Вы находитесь на странице: 1из 8

Journal of Medical Microbiology (2012), 61, 645–652

DOI 10.1099/jmm.0.041764-0

Correspondence Kirstin J. Edwards Kirstin.edwards@hpa.org.uk

Received 15 December 2011 Accepted 9 February 2012

Utility of real-time amplification of selected 16S rRNA gene sequences as a tool for detection and identification of microbial signatures directly from clinical samples

Kirstin J. Edwards, Julie M. J. Logan, Sally Langham, Craig Swift and Saheer E. Gharbia

Department of Bioanalysis and Horizon Technologies, Health Protection Agency Microbiology Services Division Colindale, 61 Colindale Avenue, London NW9 5EQ, UK

The potential of incorporating a real-time PCR for amplification and detection of 16S rRNA gene signatures directly from clinical samples was assessed as a tool for diagnostics. Universal PCR primers spanning short variable regions (~500 bp) were optimized for real-time PCR and tested in comparison with a longer fragment (~1400 bp) generated from block-based amplification. Real-time PCR had improved sensitivity of detection (8 % increase), decreased amplification time and simplified downstream processing. The real-time PCR primers also offered an improvement in detection of bacteria from samples that demonstrated inhibition with the block-based primers and in the resolution of mixed-sequence traces. In addition to testing primer sensitivity, the effect of amplifying and sequencing two different variable regions of the 16S rRNA gene on organism identification was compared. By amplifying and sequencing a shorter variable region, the number of species that were identified to the species level was reduced by 18 %.

INTRODUCTION

16S rRNA gene amplification and sequencing has been well documented as a useful tool for detection and identifica- tion of bacteria (Bosshard et al. , 2003; Drancourt et al. , 2000; Mignard & Flandrois, 2006; Tang et al., 1998; Woo et al. , 2008). Several strategies and targets have been used based on the utility of the entire gene or specifically targeting one or more variable regions within the gene. As well as proving fundamental in taxonomy, where 16S rRNA gene sequencing has been used repeatedly to clarify relationships within a wide range of taxa (Snel et al., 1999:

Stackebrandt & Goebel, 1994), the 16S rRNA gene has also been used as a tool for probing the diversity of microbial communities (Baker et al. , 2001; Hill et al. , 2000; Hohn et al., 2002). Here, it was used to identify both cultivable and uncultivable bacteria, the latter due to difficulty in mimicking the environmental habitats and nutritional requirements of these organisms in the laboratory. Similar studies have resulted in the discovery of several new taxonomic groups through direct amplification of 16S rRNA gene fragments (Drancourt et al. , 2000, 2004; Woo et al., 2008).

With the increase in molecular methods and supporting technologies in laboratories, the use of 16S rRNA gene PCR and sequencing has been exploited in clinical microbiology.

Abbreviations: CP, crossing point; LC, LightCycler; V, variable region.

Initially, this was used for the identification of bacterial isolates (Bosshard et al., 2003; Drancourt et al. , 2004; Mignard and Flandrois, 2006; Zucol et al. , 2006) but increasingly has been employed for direct detection in clinical samples and in particular the analysis of culture- negative samples (Harris & Hartley, 2003; Heikens et al. , 2005; Kommedal et al., 2009; Zucol et al. , 2006). Direct detection has allowed identification of bacteria that have not survived transportation, are suppressed by antibiotic treatment and have complex growth requirements. Several groups have published methodologies for block-based analysis (also known as end-point analysis) of different sample types and have reported the advantages and pitfalls of using 16S rRNA gene PCR and sequencing (Corless et al., 2000; Janda & Abbott, 2007; Millar et al. , 2002). Unlike real- time PCR, block-based assays require agarose gel electro- phoresis to visualize the PCR product.

Real-time PCR offers many advantages for detection of the 16S rRNA gene from clinical samples. For other gene targets, real-time PCR has demonstrated increased sens- itivity over block-based assays (Apfalter et al., 2003; Desjardin et al., 1998; Lienard et al. , 2011), which could be vital when detecting low numbers of bacteria in culture- negative samples. Analysis can be semi-quantitative, as there is a direct correlation between the amount of starting material and the threshold crossing point (CP), which is more informative than visualizing the products on an

Downloaded from www.microbiologyresearch.org by

IP: 115.178.237.126 On: Tue, 18 Oct 2016 06:57:31

041764 G 2012 SGM Printed in Great Britain

645

K. J. Edwards and others

agarose gel (Ririe et al. , 1997). The speed of amplification in real-time PCR is generally faster than a block-based approach and it is not necessary to use agarose gel electrophoresis to visualize the products. This reduces the overall processing time, which in turn may have a positive impact on patient care. The aim of this study was to increase sensitivity, speed of amplification and speed/ease of use of downstream processing. In this paper, we report our findings using a LightCycler (LC; Roche) for amplification of two alternative 500 bp 16S rRNA gene fragments covering different variable regions and compar- ison of these with a block-based assay (amplifying a 1400 bp fragment) currently used in our laboratory. Bacterial identification resulting from sequencing the different variable regions was also assessed.

