et al. 2015), is directly related to human health and its related phenotypes. For example, the human gastrointestinal microbiome and the microbial metabolites are of great significance in the development of gastrointestinal disorders (Chen et al. 2021), type 2 diabetes (Herrema and Niess 2020), cardiovascular disease (Witkowski et al. 2020), and autoimmune disease (Mariño et al. 2017). Specifically, the primary metabolites produced by bacterial fermentation of dietary fiber in the gastrointestinal tract—short-chain fatty acids (SCFAs)—are speculated to have an influence on systemic inflammation and central nervous function (Dalile et al. 2019).

In the microbiome research area, each laboratory decides the appropriate protocols that fit its practical situation, including the experimental design, laboratory space, staff, and financial cost. However, the bias of sample collection, DNA extraction, and library construction possibly introduced into laboratory workflow may obscure the actual composition of a microbial community, resulting in analysis deviation (Tourlousse et al. 2021). A series of false or improper technical operations may lead to inevitable and non-repeatable variations of microbiome characteristics. The microbial composition of stool samples is highly susceptible to storage-temperature fluctuations and freeze–thawing (Choo et al. 2015). Variation also occurs in the DNA extraction of aliquots from the same sample at different storage periods. Because the lysis of Gram-positive bacteria (Bruslind 2018) and eukaryotic flora in the intestine are difficult, the DNA extraction efficiency of these microbes may directly affect the microbial community composition (Hugerth and Andersson 2017). As no universal primer pair can target all bacteria (Fredriksson et al. 2013), the choice of the primer and the amplification region results in further variations during polymerase chain reaction (PCR) amplification in the total microbial diversity (Darwish et al. 2021; Johnson et al. 2019).

Here, we propose practical operation protocols for sample collection, DNA extraction, and library construction based on microbial samples from the human nasal cavity, oral cavity, skin, and stool of healthy adult participants in the “International Human Phenome Project (phase I)”.

Materials and Equipments

Materials for Sample Collection

Disposable sterile iCleanhcy® Specimen Collection Swabs (CY-98000, HCY Technology, China); Screw Cap Microtubes (B80180, BOOPU Biotechnology, China); 5 mL Funnel-type saliva sample collection tube (LF005-5, BIORISE, China); In-house stool collection kit including a pair of disposable gloves (Yangzi Lide Medical Equipment, China), two 25 mL labeled collection tubes with plastic scoop embedded into the screw cap (LF0025, BIORISE, China), one feces catcher designed for pets (Jujukong Pet Food, China), one paper bowl (TH7661, Tinghao, China), two wooden sticks (Shangji, China), and a plastic sealed bag; Sterile specimen collection fluid-1 (SCF-1) solution (0.15 M NaCl and 0.1% Tween 20).

Materials for DNA Extraction and Library Construction

DNeasy 96 PowerSoil Pro QIAcube HT Kit (47021, Qiagen, Germany); Buffer ATL (939011, Qiagen, Germany); QIAcube HT Plasticware (950067, Qiagen, Germany); Premix Ex Taq DNA Polymerase (RR003A, TaKaRa Biotechnology, Japan); DNase/RNase free H₂O (4992956, TIANGEN BIOTECH, China); AxyPrep DNA Gel Extraction Kit (AP-GX-50, Axygen, USA); KAPA LTP Library Preparation Kit (KK8233, Roche, Switzerland); PKR Y Type Adapter Kit (PKR2015/ PKR2016/ PKR2017/ PKR2018, Pukairui, China); DNA Library Prep Kit (K1255, APExBIO, China); VAHTS® DNA Clean Beads (N411-01, Vazyme, China).

