evaluation of single-nucleotide primer extension for ... · potential of snupe for simultaneous...

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, May 2009, p. 2850–2860 Vol. 75, No. 90099-2240/09/$08.00�0 doi:10.1128/AEM.01910-08Copyright © 2009, American Society for Microbiology. All Rights Reserved.

Evaluation of Single-Nucleotide Primer Extension for Detection andTyping of Phylogenetic Markers Used for Investigation of

Microbial Communities�

Marcell Nikolausz,1* Antonis Chatzinotas,2 Andras Tancsics,3 Gwenael Imfeld,4 and Matthias Kastner1

UFZ, Helmholtz Centre for Environmental Research, Department of Environmental Biotechnology, Permoserstr. 15, D-04318 Leipzig,Germany1; UFZ, Helmholtz Centre for Environmental Research, Department of Environmental Microbiology, Permoserstr. 15,

D-04318 Leipzig, Germany2; Eotvos Lorand University of Science, Department of Microbiology, Pazmany Peter Setany 1/c,1117 Budapest, Hungary3; and UFZ, Helmholtz Centre for Environmental Research, Department of

Isotope Biogeochemistry, Permoserstr. 15, D-04318 Leipzig, Germany4

Received 18 August 2008/Accepted 20 February 2009

Single-nucleotide primer extension (SNuPE) is an emerging tool for parallel detection of DNA sequences ofdifferent target microorganisms. The specificity and sensitivity of the SNuPE method were assessed byperforming single and multiplex reactions using defined template mixtures of 16S rRNA gene PCR productsobtained from pure bacterial cultures. The mismatch discrimination potential of primer extension was inves-tigated by introducing different single and multiple primer-target mismatches. The type and position of themismatch had significant effects on the specificity of the assay. While a 3�-terminal mismatch has a consid-erable effect on the fidelity of the extension reaction, the internal mismatches influenced hybridization mostlyby destabilizing the hybrid duplex. Thus, carefully choosing primer-mismatch positions should result in a highsignal-to-noise ratio and prevent any nonspecific extension. Cyclic fluorescent labeling of the hybridizedprimers via extension also resulted in a significant increase in the detection sensitivity of the PCR. In multiplexreactions, the signal ratios detected after specific primer extension correlated with the original template ratios.In addition, reverse-transcribed 16S rRNA was successfully used as a nonamplified template to prove theapplicability of SNuPE in a PCR-independent manner. In conclusion, this study demonstrates the greatpotential of SNuPE for simultaneous detection and typing of various nucleic acid sequences from bothenvironmental and engineered samples.

Fast detection, differentiation, and identification of bacteriaare crucial tasks in clinical, food, and environmental microbi-ology. Cultivation-independent tools not only save time com-pared to cultivation-based techniques but also allow access tothe difficult-to-cultivate part of a microbial community. Molec-ular detection methods are usually based on hybridization ofoligonucleotide probes to signature sequences (phylogeneti-cally informative regions) in the nucleic acids (RNA or DNA)of the target microorganisms. Verification of the hybridizationevent can be accomplished by detection of hybridized labeledprobes in situ (e.g., fluorescence in situ hybridization [FISH])or ex situ (dot blot hybridization). Combining two specificoligonucleotides in a PCR increases the sensitivity of specificdetection, while real-time monitoring of the amplificationproduct formed allows quantification of the original template(for a review, see reference 17). Multiple detection can beachieved by using more than one primer pair targeting severalloci in multiplex PCR assays (for a review, see reference 32).However, the main disadvantages of FISH are its restrictedcapability for parallel analysis of several target groups in thesame sample and limitations in probe design due to differencesin accessibility of the probes to their target sites (3, 7). More-

over, detection of slowly growing bacteria with low ribosomecontents requires labor-intensive techniques (30, 36). Multi-plex PCR also has limitations for multiplexing and challengesfor primer design (32).