METHODS

Samples. A total of 213 samples from normally sterile sites and clinically suspected to contain an infectious agent were used in this study. Of these, 56 % were culture-negative and 13 % were culture- positive (i.e. an isolate was also recovered from the sample), whilst for the remainder (31 %), culture information was not provided. The range of sample types tested in the study is detailed in Table 1.

DNA extraction. Genomic DNA was extracted from all samples using a DNeasy Blood and Tissue kit (Qiagen). For different sample types, pre-treatment and lysis steps were performed (as detailed below), followed by standard extraction using the manufacturer’s instruc- tions. For tissue samples, one to three pieces of tissue approximately 2–3 mm 2 were placed in a sterile tube, proteinase K solution (Qiagen) and ATL buffer (Qiagen) were added and the tubes were incubated overnight at 56 u C. Following standard extraction, the tissue samples were resuspended in 50 m l nuclease-free water. Up to 200 ml of liquid samples was used, and if necessary the sample volume was made up to

200 m l using PBS. The samples were extracted using a standard protocol but were eluted in half of the starting volume (but not less than 50 m l) with nuclease-free water. For liquid samples that could not be pipetted (e.g. pus), a portion ~4 mm in diameter was placed into a sterile tube, proteinase K solution and ATL buffer were added and the tube was incubated overnight at 56 u C. After extraction using a standard Qiagen protocol, the DNA was resuspended in

50 m l nuclease-free water. For samples that had been fixed in wax,

either a 3 mm 2 piece of tissue, with as much wax as possible removed, or four to five shavings were placed into a sterile tube and 1.2 ml UltraClear solution (TAAB Laboratories) was added. The tubes were centrifuged at 16 000 g for 5 min and the supernatant removed. The UltraClear wash was repeated and 1.2 ml ethanol (96–100 %) was added to the pellet to remove residual UltraClear. The tubes were vortexed and centrifuged at 16 000 g for 5 min, the ethanol was removed and the ethan ol wash repeated. The pellet was resuspended in proteinase K solution and buffer AL (Qiagen) and the extraction was completed as described above. The samples were resuspended in 50 m l nuclease-free water. From a blood culture bottle, 200 m l fluid was placed into a sterile tube and 100 m l lysis buffer [ 5 M guanidine hydrochloride in 100 mM Tris/HCl (pH 8.0) ] , 400 m l sterile nuclease-free water and 800 m l 99 % benzyl alcohol were added (Fredricks & Relman, 1998). Tubes were

centrifuged at 7000 g for 5 min and 400 m l aqueous phase was transferred to a fresh tube. To this, 40 m l 3 M sodium acetate and 440 m l 2-propanol were added and the tubes were centrifuged at

16 000 g for 15 min. The supernatant was removed and 1 ml 70 %

ethanol was added to the pellet. The tubes were centrifuged at

16 000 g for 5 min and the supernatant removed. To reduce the

time required for air drying the precipitated DNA, the pellet was washed and eluted through a Qiagen DNeasy spin column as follows: the pellet was resuspended in 50 m l molecular biology-grade

water and this was added directly to the column. After incubation at room temperature for 5 min, the column was centrifuged at 6000 g for 1 min and the eluate transferred to a fresh tube (Hogg et al. , 2008). A negative extraction control (200 m l PBS) was included in each extraction run. Following processing, the extracts were stored at 2 20 u C.

Table 1. Clinical sample types

Clinical sample type

No. samples

Abscess Bronchoalveolar lavage Blood (including EDTA, serum and platelets) Blood culture bottle fluid Bone (including bone marrow) Cerebrospinal fluid Other fluid (including cyst, joint, drain, pericardial, pleural and vitreal) Pus Sputum Other Tissue Brain Heart valve Lymph node Paraffin block (lung lesion) Spine and other joint tissue Subcutaneous Other tissue Total

7

6

31

2

7

19

36

9

1

8

4

41

2

1

14

3

22

213

Identification directly from clinical samples

Primers. The primers used for PCR and sequencing are detailed in Table 2.

Block-based PCR. The PCR mixture consisted of 25 ml Go-Taq Green mastermix (Promega), 0.2 mM each of primers ANT1F and 1392R, 22 ml molecular biology-grade water and 1 ml extracted DNA. The mixture was amplified using an ABI 9700 PCR instrument using the following cycling conditions: denaturation at 95 uC for 2 min and

35 cycles of 95 uC for 45 s, 56 uC for 45 s and 72 u C for 1 min. A final

extension was performed at 72 u C for 5 min. PCR products were separated by electrophoresis on a 2 % ethidium bromide stained E-gel

(Life Technologies) following the manufacturer’s instructions and were visualized by UV transillumination.