Experimental Equipments

TissueLyser II (Qiagen, Germany); TissueLyser Adapter Set 2 × 24 (Qiagen, Germany); QIAcube HT Instrument (Qiagen, Germany); Qubit 4 Fluorometer (Thermo Fisher Scientific, USA); ETC811 PCR Thermocycler (EASTWIN, China); BG-Power600 Electrophoresis Instrument (BAYGENE, China); FR-980 Bioelectrophoresis Image Analysis System (Furi Technology, China); Biomek i5 Automated Workstation (Beckman Coulter, USA); Mastercycler@X50S PCR Thermocycler (Eppendorf, Germany); Agilent Bioanalyzer 4200 System (Agilent Technologies, USA).

Procedures

Sample Collection and Transfer

Collection and Transfer of the Nasal Cavity and Skin Microbial Samples

When sampling, the staff needs to wear a pair of gloves and a mask. Insert a sterile iCleanhcy® specimen collection swab into one nostril for about 2.5 cm and gently rotate the swab over the mucosal surfaces of the anterior nostrils two times (lasting 5 s in total). Insert the swab head into the tube, aseptically cut from the handle higher than the tube, and screw the tube cap back to its original place. Follow the same procedures with a new swab on the other nostril, combine the two swabs into one specimen, and put the tube into a 4 °C refrigerator immediately.

Skin microbiome's collection follows the sequence from the scalp, forearm, back, to the forehead (4 positions). Specific sites of skin microbiome's sampling are illustrated in Fig. 1. Skin surface specimen is also collected with a sterile iCleanhcy® specimen collection swab. Moisten the swab with sterile SCF-1 solution by dipping its tip into the solution, and stick the swab to the upper inside wall of the SCF-1 solution tube to squeeze out the excessive solution. Hold the swab shaft parallel to the surface of the skin and rub it back and forth approximately 30–50 times under firm pressure for about 30 s. Keep sampling uniform in the selected area. Insert the swab head into the tube, aseptically cut from the handle higher than the tube, and screw the tube cap back in place. Put the tube into a 4 °C refrigerator immediately. Then follow the same operation in the next position.

Fig. 1.

The locations of the sampling sites in nasal cavity and skin

After a batch of specimens is collected, transfer them into the standard sample boxes, organize them following the standard operating procedure for biobank management, and finally store them at − 80 °C.

CAUTION: Do not use the same position for both physical & chemical testing and microbial sampling: for forearms, forehead, and back, select one side to conduct physical & chemical testing and the other for microbial sampling. Skincare or cosmetic products should not be used on the skin sampling sites within 24 h before sampling. In addition, washing is not allowed for the skin sampling sites within 8 h before sampling.

Collection and Transfer of the Saliva Sample

The sample from the oral cavity here refers to saliva. Subjects rub the inside of the cheek with the tongue and spit carefully into the funnel-type collection tube (5 mL) until the saliva reach the fill line (2 mL, without foam). The staff needs to tighten the tube immediately and put it into a 4 °C refrigerator.

After a batch of specimens is collected, the staff aliquot per tube into two 2 mL microtubes (for 1 mL per tube). Then, transfer them into the standard sample boxes, organize them following the standard operating procedure for biobank management, and finally store them at − 80 °C.

CAUTION: The saliva samples collected shall be free of sundries and sputum. Subjects are not allowed to eat, drink, smoke, or chew gum within 30 min before sampling.

Collection and Transfer of the Stool Sample

An in-house stool collection kit is provided for subjects. Subjects defecate directly into the paper bowl covered with the feces catcher. (Make sure no urine, water, or other material gets in the bowl.) Use the plastic scoop embedded into the screw cap (or the wooden sticks) to collect a spoon of midstream stool (thumbnail-cover size), place the stool sample into the collection tube, and screw tight to secure the lid. Each tube is used to sample only once. Whenever two tubes are collected, put them together into the sealed bag, place it into a refrigerator (4 °C from 5:30 a.m. to 4:50 p.m. or − 20 °C from 4:50 p.m. to 5:30 a.m.), and inform the staff (see detailed instructions for the stool collection kit in supplemental materials). Most stool samples (> 80%) were collected during 8:30 a.m. to 4:50 p.m.

