| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Dermatology, School of Medicine, Keio University, Shinjuku-ku, Tokyo 160-8582, Japan
2 Microbe Division/Japan Collection of Microorganisms, RIKEN BioResource Center, Wako, Saitama 351-0198, Japan
3 Department of Biochemistry and Integrative Medical Biology, School of Medicine, Keio University, Shinjuku-ku, Tokyo 160-8582, Japan
Correspondence
Itaru Dekio
dekio{at}1999.jukuin.keio.ac.jp
Received 5 March 2007
Accepted 26 August 2007
Abbreviations: AD, atopic dermatitis; T-RF, terminal restriction fragment; T-RFLP, terminal RFLP.
Present address: Faculty of Engineering, Maebashi Institute of Technology, Kamisatori-Machi, Maebashi, Gunma 371-0816, Japan. 
The GenBank/EMBL/DDBJ accession numbers for the 16S rRNA gene sequences obtained from clone libraries are AB264551–AB264556.
 |
INTRODUCTION |
|---|
 TOP INTRODUCTION METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In contrast, molecular techniques allow us to detect a variety of unidentified micro-organisms in human organs, such as the oral cavity, gut, prostate and vagina. These methods enable us to detect bacteria that are difficult to culture and isolate because of technical difficulties. The gene used is a small subunit rRNA-encoding gene known to possess species-specific evolutionary variation of the gene structure (Nelson et al., 2000). This RNA-encoding gene is present in all known bacteria and its conserved region is suitable for amplification; thus, it is convenient for identification purposes (Fredricks, 2001). For the first time, we have reported that clone libraries of the 16S rRNA gene enable us to detect the total profile of bacterial members in skin microbiota, which include uncultured bacteria (Dekio et al., 2005). The data indicated that up to 13 novel ‘phylotypes’ (yet-to-be-cultured micro-organisms) and 9 novel species exist in the facial skin of healthy humans, in addition to the 10 previously recognized species. A similar molecular analysis from the forearm skin of healthy humans detected 30 phylotypes (Gao et al., 2007), indicating that such molecular methods will inevitably be required to profile microbiota at various sites of the human skin. However, attempts to characterize the profile of microbiota in skin diseases using molecular methods have not yet been reported.
Bacterial 16S rRNA gene-based terminal RFLP (T-RFLP) is one of the molecular methods able to visualize complex microbial communities (Liu et al., 1997; Sakamoto et al., 2003). This molecular approach, which detects a fluorescently labelled fragment of the PCR-amplified 16S rRNA gene, has enabled rapid analysis of the total composition of bacterial samples that include a variety of micro-organisms. We applied T-RFLP analysis to swab-scrubbed skin samples of AD patients and normal subjects to compare the bacterial communities present in the two groups.
|
METHODS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The AD group included outpatients attending the dermatological clinic of Keio University Hospital and were diagnosed based on the criteria of Hanifin & Rajka (1980). At the time of sampling, they continued to receive the same treatment for more than 2 weeks. The AD severity scores were calculated using the SCORAD scoring system (European Task Force on Atopic Dermatitis, 1993). Patients with scores of 
40 were defined as having severe AD, those with scores of <40 but 15 as moderate AD, and those with scores of <15 as mild AD. The healthy subjects were requested to complete a questionnaire to confirm that they had no history of skin diseases, allergic rhinitis or bronchial asthma. The participants of both groups did not use medicated soap or cosmetics in their daily life. They were advised not to wash, touch or apply cosmetics to the forehead skin (the sampling area) for at least 12 h prior to commencement of the study.
|
Collection of samples. The procedure was performed as described previously (Dekio et al., 2005). Briefly, the open end of a 4.9 cm2 sterile plastic cylinder was manually placed on the forehead skin. The skin within the enclosed area was scrubbed using a sterile swab moistened with dilution liquid. The tip of the swab was then broken against the wall of a glass tube containing 1 ml dilution liquid, and the tube was immediately capped and shaken to suspend the bacteria. In addition, four control samples were prepared by scrubbing only the inner wall of the cylinder with moistened swabs. A volume of 300 µl suspension was used for culture analysis, and a volume of 250 µl was used for T-RFLP analysis.