Real-time PCR. Real-time PCR was performed using a LC 2.0 in glass capillaries (Roche) or a LC 480 in an optical 96-well plate (Roche). The reaction mix consisted of 10 ml TaKaRa Ex Taq (Lonza), 0.4 mM each primer, 4.4 ml molecular biology-grade water and 2 ml extracted sample. The TaKaRa Ex Taq master mix contained SYBR Green I, which was used for real-time monitoring and detection of PCR products. Two sets of primers were used in the study, the first used forward CLSI and reverse Bosshard (Bosshard LC assay) (Table

2) with the following cycling conditions: denaturation at 95 uC for

10 s and 40 cycles of 95 u C for 10 s, 64 uC for 15 s and 72 u C for 20 s.

The second set of primers was Kommedal forward and Kommedal reverse (Kommedal LC assay) with cycling conditions of: denatura- tion at 95 uC for 10 s and 40 cycles of 95 uC for 15 s, 70 u C for 10 s and 72 uC for 20 s.

PCR controls. Each PCR run included a positive control containing 1 m l Escherichia coli DNA at 50 ng m l 2 1 (Sigma) in the block-based

assay or 1 m l E. coli DNA at 50 pg m l 2 1 (Sigma) in the LC assays, and

a negative control of molecular biology-grade water. Extraction

negative controls were also amplified. To control for PCR inhibition,

a second PCR was performed where an identical reaction mix was

spiked with 1 m l E. coli DNA at 1 ng m l 2 1 for the block-based assay

or 0.5 pg m l 2 1 for the LC assays. For any reactions where amplification was not observed in either the extract or the spiked sample, the original nucleic acid template was diluted 1 : 10 and the PCR repeated.

Sequencing. The LC 2.0 reactions were spun out of the glass capillary into a 1.5 ml Eppendorf tube and these and the LC 480 reactions were transferred to an optical 96-well plate (Life Technologies). All PCR products were purified for sequencing using Agencourt Ampure beads (Beckman Coulter) using an NXP Biomek

Robot following the manufacturer’s protocol (Beckman Coulter). The samples were resuspended in a total volume of 40 m l molecular biology-grade water with 30 ml transferred by the robot to a new 96- well plate for the sequencing set-up. The block-based PCRs were sequenced using primers 357F and 3R, the Bosshard LC products were sequenced using Bosshard forward and reverse primers, and the Kommedal LC products were sequenced with the Kommedal forward and reverse primers (Table 2). Sequencing reactions were performed using an ABI Prism 1.1 or 3.1 Big Dye sequencing kit and analysed on an ABI 3730 DNA Analyzer (Life Technologies).

Identification. Bacterial identification was made by forming contigs of forward and reverse sequences and comparing the contig consensus sequence with sequences in GenBank using the NCBI BLAST algorithm (Altschul et al. , 1990). Criteria of 98 % sequence identity and fewer than 5 bp different were considered sufficient to identify the organism to the species level (Harris & Hartley, 2003). If species-level identification was not possible, then genus-level identification was made. Identification was confirmed by comparison of the sequence data with the ribosomal database project (http://rdp.cme.msu.edu/) using the Seq match analysis tool and using the same analysis criteria as above (Cole et al. , 2007, 2009).

RESULTS

For phase 1 of the study, extracts were prepared from a total of 142 samples and amplified using the block-based assay. Of these samples, 43 were positive and 99 were negative. Nucleic acid extracts were stored at 2 20 u C and subsequently retested using real-time LC PCR over a time period of up to 2 years following the original extraction. For this, two amplicons were generated separately corres- ponding to two different variable (V) regions within the 16S rRNA gene. The Bosshard-based primers, which amplified the region spanning V1–V3, resulted in 54 po- sitive samples and 88 negative samples. The Kommedal- based primers, which amplified regions V3–V4, detected 49 positive and 93 negative samples. Nine of the samples detected using LC PCR and not detected using block-based PCR showed a CP greater than cycle 26, which indicates a weak sample with a low amount of bacterial DNA present, and two samples contained substances that were inhibitory

Table 2. 16S rRNA gene PCR and sequencing primers

Primer

Sequence (5 §A 3§)

Position (based on E. coli numbering

16S rRNA gene variable region

Reference

ANT1F

AGAGTTTGATCCTGGCTCAG

8

V1–V9

Bosshard et al. (2002) Lane (1991) Lane (1991) Rajendram et al. (2006) CLSI (2008) Bosshard et al. (2002); Edwards et al. (1989) Bosshard et al. (2002); Edwards et al. (1989) Kommedal et al. (2008) Kommedal et al. (2008)