Within 30 min after the staff is informed, the samples collected should be aliquoted into four frozen storage tubes and then stored at − 80 °C until used.

CAUTION: Day-shift staff collect the collection tubes that contain the remaining samples and perform sterilization.

DNA Extraction and Library Construction

DNA Extraction and Amplicon Sequencing of Nasal, Skin and Saliva Samples

According to the experimental protocol, DNA is extracted using the DNeasy 96 PowerSoil Pro QIAcube HT Kit and Buffer ATL on the QIAcube HT instrument. The TissueLyser II and TissueLyser Adapter Set 2 × 24 are used for two times of 5-minute shaking (10 min in total) at the frequency of 25 Hz (A reorientation is needed before another 5-minute shaking). The Qubit 4 Fluorometer determines DNA yield.

All primers used in this study are from Sangon Biotech (China). As a marker, hypervariable region V4 of the 16S ribosomal RNA (rRNA) gene is used to classify the bacterial taxonomy composition of specimens from the nasal cavity, oral cavity, and skin. Barcoded forward primer 515F (5′-GTGYCAGCMGCCGCGGTAA-3′) (Apprill et al. 2015) and reverse primer 806R (5′-GGACTACNVGGGTWTCTAAT-3′) (Parada et al. 2016) are used to amplify 16S rRNA gene V4 region. The individual 50 μL reaction system is described in Table 1. Specific procedures for the 16S-V4 PCR system are described in Table 2.

Table 1.

16S-V4 (ITS1) PCR system

Composition	Adding volume (μL)
Premix Ex Taq DNA Polymerase	25
template DNA	10
forward primer (10 μmol/L)	1
reverse primer (10 μmol/L)	1
DNase/RNase free H2O	13
Total	50

Table 2.

Specific procedures for 16S-V4 PCR system

Temperature (°C)	Time (s)	Cycles
94	300	1
94	30
50	30	35
72	30
94	300	1
4	∞

The intervening internal transcribed spacer 1 (ITS1) region is used for fungal taxonomy classification. Barcoded forward primer 18S-F (5′-GTAAAAGTCGTAACAAGGTTTC-3′) (Findley et al. 2013) and reverse primer 5.8S-1R (5′-GTTCAAAGAYTCGATGATTCAC-3′) (Findley et al. 2013) are used to amplify ITS1. The individual reaction system is consistent with bacterial amplification (Table 1). Specific procedures for the ITS1 PCR system are described in Table 3.

Table 3.

Specific procedures for ITS1 PCR system

Temperature (°C)	Time (s)	Cycles
94	300	1
94	30
57	30	35
72	30
94	300	1
4	∞

Productions are pooled by 48 reactions per mix and analyzed on 1.5% (w/v) agarose gel electrophoresis. Target bands are extracted using AxyPrep DNA Gel Extraction Kit. According to official instructions, preparation of Illumina sequencing library for pooled amplicon is performed with KAPA LTP Library Preparation Kit and PKR Y Type Adapter Kit. In brief, procedures are as follows: end repairing and cleanup, a-tailing and cleanup, Illumina sequencing Truseq adapter ligation, post-ligation cleanup, amplification for 20 PCR cycles, gel electrophoresis analysis, and target bands extraction. Libraries can be transported for paired-end sequencing using the Illumina Novaseq 6000 PE250 sequencing platform. Sequencing depth approaches 10 M reads per library consisting of 48 samples.