Culture analysis. Culture analysis was performed as described previously (Dekio et al., 2005). Briefly, 10–1, 10–3 and 10–5 dilutions of the samples were plated on three non-selective agar plates: trypticase soybean (TS) agar supplemented with 5 % horse blood, Eggerth–Gagnon (EG) agar and glucose-blood-liver (BL) agar. Samples plated on TS agar were cultured aerobically at 37 °C for 2–3 days, whilst those plated on EG and BL agar were cultured anaerobically at 37 °C for 7 days. On each agar plate, the cell morphology was examined microscopically by Gram staining, and colonies with comparable macroscopic features were counted to calculate the number of c.f.u. ml–1. The colonies were subcultured, the bacterial DNA extracted and their 16S rRNA genes PCR amplified using the universal primers 27F and 1492R (Lane, 1991). The anterior one-third of the 16S rRNA genes (Escherichia coli nt 27–520) was sequenced in both directions using the 27F and 520R primers (Lane, 1991). The sequences were analysed using a BLASTN search (http://www.ddbj.nig.ac.jp; Altschul et al., 1997) for identification.
DNA extraction for T-RFLP analysis. The suspension collected for T-RFLP analysis was diluted by adding 800 µl phosphate buffer solution and centrifuged at 14 000 g for 5 min to obtain a pellet. The pellet was suspended in 450 µl extraction buffer [100 mM Tris/HCl (pH 7.6), 40 mM EDTA]; subsequently, 50 µl SDS (final concentration 1 %, w/v), 500 µl phenol and 300 mg glass beads (0.1 mm diameter) were added. The cells were degraded under constant shaking using a FastPrep instrument (Bio 101) for 30 s. The mixture was extracted with equal volumes of phenol and phenol : chloroform : isoamyl alcohol (25 : 24 : 1). The lysate was subsequently purified and condensed into 20 µl distilled water using a High Pure PCR template preparation kit (Roche) for amplification of the 16S rRNA gene coding region.
Generation of T-RFLP patterns. The procedure was performed as described by Sakamoto et al. (2004) with some modifications. Two universal primers, 6-carboxyfluorescein (6-FAM)-labelled-27F (5'-6-FAM-AGAGTTTGATCCTGGCTCAG-3') and 1492R (5'-GGTTACCTTGTTACGACTT-3') (Lane, 1991), were used to amplify the bacterial 16S rRNA gene coding region. Amplification reactions were performed in a total volume of 100 µl using the following program: 1 cycle of 95 °C for 3 min, followed by 40 cycles of 95 °C for 30 s, 50 °C for 30 s and 72 °C for 1.5 min, with a final extension period of 72 °C for 10 min. The size of the amplified DNA was verified on a gel and purified using PEG solution.
The purified PCR product (2 µl) was digested with 20 U HhaI or MspI (TaKaRa) in a total volume of 10 µl at 37 °C for 3 h. The length of the terminal restriction fragments (T-RFs) was determined on an ABI PRISM 3100 genetic analyzer (Applied Biosystems) in the GeneScan mode (15 kV, 100 µA and 60 °C for 40 min for each sample) (Fig. 1). The fragment sizes were estimated using the Local Southern method of the GeneScan 3.7 software (Applied Biosystems). T-RFs with a peak height of less than 60 fluorescent units were excluded from the analysis, and the percentage of each peak area compared with the total area was calculated as described by Matsumoto & Benno (2007). The predicted T-RFLP patterns of the 16S rRNA genes of known bacterial species were obtained using the GENETYX-MAC program (Software Development).
|
The term phylotype is used for clusters of 16S rRNA gene clone sequences that differ from known species by more than 2 % and are at least 98 % similar to the members of their cluster (Suau et al., 1999; Paster et al., 2001; Dekio et al., 2005). ‘Phylotype A’ and ‘phylotype B’ refer to species presented in a previous report (Dekio et al., 2005). These were characterized by bacterial 16S rRNA gene sequences obtained from swab-scrubbed skin samples of healthy humans (GenBank accession nos AB161079 and AB161080).
|
RESULTS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
). All isolates of Staphylococcus species were coagulase-negative, and Staphylococcus aureus was not isolated from any sample examined in this study. All of the cultured bacteria were the known classical components of skin microbiota, and the total number of c.f.u. cm–2 of the samples was in the same order as that reported by Marples & McGinley (1974) and Evans (1975).
|
). The presence of each major bacterial species was reflected as one peak in the HhaI pattern and one peak in the MspI pattern; thus, we defined the two peaks of a single species as a ‘peak set’. Eighteen major corresponding peak sets were detected from the samples (Table 3). Large differences were observed among the participants with regard to the peak sets, which reflected differences in the diversity of the bacterial communities.
|
). Peak sets IV, VII, XIV and XVI were assigned to 16S rRNA sequences of well-known skin bacteria in the databases, and peak sets V, VI, IX and X were assigned to those obtained from our previous clone library analysis (Dekio et al., 2005). The computer simulation successfully identified 8 of the 18 major peak sets; among these, Stenotrophomonas maltophilia, Dietzia maris, and phylotype A and phylotype B (yet-to-be-cultured bacteria) are members that have not been reported previously to be isolated from human skin.
|
). Differences of up to 6 bp were observed between the predicted and observed peaks. A similar phenomenon was observed in an analysis of microbiota in saliva by Sakamoto et al. (2003); the reason for this is unknown.