Recently, single-nucleotide primer extension (SNuPE) wasproposed as a fast, semiquantitative multiplex detection toolfor analyzing sequence variants. This method is frequentlyused for determination of single-nucleotide polymorphismsand benefits from the fidelity of dideoxynucleoside triphos-phate (ddNTP) incorporation catalyzed by a DNA polymerase.When primer extension takes place on a solid support, themethod is called minisequencing (35, 37), while a reaction insolution is referred to as SNuPE (34) or single-base extension(15). These methods were originally developed for routinemedical diagnosis of genetic disorders (23, 35) or for use inforensic research (38). Different versions of the primer exten-sion technique have also been used recently for fast identifi-cation and genotyping of microbial strains (9, 31). Recentstudies showed that detection of a hybridization event via la-beling of the hybridized primer in the extension reaction ispossible. However, the use of this method as a detection tool inapplied and environmental microbiology has not been fullyexploited so far. Rudi and coworkers were the first workerswho used a minisequencing approach with PCR products fromenvironmental DNA to detect toxic cyanobacteria by labelingonly one of the four ddNTPs used in the reaction (27). Mul-tiplexing was accomplished by hybridizing the labeled productsto complementary oligonucleotides in an array format. In com-

* Corresponding author. Mailing address: UFZ Helmholtz Centrefor Environmental Research, Department of Environmental Biotech-nology, Permoserstr. 15, 04318 Leipzig, Germany. Phone: 49 341 2351763. Fax: 49 341 235 2492. E-mail: [email protected].

� Published ahead of print on 27 February 2009.

2850

on May 18, 2020 by guest

http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/

bination with antibody-based chromogenic visualization, ge-netic profiles of cyanobacterial diversity (28), microbial com-munities in vegetable salads (25), and Listeria strains (26) wereobtained. However, this approach is labor-intensive and time-consuming and requires specific equipment; furthermore, theprimer is restricted to certain positions since only one termi-nator nucleotide is labeled.

An alternative strategy for multiplexing in solution benefitsfrom incorporation of four differently labeled ddNTPs andattachment of mobility modifiers to the different primers. Sub-sequent separation using capillary electrophoresis and laser-induced fluorescence detection results in a very fast assay thatis easy to interpret. Determination of the incorporated nucle-otide provides additional proof of the assay specificity or mayeven provide extra phylogenetic information. The first appli-cation of primer extension with four differently labeledddNTPs in environmental microbiology was the use of thismethod by Wu and Liu (41) for multiplex detection of differentBacteroides spp. This study also addressed different method-ological issues and aspects, such as the effects of noncomple-mentary tail length, annealing temperature, cycle number, andprimer-to-template ratio on extension efficiency. In a previousstudy, Nikolausz et al. (19) reported development and appli-cation of a multiplex SNuPE assay for detection and typing of“Dehalococcoides” sp. sequences obtained from chloroethene-contaminated groundwater samples. However, there still hasnot been a systematic evaluation of factors that affect primerdesign and the discriminatory power of primer extension.Moreover, quantitative aspects of the method have not beenthoroughly addressed so far.

The present study focused on these crucial issues by in-vestigating the effects of the type, number, and position ofprimer mismatches on the extension efficiency and hence thespecificity. Furthermore, quantitative aspects of SNuPE

were investigated in a model community experiment by us-ing defined template mixtures of 16S rRNA gene PCR prod-ucts or reverse-transcribed RNA.

MATERIALS AND METHODS

Bacterial strains and cultivation. The following strains were used in this study:Pseudomonas putida mt-2, Pseudomonas stutzeri DSM50238, Thauera aromaticaK172, and Bacillus subtilis DSM402. Strains were cultivated on nutrient agar(DSMZ medium 1).

DNA and RNA isolation, reverse transcription-PCR, PCR, and nucleic acidquantification. DNA was isolated from strains using a DNAeasy tissue kit (Qia-gen, Hilden, Germany) according to the manufacturer’s instructions for gram-

FIG. 1. Detection of target sequences obtained from strains of T. aromatica (Ta), P. putida (Pp), and B. subtilis (Bs) via hybridization with specificprimers and labeling by SNuPE. The AmpliTaq FS polymerase incorporates a labeled dideoxynucleotide, which terminates the reaction and results ina fluorescently labeled product. Multiplex reactions are carried out with different target-specific primers that also differ in length. The primers andproducts are subsequently separated by capillary electrophoresis, and the extended products are detected with laser-induced fluorescence.