Sequences of 16S rRNA gene or ITS amplicons are processed using the Quantitative Insights Into Microbial Ecology 2 (QIIME2) program (Bolyen et al. 2019). Briefly, raw fastq files are split according to the barcode sequence and converted into QIIME2 compatible files. For ITS amplicons, the conserved regions around the ITS amplicons are first trimmed using the ITSxpress (Rivers et al. 2018) plugin of QIIME2. Using Divisive Amplicon Denoising Algorithm 2 (DADA2) (Callahan et al. 2016) plugin of QIIME2 to perform the denoising and calling of amplicon sequence variants (ASVs). The truncation parameters for DADA2 in amplicons of 16S rRNA gene are estimated using FIGARO software (Weinstein et al. 2019); parameters of quality scores and expected number of errors are evaluated in the analysis of ITS amplicons considering their significant differences in length — a hard trim may not be proper. Taxonomic assignments are made using the UNITE database (Kõljalg et al. 2005) for ITS sequences and using the SILVA database (Quast et al. 2013) for 16S rRNA gene sequences via plugins of feature-classifier (Bokulich et al. 2018) and classify-sklearn (Pedregosa et al. 2011).

CAUTION: All experimental materials are sterile and UV-irradiated. The whole process of skin sample handling is carried out in the vertical flow clean bench. After each experiment, the clean bench is ventilated and exposed to UV irradiation. In addition, the PCR amplification products of negative control using the sterile water as templates are checked with gel electrophoresis. The PCR products collected are pooled together for sequencing. Based on the sequencing data of negative control, the major contaminating bacterial genus in our study was Sphingomonas (data not shown), which is consistent with previous studies (Salter et al. 2014; Stinson et al. 2019).

Skin microbial samples may also be analyzed through shotgun metagenomic sequencing. Nevertheless, researchers should check whether each sample's biomass meets the requirement of library construction for metagenomic sequencing.

We suggest applying the same primer set to a comparative analysis of different body sites. In our study, V4 region is chosen. If only skin sites bacteria are to be studied, we suggest using primer 27F (5′-AGAGTTTGATCMTGGCTCAG-3′) (Weisburg et al. 1991) and 338R (5′-TGCTGCCTCCCGTAGGAGT-3′) (Amann et al. 1995). For more details on the experimental process, please refer to the previous publication (Zhu et al. 2019).

DNA Extraction and Shotgun Metagenomics Sequencing of Stool Sample

As the protocol instructs, the DNA extraction of stool samples is consistent with that of nasal and skin swabs and saliva samples.

According to the experimental protocol (supplemental materials), library preparation for shotgun metagenomics sequencing is performed with the DNA Library Prep Kit on the Biomek i5 automated workstation. In brief, DNA fragmentation and end-linking are performed by incubating Transposase Mix with input DNA at 55 °C for 10 min. The tagged fragments can be further amplified with N5 (N5XX) and N7 (N7XX), and the amplified products need to undergo size selection and purification. A Qubit Flex fluorometer and the Agilent Bioanalyzer 4200 system are applied to confirm the distributions of the expected insert size. Metagenome sequencing is performed using the Illumina Novaseq 6000 PE250 sequencing platform. Each sample may generate approximately 10 Gb of 150 bp paired-end reads.

Pilot Studies for the Preparation of the Protocol

Choice of Swabs or Strip Tapes and Amplification Cycle Number

In our pilot study of four subjects, four types of sterile iCleanhcy® specimen collection swabs moistened with SCF-1 solution, together with two types of strip tapes, were used to collect microbiome samples from the scalp, forehead, and forearm skin surfaces. After DNA extraction, based on the recommendations of previous publications (Salter et al. 2014), the target DNA segments were amplified after 30 or 35 PCR cycles. However, the target bands from some subjects were weak after 35 PCR cycles (e.g., columns d and e in Subjects 3 and 4 in Fig. 2a) while did not appear after 30 PCR cycles (e.g., Subjects 3 and 4 in Fig. 2b). Because the sampling method of strip tapes was more involved with manual operation (may cause contamination easily) and its less favorable performance in PCR products, swabs were chosen over strip tapes. In addition, due to the brightest target bands produced with the iCleanhcy® specimen (CY-98000) after 35 PCR cycles in most of the products (column e in Fig. 2a), we decided to use iCleanhcy® specimen (CY-98000) collection swabs for the collection of skin microbiome samples and apply 35 PCR cycles in the amplification procedure.

Fig. 2.