1392R

ACGGGCGGTGTGTACAAG

1398

V1–V9

357F

CTCCTACGGGAGGCAGCAG

339

V3–V7

3R

GTTGCGCTCGTTGCGGGACT

1093

V3–V7

Forward CLSI

TTGGAGAGTTTGATCMTGGCTC

4

V1–V3

Reverse Bosshard

GTATTACCGCTGCTG

536

V1–V3

Forward Bosshard

AGAGTTTGATCMTGGCTCAG

8

V1–V3

Kommedal forward

CGGCCCAGACTCCTACGGGAGGCAGCA

330

V3–V4

Kommedal reverse

GCGTGGACTACCAGGGTATCTAATCC

810

V3–V4

K. J. Edwards and others

to the PCR with block-based primers but not with the Bosshard- or Kommedal-based primers. There were no strong positive results (CP , 25) detected by the LC assays that were not detected by the block-based assay. Using the block-based primers, there were 13 extracts that showed mixed sequencing trace files compared with only four with either the Bosshard-based primers or the Kommedal-based primers. Five samples were detected and identified using the Bosshard-based primers but not the Kommedal-based primers, and the organisms identified included three Staphylococcus species, one Streptococcus species and one presumptive E. coli / Shigella species.

After excluding the mixed samples, using the block-based primers (V3–V7) it was possible to identify 20 samples (67 %) to the species level and nine (30 %) to the genus level, with one (3 %) remaining unidentified even to the genus level. By amplifying and sequencing the V1–V3 region (Bosshard primers), it was possible to identify 31 samples (62 %) to the species level and 16 (32 %) to the genus level, with three (6 %) remaining unidentified. By sequencing across variable regions V3 and V4 (Kommedal primers), it was possible to identify 25 samples (56 %) to the species level and 16 (36 %) to the genus level, with four (8 %) remaining unidentified. From the samples that were unidentified, all three sequenced regions were unable to differentiate between Escherichia species and Shigella species, and the V3–V4 region was unable to differentiate between Bacillus species and Burkholderia terricola (Table 3).

In phase 1 of the study, the Bosshard-based primers amplifying V1–V3 resulted in more positive samples (4 % increase) and more species-level identifications (6 % increase) than the Kommedal-based primers (amplifying V3–V4) and as a result were selected for phase 2 of the study. In this phase, a blind comparison was made by testing a further 71 samples by both block-based primers and the Bosshard-based primers, with both PCRs being performed at the same time rather than retrospectively. The block-based primers resulted in 16 positives compared with 23 using the Bosshard-based primers; both also detected one mixed sample and the remainder of the samples were negative. The Bosshard-based primers detected six samples not amplified by the block-based primers, all of which were weak positives as indicated by the LC CP. As before, this demonstrated the increased sensitivity afforded by the smaller PCR product and use of real-time PCR. Amplifying and sequencing across V1–V9 resulted in ten (62 %) species-level identifications and six (38 %) genus-level identifications, whereas amplifying and sequencing V1–V3 resulted in ten (44 %) species-level identifications and 12 (52 %) genus-level identifications, with one (4 %) sample unidentified (Table 3). The negative controls always demonstrated amplification, with CP values ranging from 28 to 35 (no. tested 580; median 5 29). Following sequencing, from the majority of controls a single sequence trace was not obtained that could be used to identify a single organism. The organisms that were

identified included Bacillus species ( n 5 4), Microbacterium species ( n 5 2) and Staphylococcus species ( n 51).

Analysis of the overall data revealed a range of different bacteria detected directly by 16S rRNA gene PCR and sequencing. The total numbers assigned to each genus/ species are detailed in Table 3 and in total 18 different genera were represented of which 73 % were Gram-positive bacteria and 27 % were Gram-negative bacteria. Streptococcus species accounted for 26 % of the total organisms identified and only 7/19 could be identified to the species level. Species of Staphylococcus accounted for 14 % (over half were only identified to the genus level), species of Haemophilus accounted for 14 % and species of Bacillus accounted for 15 %, whilst the remainder of the organisms were only identified in one to three samples.

DISCUSSION

Using the LC for amplification of 16S rRNA gene sig- natures directly from clinical samples improved assay

sensitivity, decreased the overall time taken to perform the test and simplified the downstream processing. The LC assays resulted in identifying more positive samples than the block-based assay, and this improvement in sensitivity could partly be attributed to the decrease in product size.