TABLE 1. Primers used to assess the effects of mismatch type,position, and number on the primer extension efficiency

Primer Sequencea

PAE997rc........................5� (T)9 TGC AGA GAA CTT TCC AGA 3�

PAE997_mm1 ................5� (T)9 TGC AGA GAA CTT TCC AGC 3�

PAE997_mm2a ..............5� (T)9 TGC AGA GAA CTT TCC AGG 3�PAE997_mm2b ..............5� (T)9 TGC AGA GAA CTT TCC GGA 3�PAE997_mm2c...............5� (T)9 TGC AGA GAG CTT TCC AGA 3�PAE997_mm2d ..............5� (T)9 TGC GGA GAA CTT TCC AGA 3�PAE997_mm2e ..............5� (T)9 TGC GGA GAG CTT TCC AGA 3�PAE997_mm2f ...............5� (T)9 TGC AGA GAG CTT TCC GGA 3�PAE997_mm3a ..............5� (T)9 TGC AGA GAA CTT TCC AGT 3�

PAE997_mm3b ..............5� (T)9 TGC AGA GAA CTT TCC TGA 3�PAE997_mm3c...............5� (T)9 TGC AGA GAT CTT TCC AGA 3�PAE997_mm3d ..............5� (T)9 TGC TGA GAA CTT TCC AGA 3�PAE997_mm3e ..............5� (T)9 TGC TGA GAT CTT TCC AGA 3�PAE997_mm3f ...............5� (T)9 TGC AGA GAT CTT TCC TGA 3�

PAE997_BTEX..............5� (T)9 TCC AAT GAA CTT TCT AGA

a The mismatch positions are indicated by bold type.

VOL. 75, 2009 SNuPE FOR DETECTION AND TYPING OF PHYLOGENETIC MARKERS 2851


http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/

positive cells. DNA was eluted in 120 �l of RNase-free distilled water. RNA wasisolated from strains using an RNeasy mini kit (Qiagen) as described previously(20) and was quantified by using the Ribo Green RNA quantification reagent(Molecular Probes, Eugene, OR) along with the rRNA standard supplied withthe kit. Fluorescence was determined using a fluorimeter (Viktor 2 1420 multi-label counter; Wallac, Finland). Each PCR was performed using a 50-�l (finalvolume) mixture with HotStar Taq polymerase with the supplied buffer (Qiagen)on a Mastercycler gradient (Eppendorf, Hamburg, Germany) and primers 27F(13) and 1378R (8). The conditions used for PCR amplification of the 16S rRNAgenes were as follows: initial denaturation at 95°C for 15 min, followed by 30cycles of denaturation for 30 s at 95°C, primer annealing at 52°C for 30 s, andchain extension at 72°C for 70 s and then a final extension at 72°C for 30 min.

PCR products were purified (QIAquick PCR purification kit; Qiagen) andquantified by using the Pico Green double-stranded DNA quantification reagent(Molecular Probes) and a lambda phage DNA standard supplied with the kit.Unincorporated primers and deoxynucleoside triphosphates from 42 �l of puri-fied PCR products were treated with 12 U shrimp alkaline phosphatase (SAP)(Fermentas, Vilnius, Lithuania) and 6 U exonuclease I (Fermentas) in the SAPbuffer provided in a 60-�l (final volume) mixture. Reaction mixtures were mixedand incubated at 37°C for 1 h, which was followed by enzyme inactivation at 75°Cfor 15 min.

RNA isolated from different strains were mixed at defined ratios and subjectedto reverse transcription. The reaction was carried out using an Omniscript re-verse transcription kit (Qiagen) with primer 1378R in a 20-�l (final volume)mixture according to the manufacturer’s instructions, including a denaturation

step at 65°C for 5 min. The amount of RNA used per reaction ranged from 60ng to 180 ng. Reverse transcription was carried out at 42°C for 2 h. Treatment ofthe cDNA was performed by using only SAP as described above.