Assignment of T-RFLP peaks using clone library analysis
Despite the extensive computer simulation, 10 of the 18 major peak sets remained unassigned. To identify the organisms producing these peaks, we performed gel electrophoresis using a mixture of the T-RFs used for T-RFLP analysis, constructed clone libraries from the excised gel bands and sequenced them. Based on the obtained nucleotide sequences (GenBank accession nos AB264551–AB264556), we assigned the peak sets to Alcaligenes xylosoxidans (peak set XII), Staphylococcus species (peak set XIII) and P. acnes (peak set XVII). So the clone library analysis successfully identified three peak sets; among these, A. xylosoxidans has not been reported previously to be isolated from human skin.
At this stage, 7 of the 18 peak sets remained unidentified. These peaks were difficult to assign on the basis of gel extraction and data from clone libraries; some peaks were too short for the analysis, others were difficult to excise from the gel because the bands on the gel were in extremely close proximity to other bands.
Heterogeneity in the distribution of bacterial species among individual samples
We detected species that were observed more frequently in AD patients than in the controls and vice versa. In AD patients, Stenotrophomonas maltophilia (peak set VI) was significantly more common (5/13 AD patients and 0/10 controls, P = 0.045; Table 5). On the other hand, D. maris (peak set X) was significantly more common in the controls (2/13 AD patients and 8/10 controls, P = 0.003; Table 5).
|
High detection rate of Streptococcus species and phylotypes
Streptococcus species (peak set XIV), which has been considered previously to reside only in infected lesions in human skin, was detected in both AD patients and healthy controls. We detected this species in 7/13 (54 %) patients and in 8/10 (80 %) controls.
Other observations in non-typical participants
Patients AS1, AS4, AM2 and AM3 continuously applied topical tacrolimus (Protopic ointment). Although this drug is known to reduce the number of Staphylococcus species on the skin (Pournaras et al., 2001; Remitz et al., 2001), the number of Staphylococcus species in the skin of these patients appeared to be the same as that observed in other patients in the analysis of both the molecular T-RF peaks and cultured colonies. This ointment thus appears to be less effective in reducing the number of Staphylococcus species in AD patients in vivo.
For two patients (AS3 and AM9), samples were taken twice over 7 days. Although culture analysis revealed a slight change, the T-RFLP patterns and the appearance of the skin showed no apparent change. This suggests that the skin microbiota do not change in a short time period in patients, irrespective of a change in the ointment used, unless a change in the skin condition is observed.
Participant C1, a healthy control, had a unique set of micro-organisms; only Staphylococcus species, Streptococcus species and P. acnes were detected, and the largest peak set corresponded to Staphylococcus species. This pattern was specific to AD patients. The culture result supported the finding that the count of Staphylococcus species was atypically as high as that of Propionibacterium species. However, this individual did not have any allergic diseases and was experiencing a very good skin condition. This seemingly unexplainable phenomenon may have occurred due to the unique host–parasite relationship in this individual; a molecular analysis of her faeces revealed that she had a unique set of micro-organisms in the intestine when compared with that observed in other healthy individuals (data not shown).
|
DISCUSSION |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
We used forehead skin as the sampling site. According to published studies using culture techniques, skin microbiota vary largely among different sites (Marples, 1982). To examine differences between the data obtained in this study and those in a previous report on the 16S rRNA gene-based bacterial profiling of forehead skin samples in healthy individuals (Dekio et al., 2005), we used the same sampling site. The selected site has one of the greatest bacterial counts on the human skin and is considered to be stable despite environmental changes (Evans, 1975; Dekio et al., 2005). Such a high population appears to be beneficial to reduce the contamination effect from the environment. As a result, the number of c.f.u. cm–2 determined in our study was comparable to that in published reports (Marples & McGinley, 1974; Evans, 1975) (Table 2), suggesting that the sampling procedure was reliable for the analyses of physiological steady-state conditions with minimal contamination.