A smaller amplicon size generally improves the PCR

efficiency, as there is less depletion of the PCR compo- nents, e.g. Taq enzyme or dNTPs, resulting in an assay that is more sensitive. During amplification of smaller ampli- cons, there is less likelihood of secondary structures forming that can reduce efficiency by impairing transcrip- tion, and there are more complete products produced

during each cycle of PCR, both of which contribute to the increased efficiency and sensitivity. A decrease in assay time can only be achieved by targeting a smaller product; larger products can be amplified on the LC but the hold times used cannot be reduced and overall time remains comparable

to block-based amplification. The LC amplification was

completed in 1 h compared with 2.5 h for the block-based

amplification, plus a further reduction in processing time as

an agarose gel was not required to visualize the products.

Amplification was observed on the LC as a curve, and the software calculated the CP of when the amplicon amount exceeded the background threshold and could be detected. This CP value was directly correlated with the amount of starting material in the sample and could be used to provide information on the subsequent amount of template for sequencing plate preparation and for result calling. In this study, 36 % of the positive samples amplified with a CP value of ,20 cycles, which indicates a large amount of bacterial DNA present in the extract (corresponding to a strong band on an agarose gel), and 14 % of the positive samples were amplified with a CP value of .27, which indicates a very low level of bacterial DNA in the sample.

Three of the extracts produced amplicons only by LC PCR but required tenfold dilution before amplification was

http://jmm.sgmjournals.org

Identification directly from clinical samples

649

Table 3. Number of organisms identified to genus and species level and LC CP ranges

ND, Not detected; NT , not tested.

Group

Genus

Total no. identified

No. identified to species level

 

LC CP*

 

Block-based V1–V9

Bosshard V1–V3

Kommedal V3–-V4

Bosshard V1–V3 Kommedal V3–V4

 

region

region

region

Gram positive

Actinomyces

1

ND

1

1

27

29

Anaerococcus

3

3

0

0

18–21

20

Bacillus

11

4

2

0

14–17

13–28

Gordonia

1

0

0

0

12

12

Lactobacillus

1

1

1

1

22

24

Mycobacterium

2

0

0

0

20–21

20

Nocardia

1

ND

1

NT

27

28

Propionibacterium

1

1

1

1

11

28

Staphylococcus

10

3

5

3

16–29

21–28

Streptococcus

19

5

7

3

14–27

15–29

Gram negative

Aggregatibacter

2

2

2

2

22

21

Capnocytophaga

2

22

NT

21–26

NT

Fusobacterium

2

1

1

0

13–21

23

Haemophilus

10

6

10

8

12–23

18–29

Kingella

1

1

1

1

25

25

Moraxella

1

ND

1

1

20

15

Pasteurella

1

1

1

1

21

20

Pseudomonas

1

00

NT

16–25

25

Unnamed

Escherichia/Shigella

3

0

0

0

21–25

27–29

Cetobacterium/Fusobacterium

1

ND

0

NT

27

NT

*The range of values is presented when different CPs were obtained from different samples.

Downloaded from www.microbiologyresearch.org by IP: 115.178.237.126 On: Tue, 18 Oct 2016 06:57:31

K. J. Edwards and others

observed. The original extracts and the dilutions demon- strated inhibition (a second reaction spiked with E. coli DNA did not produce an amplicon) with the block-based PCR. All three extracts were from sample types known to cause problems due to PCR inhibitors (e.g. chemicals such as sodium polyanetholesulfonate), blood culture fluid and paraffin-embedded tissue (Akane et al., 1994; Fredricks & Relman, 1998; Holodniy et al., 1991). The bacterial load in these samples would have been reduced following dilution and may have been beyond the range of the block-based primers. Additionally, some of the samples may have contained smaller sheared fragments of DNA resulting from sample preparation (e.g. paraffin-embedded tissues), which can result in shearing into fragments of , 650 bp following Qiagen extraction, and amplification with the block-based primers would therefore have been greatly reduced.

The inherent problem observed in amplification of the 16S rRNA gene is background contamination of reagents with low levels of bacteria, which can be co-amplified (Corless et al., 2000; Harris and Hartley, 2003), and was observed in all assays. The LC assays were adapted for use on the LC from a SmartCycler method published by Kommedal et al. (2008). The authors used the following definition: ‘a positive sample was defined as a sample reaching the fluorescence threshold value (CP) ¢ three cycles before the negative control and a sample was also defined as positive if it reached the CP fewer than three cycles before the negative control but if the subsequent melting curve analysis showed a single distinct peak clearly different from the negative control’. Both of these criteria were considered in this study, but the negative-control CP value showed considerable run-to-run variation and all samples were sequenced. This identified amplicons from 11 samples that would have been considered negative by the above criteria, but subsequent sequencing identified clinically relevant organisms. Analysis of the melting-curve data for these late-amplified samples did not reveal a melting curve that was significantly different from the negative controls and could not be used to differentiate these positive samples. In this study, the following criteria were used: all samples were sequenced regardless of CP, but caution was used with samples that amplified in less than three cycles from the negative control. In addition, samples that amplified after cycle 25 were used neat in subsequent sequencing reaction, and samples that amplified before cycle 25 were diluted 1 : 4 in the sequencing reaction, whilst melting-curve data were not used in the analysis and are no longer performed.