Primer design. Oligonucleotides that were specific for the genera of the ref-erence strains and targeted 16S rRNA were selected using the search engine anddatabase of probeBase (http://www.microbial-ecology.net/probebase/) (16). Thespecificities of the primers were further evaluated by using the Probe Matchfunction of Ribosomal Database Project II (5). The reverse complements of theoriginal probes were hybridized to reverse-transcribed RNA. The target posi-tions and sequences of the primers used in the model community experimentsare shown in Fig. 1 and 4 and Table 1.

SNuPE. Cyclic primer extension reactions were performed in 10-�l (finalvolume) mixtures containing 5 �l of SNaPshot multiplex kit reagent (AppliedBiosystems), 4 �l of purified PCR products or cDNA, and 1 �l of primer solutionor primer mixture (10 �M of each primer). The SNuPE reactions were carriedout using 35 cycles of denaturation at 96°C for 10 s, annealing at 55°C for 5 s, andextension at 60°C for 30 s. The volume of PCR products added and the annealingtemperature were different in different experiments. In order to remove unin-corporated ddNTPs, 1 U of SAP was added to each reaction mixture andincubated at 37°C for 1 h, which was followed by inactivation at 75°C for 15 min.The copy number of the PCR product for each SNuPE reaction was calculatedbased on the measured concentration and average length of the amplicons.SNuPE reactions were performed in duplicate to ensure reproducibility, while atleast triplicate reactions were carried out when peak area ratios were calculated;100 relative fluorescence units was used as the detection limit.

FIG. 2. Effects of terminal and internal primer mismatches at annealing temperatures of 50°C (A) and 55°C (B) on the specificity of the primerextension assays. The P. putida PCR product was targeted with perfectly matching primer PAE997rc (panels a and f). Mismatch discrimination wasachieved by replacement of the perfectly matching terminal base adenine by cytosine (panels b and g), guanine (panels c and h), or thymine (panelsd and i). Primer PAE997_BTEX designed to detect Pseudomonas veronii had four internal mismatches compared with the target DNA (panels eand j). The vertical axis indicates fluorescence intensity (in relative fluorescence units); the horizontal axis indicates the sizes of the extendedproducts. The numbers indicate the lengths (in bases) of the internal standards (black peaks).

2852 NIKOLAUSZ ET AL. APPL. ENVIRON. MICROBIOL.


http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/

Capillary electrophoresis and analysis. A 0.5-�l portion of posttreated exten-sion products was mixed with 9 �l of formamide and 0.5 �l of the GeneScan-120LIZ internal size standard (Applied Biosystems). The mixture was denatured at95°C for 5 min and quickly cooled on ice. DNA fragment separation was per-formed with an ABI PRISM 3100 genetic analyzer using a 36-cm capillary filledwith denaturing POP6 polymer with an E5 filter set (Applied Biosystems). Thefollowing parameters were used for electrophoresis: injection time, 10 s; elec-trophoresis voltage, 15 kV; run temperature, 60°C; and run time, 24 min. Dataanalysis was performed using the GeneMapper software (version 3.5; AppliedBiosystems). Peak areas were normalized based on the average signal strength ofthe applied internal standard (35 nucleotides), and the area data were used forcomparison of different samples.

Statistical analysis of single and multiplex SNuPE reaction data. The robust-ness of the single SNuPE assay was assessed by measuring the extents of thefluorescent signals generated (peak areas) with dilution series of the PCR prod-ucts from the different strains (T. aromatica, P. putida, and B. subtilis). Ninety-fivepercent confidence intervals were determined for the slopes of the correspondinglinear regressions. The linear models were considered to significantly fit the dataat a P value of �0.001 using F statistics (analysis of variance). All data werefound to significantly fit the linear models. The linear models were further usedas calibration curves to evaluate defined template mixtures of PCR amplicons ofthe three strains at various ratios in multiplex SNuPE assays. The calculatedfrequency distributions of various ratios of the PCR products were compared tothe distribution of the retrieved ratios of the mixed target templates in themultiplex SNuPE assays using a one-dimensional �2 “goodness-of-fit” test at asignificance threshold (�) of 0.5.