The agarose gel electrophoresis of PCR products by 35 cycles (a) and 30 cycles (b) for four subjects. Columns “a” to “f” represent the skin microbiome templates collected by Cuderm D-Squame sampling disks D100 (a) and D301 (b), iCleanhcy® specimen swabs CY-93050 (c), CY95000 (d), and CY98000 (e), SOCO sterilized polyester swab SC718A (f), respectively. Column “n” represents the negative control of sterile water template (The gel image of subjects 1–3 and that of subject 4 were from two different gels)

Choice of Moistening of the Swabs

Several types of un-pre-moistened sterile iCleanhcy® specimen collection swabs were applied to collect microbiome samples from one subject's scalp, forehead, and forearm skin. An absence of target bands observed on agarose gel in all PCR products of the 16S rRNA gene indicated the necessity of pre-moistening of the swabs. However, excessive solution concentration in swabs might also decrease the efficiency of the process. Therefore, after dipping the swab's tip into the solution, we stuck the swab to the upper inside wall of the SCF-1 solution tube to squeeze out the excessive solution.

Choice of DNA Extraction Protocols

In this pilot study, nine protocols were applied to extract DNA from stool, saliva, and mock microbial communities, namely, QIAamp BiOstic Bacteremia DNA Kit; QIAamp Fast DNA Stool Mini Kit; Magnetic Soil and Stool DNA Kit; Roche MagNA Pure 24 Total NA Isolation Kit; QIAamp PowerFecal Pro DNA Kit; semi-automated DNeasy PowerSoil HTP 96 Kit; TIANamp Stool DNA Kit; ZymoBIOMICS DNA Miniprep Kit (ZYMO); and non-commercial protocol Q recommended by the International Human Microbiome Standard (Dore et al. 2015) with some revisions (Costea et al. 2017).

DNeasy PowerSoil HTP 96 Kit with bead-beating step had the greatest performance in terms of DNA yield, purity, stability, the least amount of bias in positive mock community controls, the efficiency of extracting DNA from Gram-positive bacteria, and the ability to scale up extraction pipelines to many types of samples using the matched automated extraction machine (Pu et al. manuscript in progress).

Choice of PCR Polymerases, Barcodes, Sequencing Company and Check of Contamination

In this pilot study, three commonly used enzyme kits were tested for with 35-cycle PCR processes, i.e., Premix Ex Taq DNA Polymerase (TaKaRa Biotechnology, Japan), KAPA HiFi HotStart DNA Polymerase (Roche, Switzerland), and VAHTS HiFi Amplification DNA Polymerase (Vazyme, China). The above confirmed specific sterile iCleanhcy® specimen collection swab was applied to scalp and forearm skin microbiome sampling. All PCR products were checked on agarose gel electrophoresis. For VAHTS HiFi Amplification DNA Polymerase, an absence of target bands observed during the amplification of forearm skin samples indicated its poor amplification ability. While for KAPA HiFi HotStart DNA Polymerase and Premix Ex Taq DNA Polymerase, the sequencing data analysis revealed no major difference in their application to duplicate skin microbiome samples (Fig. 3). Two different barcodes were added to the forward primer to amplify the duplicate skin microbiome samples with each type of polymerase, and no significant differences were observed between the results from these two barcodes (Fig. 3). Simultaneously, no apparent differences were observed between the sequencing data of the duplicate library generated by the two sequencing companies (Fig. 3).

Fig. 3.