The background contamination of the reagents can also cause a problem with mixed-sequence traces. If the bacterial load of the clinically significant bacteria is sufficiently high, then it is preferentially amplified and the background contamination is not observed. However, when there are only low levels of clinically significant bacteria, then the background is co-amplified and a mixed trace can be observed. The master mix used for amplification with the block-based primers was different

from that used for the LC, which requires a specialized master mix optimal for fast amplification. In this study, there were significantly more mixed traces detected using the block-based assay. The LC reagents do have a low level background contamination, as demonstrated by late amplification curves with high CP values in the negative controls; however, this was not amplified in sufficient amounts or was not composed of a dominant organism and so mixed-sequence traces were not observed. There were four extracts that showed mixed-sequence traces with all assays and represented genuine mixed bacterial infections. Specialized Taq enzymes and master mixes are available with a guaranteed low background DNA level. The use of these may improve the background contam- ination observed but will increase the cost of performing the assay.

There are a large number of publications on the use of 16S rRNA gene sequencing for bacterial identification, but there is no consensus on the optimal region of the gene to target. Therefore, it was important in this study to review the differences observed from amplifying and sequencing different regions. The Bosshard-based primers amplified the region spanning V1–V3, whereas the Kommedal-based primers amplified the region spanning V3–V4. The Bosshard-based primers performed slightly better than the Kommedal-based primers for both sensitivity and resolution. In the paper by Kommedal et al. (2009), the authors reported that the Kommedal-based primers bound poorly to a range of bacteria including Chlamy- dia abortus , Chlamydia psittaci , Chlamydophilia pneumo- niae , Coxiella burnetti , Dermabacter hominis , Leuconostoc species, Microbacterium species and Propionibacterium species. Thus, based on this evidence as well as the results obtained in this study, the Bosshard-based primers are more suitable for amplification from clinical samples.

By amplifying and sequencing the variable region V1–V3, 46 % of samples could only be identified to the genus level (Table 3). There are a number of genera that do not resolve well with 16S rRNA gene sequencing, such as Streptococcus ( pneumoniae / mitis/ oralis / intermedius ), Staphylococcus ( aur- eus / epidermidis / capitis / caprae) and some Bacillus species (Kommedal et al. , 2008; Harris & Hartley, 2003), and these organisms represented the most prevalent organisms found in this study (Table 3). Therefore, for the majority of samples, reducing the size of the region sequenced did not have a negative impact on identification, and it would be more beneficial to sequence alternative gene targets (e.g. rpoB or secA ) rather than sequence a larger region of the 16S rRNA gene if species-level identification was clinically important. There were four samples where the V1–V3 region could not identify the organism to the genus level ( Cetobacterium/Fusobacterium and Escherichia/Shigella spe- cies). E. coli and Shigella dysenteriae are genotypically close enough to be considered the same species but have been kept separate for clinical reasons (Clarridge, 2004), and for the sample containing either Cetobacterium or Fusobac- terium , a larger region of the 16S rRNA gene sequence is

Identification directly from clinical samples

required to resolve these to the genus or species level, delaying the clinical result by 1–2 days.

Considerable work has already been undertaken to show the utility of 16S rRNA gene PCR and sequencing for identification of bacteria. It has been shown previously to be superior to phenotypic identification methods and nearly always results in an identification (Clarridge, 2004; Janda & Abbott, 2007). The application for clinical samples and in particular culture-negative samples has also been demonstrated (Harris & Hartley, 2003; Kommedal et al. , 2009), as have the pitfalls of quality of sequence databases, contamination of reagents and lack of resolution for some genera. The work in this study was undertaken to establish whether using a real-time PCR-based approach would improve the sensitivity and speed of detection. We have demonstrated that amplification can easily be performed on the LC and that this approach offers a considerable improvement in sensitivity of detection. In our hands, it was relatively easy to differentiate amplification of infection-causing organisms against the background con- tamination of the reagents, and fewer mixed-sequence traces were observed. Amplification of a smaller region of the gene affected the resolution of the data but not significantly. Instead, it improved the sensitivity, allowed amplification from difficult sample types and improved the overall time taken to perform the analysis. Using this approach, it is possible to perform amplification and sequencing, leading to identification in , 24 h following extraction. It is also possible that, by employing different extraction methods and techniques, the overall time could be reduced further and the sensitivity of detection improved, offering an additional improvement for detec- tion of bacteria directly from clinical samples.

REFERENCES

Akane, A., Matsubara, K., Nakamura, H., Takahashi, S. & Kimura, K.