RESULTS AND DISCUSSION

Specificity of the SNuPE assay and effect of primer mis-match. One of the most important requirements for an effi-cient and reliable molecular detection technique is the abilityto specifically detect one type of sequence in the presence ofclosely related types of sequences. Our previous study showedthat discrimination between closely related sequences could beachieved with two mismatches in the SNuPE primer annealingregion of the nontarget sequence (19). In order to furtherinvestigate the effect of the nature and position of mismatcheson the efficiency, specificity, and the discrimination potential ofSNuPE, 16S rRNA gene PCR amplicons of P. putida were usedas templates in single reactions. Instead of alteration of thetemplate used for the SNuPE reaction, the Pseudomonas-spe-cific primer (PAE997rc) was redesigned to obtain 14 furtherprimer variants (Table 1). Single and multiple base alterationswere introduced at different positions in the primer sequence.First, variants with primer-template mismatches at the 3� endof the primer sequence were synthesized, since primer exten-sion is expected to be most sensitive to a terminal mismatch. Inaddition, internal single and multiple mismatches were intro-duced close to the 3� end, as well as in the middle of theprimer, to investigate to what extent internal mismatches po-tentially affect the specificity of SNuPE.

In the first experiment, primer extension with single primer-template mismatches at the 3� end was investigated at twodifferent annealing temperatures (Fig. 2) since early studies ofthe fidelity of DNA polymerases indicated the importance ofmismatched primer termini (22, 40). SNuPE with the specificprimer resulted in a strong specific peak, whereas only very lowsignals were obtained with the mismatch primers. A higherannealing temperature (55°C) further decreased the very lownonspecific signals, as expected. Various non-Watson-Crickbase pairings destabilize duplex formation at different ratios.Previous studies showed that G-T and G-A mismatches arerelatively stable (11, 14), which is in good agreement with our

results. When the 3� end of the primer was changed to G,resulting in a terminal G-T mismatch, a product peak area thatwas 2.3% of the specific product peak area was obtained at anannealing temperature of 55°C. On the other hand, a terminal

FIG. 3. Effects of the type and position of (A) single strongly bind-ing (G-T), (B) single weak (T-T), and (C) double mismatches on theprimer extension efficiency. Primer sequences are shown in Table 1.For easier measurement of nonspecific extension products, 55 cycleswere used. The results are expressed as averages for triplicate exper-iments and are based on comparisons with the results for perfectlymatching primer extension. The error bars indicate standard devia-tions for average values. U.D., under the detection limit (100 relativefluorescence units).



http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/

T-T mismatch (very weak base pairs [11]) resulted in signalsunder the detection limit, similar to the results obtained withthe primer with four internal mismatches. This is in accordancewith several previous studies that showed the importance of a3� end mismatch for specific discrimination in primer extensionduring PCR amplification (2). However, there is little agree-ment on which 3� terminal primer-template mismatches aremore readily extended by Taq DNA polymerase. Kwok et al.(12) observed a minimal effect of T-G or T-T terminal mis-matches on the PCR product yield, while Okayama and co-workers have not obtained any PCR products with T-G andT-T terminal mismatches (21). The major difference betweenPCR and SNuPE is that SNuPE mismatch products are noteffective templates for the next cycles of enzymatic reactions,while the mismatch products introduced in the early PCRcycles are readily amplified in the later cycles.

In addition, little is known about the effect of internal mis-matches on primer extension. The effect of nonterminal andmultiple mismatches on the stability of primer annealing wasinvestigated in detail by introducing strong and weak mis-matches at different positions and in different numbers (Table1). In this second experiment, the cycle number was increasedfrom 35 to 55 in order to increase the probability of nonspecificextension events (2) and to facilitate further evaluation. Thepeak areas of the mismatched extension products were rela-tively small compared to the peak area of the perfectly match-ing product (Fig. 3). Three independent SNuPE reactions wereperformed in all cases, and the average values are shown in Fig.3. The 3�-terminal position again had a pronounced detrimen-tal effect on the extension efficiency, and there was a consid-erably larger signal in the case of the more strongly bindingG-T mismatch (Fig. 3A) than in the case of the T-T mismatch(Fig. 3B). The primer extension efficiency increased when themismatch was moved toward the 5� end of the primer se-quence. However, the opposite trend was observed with prim-ers having a mismatch at position �9 with respect to the3�-terminal position (i.e., the middle of the primer sequence).