Staphylococcus species were detected as two different peak sets (peak sets VII and XIII) in our T-RFLP analysis. This was probably attributable to strain variation of nt 238, which is located in the HhaI restriction site. Analysis of the sequence data for several strains in the nucleotide databases revealed that Staphylococcus aureus should be detected only as peak set VII, but other Staphylococcus species strains should be detected as either peak set VII or XIII. The sequence data from our clone library analysis (GenBank accession nos AB264552–AB264555) verified this possibility. In contrast, P. acnes was also detected as two different peak sets (XVI and XVII), but this was unlikely to have resulted from a difference at the restriction point as in the case of Staphylococcus species. The sequence of our clone library from peak set XVII (fragment sizes 781 bp/165 bp) (GenBank accession no. AB264556) exhibited the HhaI-restrictive sequence at nt 671 and not nt 781; therefore, there was no apparent reason for the generation of the 781 bp fragment. One hypothesis is that the conformational structure of the single strands allowed cleavage by the restriction enzyme at this point (‘pseudo-T-RF’, discussed below).
Using the computer simulation and clone library technique, we successfully detected five species/phylotypes that have not been reported previously to be isolated from human skin; Stenotrophomonas maltophilia, D. maris, A. xylosoxidans, and phylotype A and phylotype B. Stenotrophomonas maltophilia is an environmental bacterium isolated from soil and crops (Hauben et al., 1999). As this bacterium is related to dermatitis in sheep (London & Griffith, 1984), its presence may worsen the condition of atopic skin. Furthermore, it is an emerging pathogen in the respiratory and urinary tracts of immunocompromised hosts (Vartivarian et al., 1996; Senol, 2004); our findings suggest that the skin microbiota of AD patients serve as a reservoir for such opportunistic pathogens. D. maris is a common bacterium found in ponds and on the surface of fish. It is possible that it exists commonly in humans and maintains a healthy environment on the skin. D. maris was observed to coexist with unspecified bacteria (peak set XI), suggesting a symbiotic relationship between them.
Compared with previous culture studies, our molecular analysis detected a higher rate of Streptococcus species. This species has been reported previously to occur only in infected lesions in human skin and not on normal atopic or healthy skin. In our study, peaks corresponding to Streptococcus species were much smaller relative to those corresponding to Staphylococcus species in every sample. This suggests the coexistence of a small number of Streptococcus species and a large number of Staphylococcus species. The reason for the low detection rate of Streptococcus species by culture analysis is unknown, but we hypothesize that the ratio of the population of Streptococcus species to that of Staphylococcus species is very small in culture plates; therefore, their colonies remained unrecognized. Our molecular method enabled the detection of even these minor components of microbiota.
Despite the use of computer simulation and the clone library technique to assign peak sets to bacterial species, 7 of the 18 major peak sets remained unidentified. Whilst these could be real micro-organisms, such peaks could also be ‘pseudo-T-RFs’, which have been described by Egert & Friedrich (2003). Single-stranded DNA molecules, which result from a large number of PCR cycles, form partial double strands within themselves. New restriction sites are generated after the formation of a complex structure consisting of covalent bonds within a strand and hydrogen bonds between the two strands of the partial double strand; this results in the generation of ‘pseudo-T-RFs’. To avoid such peaks, pre-treatment using single-strand-specific mung bean nuclease should be performed before capillary electrophoresis.
In previous analyses of skin microbiota in AD patients, interest was limited to the number and ratio of Staphylococcus species to P. acnes (Leyden et al., 1974; Ogawa et al., 1994). However, such culture analyses were unable to categorize participants clearly based on the condition of their skin, and only tendencies were observed. In contrast, our method enabled a clearer clustering of skin microbiota, although a small number of exceptions existed. This suggests that bacteria other than Staphylococcus species and P. acnes should also be analysed in AD patients. Furthermore, a total analysis of skin microbiota should also include analysis of Stenotrophomonas maltophilia, D. maris, Streptococcus species and phylotype A.
Our aim is to search for novel micro-organisms that are beneficial or harmful to AD patients. Our high-throughput molecular analysis successfully captured the micro-organisms that were relatively specific to AD patients and the controls. However, it is impossible to determine whether the composition of microbiota is the cause, a worsening factor or only the result of AD, as the data are solely observational.
|
ACKNOWLEDGEMENTS |
|---|
|
REFERENCES |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | J MED MICROBIOL | MICROBIOLOGY | J GEN VIROL | ALL SGM JOURNALS |