The composition of top 20 bacterial taxa amplified by three types of PCR polymerases with different barcodes for forehead skin samples (a, b) and forearm samples (c, d). The samples in a and c were sequenced by Novogene Co., Ltd. (China), while those in b and d were sequenced by GENEWIZ, Inc (China)

To further examine the kit contaminants from KAPA HiFi HotStart DNA Polymerase and Premix Ex Taq DNA Polymerase, we selected one strain in the genus Limimaricola isolated from Mariana Trench, usually undetected in the human microbiome, as a positive control. We established a pure culture of this strain on agar plates, checked and confirmed the purity of this positive control using 16S rRNA gene sequencing. Serial dilution of this strain was conducted, and DNA extraction and amplification of the 16S rRNA gene of these diluents were performed with these two enzyme kits separately following the aforementioned procedures. The sample with the highest dilution fold (~500 cells) presenting a clear target gel band was selected for sequencing, to maximize the difference in the relative abundance of the contaminants between these two kits. The sequencing data revealed a higher percentage of the other bacteria besides Limimaricola in samples amplified using the KAPA HiFi HotStart DNA Polymerase kit than that using the Premix Ex Taq DNA polymerase kit (Fig. 4). This observation suggested a higher presence of contaminants (such as Anarobacillus and Acidovorax) in the KAPA HiFi HotStart DNA Polymerase kit. Therefore, Premix Ex Taq DNA Polymerase was selected as the suitable PCR enzyme kit in the 16S amplification procedure. Among the amplicon sequences using Premix Ex Taq DNA polymerase, Sphingomonas was one of the main contaminants (Fig. 4). This observation was consistent with the sequencing results of negative control in our study protocol using Premix Ex Taq DNA polymerase as mentioned in the CAUTION of “DNA extraction and amplicon sequencing of nasal, skin and saliva samples” section.

Fig. 4.

The relative abundance of the top 20 bacterial genera amplified by the two candidate PCR polymerases kits of the diluted samples of Limimaricola under the protocol for the skin microbiome procedures

Discussion

This study is based on human microbial samples from the nasal cavity, oral cavity, skin, and stool, and provides a detailed operation guide covering the process of managing microbiome samples from sample collection, transportation, and preservation to DNA extraction and library construction.

This current study is aimed to provide a producible protocol of sample handling for future studies in the human microbiome, and there are limitations in the pilot studies. For example, variations caused by operators were not assessed, and operators' technical stability is essential to the success of the whole experimental process. Operators should be relatively fixed, fully trained, and consistently monitor the quality control process. In addition, the pilot studies were conducted for in-house assessments and the included sample sizes were too small to allow for a formal statistical analysis. A continuous update and optimization of the aforementioned procedure standards are needed to further reduce experimental variations and improve the stability and predictability of the operation process.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

SCFAs

Short-chain fatty acids

PCR

Polymerase chain reaction

SCF-1

Specimen collection fluid-1

rRNA

Ribosomal RNA

ITS

Internal transcribed spacer

QIIME2

Quantitative insights into microbial ecology 2

DADA2

Divisive Amplicon Denoising Algorithm 2

ASVs

Amplicon sequence variants

Authors' Contributions

All authors contributed to the study conception and design. YTW, RYZ, YNP, YJP, CL, YRW, XMW and DQW completed the whole procedures of sample collection, DNA extraction, library construction and pilot studies of the protocol. Data collection and analysis were performed by DQW and YRW. The first draft of the manuscript was written by YTW and all authors commented on previous versions of the manuscript. YZ, ZXQ and GPZ provided insightful suggestions to improve the framework and enrich the content of the manuscript. All authors read and approved the final manuscript.

Funding

This study was supported by the National Key Research and Development Program of China (2021YFA1301000) and the Shanghai Municipal Science and Technology Major Project (2017SHZDZX01).

Data Availability Statements

Raw sequence data are available in the National Omics Data Encyclopedia under the project ID OEP003565 (https://www.biosino.org/node/project/detail/OEP003565), with sample IDs OES23343-OES23362 (for the skin samples in Fig. 3) and OES23395-OES23397 (for the pure strain sample in Fig. 4).

Declarations

Ethics approval

This study was approved by the Ethics Committees of the School of Life Sciences and Zhongshan Hospital, Fudan University, Shanghai, China, and carried out following the principles of the Declaration of Helsinki.

Consent to participate

Written informed consent was obtained from each participant before enrollment.

Consent for publication

Not applicable.