(1994). Identification of the heme compound copurified with deoxyribonucleic acid (DNA) from bloodstains, a major inhibitor of polymerase chain reaction (PCR) amplification. J Forensic Sci 39,

362–372.

Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J.

(1990). Basic local alignment search tool. J Mol Biol 215, 403–410.

Apfalter, P., Barousch, W., Nehr, M., Makristathis, A., Willinger, B., Rotter, M. & Hirschl, A. M. (2003). Comparison of a new quantitative

ompA -based real-time PCR TaqMan assay for detection of Chlamydia pneumoniae DNA in respiratory specimens with four conventional PCR assays. J Clin Microbiol 41 , 592–600.

Baker, G. C., Gaffar, S., Cowan, D. A. & Suharto, A. R. (2001).

Bacterial community analysis of Indonesian hot springs. FEMS Microbiol Lett 200, 103–109.

Bosshard, P. P., Zbinden, R. & Altwegg, M. (2002). Turicibacter

sanguinis gen. nov., sp. nov., a novel anaerobic, Gram-positive bacterium. Int J Syst Evol Microbiol 52, 1263–1266.

Bosshard, P. P., Abels, S., Zbinden, R., Bo¨ ttger, E. C. & Altwegg, M.

(2003). Ribosomal DNA sequencing for identification of aerobic Gram-positive rods in the clinical laboratory (an 18-month evaluation). J Clin Microbiol 41 , 4134–4140.

Clarridge, J. E., III (2004). Impact of 16S rRNA gene sequence analysis for identification of bacteria on clinical microbiology and infectious diseases. Clin Microbiol Rev 17, 840–862.

CLSI (2008) . Interpretive Criteria for Identification of Bacteria and Fungi by DNA Target Sequencing . Approved Guideline MM18-A. Wayne, PA: Clinical and Laboratory Standards Institute.

Cole, J. R., Chai, B., Farris, R. J., Wang, Q., Kulam-Syed-Mohideen,

A. S., McGarrell, D. M., Bandela, A. M., Cardenas, E., Garrity, G. M.

& Tiedje, J. M. (2007). The ribosomal databa se project (RDP-II):

introducing myRDP space and quality controlled public data. Nucleic Acids Res 35 (Database issue), D169–D172.

Cole, J. R., Wang, Q., Cardenas, E., Fish, J., Chai, B., Farris, R. J., Kulam-Syed-Mohideen, A. S., McGarrell, D. M., Marsh, T. & other

authors (2009). The Ribosomal Database Project: improved align- ments and new tools for rRNA analysis. Nucleic Acids Res 37 (Database issue), D141–D145.

Corless, C. E., Guiver, M., Borrow, R., Edwards-Jones, V., Kaczmarski, E. B. & Fox, A. J. (2000). Contamination and sensitivity

issues with a real-time universal 16S rRNA PCR. J Clin Microbiol 38,

1747–1752.

Desjardin, L. E., Chen, Y., Perkins, M. D., Teixeira, L., Cave, M. D. &

Eisenach, K. D. (1998). Comparison of the ABI 7700 system (TaqMan) and competitive PCR for quantification of IS6110 DNA in sputum during treatment of tuberculosis. J Clin Microbiol 36,

1964–1968.

Drancourt, M., Bollet, C., Carlioz, A., Martelin, R., Gayral, J. P. &

Raoult, D. (2000). 16S ribosomal DNA sequence analysis of a large collection of environmental and clinical unidentifiable bacterial isolates. J Clin Microbiol 38 , 3623–3630.

Drancourt, M., Berger, P. & Raoult, D. (2004). Systematic 16S rRNA

gene sequencing of atypical clinical isolates identified 27 new bacterial species associated with humans. J Clin Microbiol 42, 2197–2202.

Edwards, U., Rogall, T., Blo¨ cker, H., Emde, M. & Bo¨ ttger, E. C.

(1989). Isolation and direct complete nucleotide determination of entire genes. Characterization of a gene coding for 16S ribosomal RNA. Nucleic Acids Res 17 , 7843–7853.

Fredricks, D. N. & Relman, D. A. (1998). Improved amplification of

microbial DNA from blood cultures by removal of the PCR inhibitor sodium polyanetholesulfonate. J Clin Microbiol 36, 2810–2816.

Harris, K. A. & Hartley, J. C. (2003). Development of broad-range 16S

rDNA PCR for use in the routine diagnostic clinical microbiology service. J Med Microbiol 52, 685–691.

Heikens, E., Fleer, A., Paauw, A., Florijn, A. & Fluit, A. C. (2005).

Comparison of genotypic and phenotypic methods for species-level identification of clinical isolates of coagulase-negative staphylococci. J Clin Microbiol 43, 2286–2290.

Hill, G. T., Mitkowski, N. M., Aldrich-Wolfe, L., Emele, L. R., Jurkonie,

D. D., Ficke, A., Maldonado-Ramirez, S., Lynch, S. T. & Nelson, E. B.

(2000). Methods for assessing the composition and diversity of soil microbial communities. Appl Soil Ecol 15, 25–36.

Hogg, G. M., McKenna, J. P. & Ong, G. (2008). Rapid detection of

methicillin-susceptible and methicillin-resistant Staphylococcus aureus directly from positive BacT/Alert blood culture bottles using real-time polymerase chain reaction: evaluation and comparison of 4 DNA extraction methods. Diagn Microbiol Infect Dis 61, 446–452.

Hohn, M. J., Hedlund, B. P. & Huber, H. (2002). Detection of 16S

rDNA sequences representing the novel phylum ‘‘Nanoarchaeota’’:

indication for a wide distribution in high temperature biotopes. Syst Appl Microbiol 25 , 551–554.

Holodniy, M., Kim, S., Katzenstein, D., Konrad, M., Groves, E. &

Merigan, T. C. (1991). Inhibition of human immunodeficiency virus gene amplification by heparin. J Clin Microbiol 29, 676–679.

K. J. Edwards and others

Janda, J. M. & Abbott, S. L. (2007). 16S rRNA gene sequencing for

bacterial identification in the diagnostic laboratory: pluses, perils, and pitfalls. J Clin Microbiol 45 , 2761–2764.

Kommedal, O., Karlsen, B. & Saebø, O. (2008). Analysis of mixed

sequencing chromatograms and its application in direct 16S rRNA

gene sequencing of polymicrobial samples. J Clin Microbiol 46, 3766–

3771.

Kommedal, O., Kvello, K., Skja˚ stad, R., Langeland, N. & Wiker, H. G.

(2009). Direct 16S rRNA gene sequencing from clinical specimens, with special focus on polybacterial samples and interpretation of mixed DNA chromatograms. J Clin Microbiol 47, 3562–3568.

Lane, D. J. (1991). 16S/23S rRNA sequencing. In Nucleic Acid Techniques in Bacterial Systematics , pp. 115–175. Edited by E. Stackebrandt & M. Goodfellow. London: John Wiley & Sons Ltd.

Lienard, J., Croxatto, A., Aeby, S., Jaton, K., Posfay-Barbe, K.,

Gervaix, A. & Greub, G. (2011). Development of a new Chlamydiales- specific real-time PCR and its application to respiratory clinical samples. J Clin Microbiol 49 , 2637–2642.

Mignard, S. & Flandrois, J. P. (2006). 16S rRNA sequencing in routine

bacterial identification: a 30-month experiment. J Microbiol Methods 67, 574–581.

Millar, B. C., Xu, J. & Moore, J. E. (2002). Risk assessment models and

contamination management: implications for broad-range ribosomal DNA PCR as a diagnostic tool in medical bacteriology. J Clin Microbiol 40, 1575–1580.

Rajendram, D., Ayenza, R., Holder, F. M., Moran, B., Long, T. & Shah,

H. N. (2006). Long-term storage and safe retrieval of DNA from microorganisms for molecular analysis using FTA matrix cards. J Microbiol Methods 67 , 582–592.

Ririe, K. M., Rasmussen, R. P. & Wittwer, C. T. (1997). Product

differentiation by analysis of DNA melting curves during the polymerase chain reaction. Anal Biochem 245, 154–160.

Snel, B., Bork, P. & Huynen, M. A. (1999). Genome phylogeny based

on gene content. Nat Genet 21 , 108–110.

Stackebrandt, E. & Goebel, B. M. (1994). Taxonomic note: a place for

DNA–DNA reassociation and 16S rRNA sequence analysis in the

present species definition in bacteriology. Int J Syst Bacteriol 44, 846–

849.

Tang, Y.-W., Ellis, N. M., Hopkins, M. K., Smith, D. H., Dodge, D. E. &

Persing, D. H. (1998). Comparison of phenotypic and genotypic techniques for identification of unusual aerobic pathogenic gram- negative bacilli. J Clin Microbiol 36 , 3674–3679.

Woo, P. C., Lau, S. K., Teng, J. L., Tse, H. & Yuen, K.-Y. (2008). Then

and now: use of 16S rDNA gene sequencing for bacterial identification and discovery of novel bacteria in clinical microbiology laboratories. Clin Microbiol Infect 14 , 908–934.

Zucol, F., Ammann, R. A., Berger, C., Aebi, C., Altwegg, M., Niggli, F. K.

& Nadal, D. (2006). Real-time quantitative broad-range PCR assay for detection of the 16S rRNA gene followed by sequencing for species identification. J Clin Microbiol 44, 2750–2759.