A similar result was obtained by Bru and coworkers (4) whenthey investigated the effect of mismatch positions on real-timePCR efficiency. The mismatches closer to the 5� end did notsignificantly affect the PCR efficiency, while the mismatchescloser to the 3� end decreased the efficiency by 1 to 3 orders ofmagnitude. Mismatch positions in the middle part of theprimer had a detrimental effect on PCR efficiency greater thanthe general trend. This was hypothesized to occur as a result ofchanges in the secondary structure of the primer. On the otherhand, when the target DNA was altered instead of the primer,some base conversions in the middle of the primer binding siteof the template had a greater effect than expected from thetrend, which cannot be explained by alteration of the second-ary structure of the primer.

The specificity of the primer extension in both SNuPE andPCR depends on the primer annealing and the fidelity of theDNA polymerase. While the formation and the thermal sta-bility of a primer-target duplex are influenced mainly by mis-matches in the middle part, strand extension and hence PCRamplification efficiency are influenced more by mismatches atthe 3� end of the primer. Our results demonstrate that bothfactors are likely to be responsible for the discriminating powerof SNuPE, as shown in Fig. 3C. With two mismatches, oneclose to the 3� end and one close to the middle part of theprimer, no detectable signal was observed, not even in the caseof strongly binding G-T mismatches. Thus, two well-chosenprimer mismatch positions result in a high signal-to-noise ratioand prevent any nonspecific extension regardless of the mis-match type even under nonstringent conditions.

To demonstrate the efficient mismatch discrimination poten-tial of the SNuPE method, new assays were designed (Fig. 4) todiscriminate two closely related strains, a P. putida strain and aP. stutzeri strain. Primer Pseudo_SNP was designed to hybrid-ize to both sequences in the same target region, which, how-ever, could be extended with different signature nucleotidesspecific for each target strain. Primers Pp_SNP and Ps_SNPare specific for P. putida and P. stutzeri, respectively. Only one

FIG. 4. Primer design strategy for demonstration of the one-mismatch discrimination potential of SNuPE. Differences between the 16S rRNAgenes of two closely related Pseudomonas species (P. putida and P stutzeri) are highlighted with colors in the sequence alignment of the target regionof the gene. The colors are the colors assigned for the nucleotide analogues used in the SNuPE assays. The primer design benefited from the fidelityof nucleotide incorporation of the Taq polymerase (primer Pseud_SNP) or from the discrimination potential due to the detrimental effect of 3�end mismatch on extension efficiency (primers Pp_SNP and Ps_SNP). The expected products for both microorganisms are also shown. The colorsof the incorporated ddNTPs are the colors of the expected peaks.



http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/

nucleotide difference was targeted in this second assay, whilethe nucleotides that were potentially incorporated were alsodifferent in the two target sequences to allow unambiguousevaluation of the extended products. The single-nucleotideincorporation proved to be very specific, since a nonspecificproduct could not be detected when primer Pseudo_SNP wasused alone (Fig. 5A). When two specific primers (Pp_SNP andPs_SNP) and only one of the target DNA templates were usedin the extension reaction (Fig. 5B), only specific signals weredetected. The lack of even traces of a nonspecific signal indi-cates that in precise primer design absolute one-mismatch dis-crimination can be achieved. In this case this meant that therewas a perfectly matching 3�-terminal position with strong basepairing (C-G) compared with a weakly binding mismatch(C-A) or a moderate mismatch (G-T).

Assessment of sensitivity. Besides the specificity, the sensi-tivity of a detection method is a critical issue. The overallsensitivity of SNuPE strongly depends on the preceding PCRamplification. The sensitivity of the primer extension assay wasevaluated with a dilution series containing known quantities ofPCR products as templates and a mixture of three primers

(Fig. 1). As expected, only the specific signal was observed(data not shown). Moreover, the signal strength decreasedproportionately with decreasing template concentration. Thedetection limit was estimated to be between 2 � 106 and 2 �107 copies of PCR amplicons �l�1 for the P. putida targetsequence. The limits of detection were slightly higher for T.aromatica and B. subtilis and were estimated to be around1.4 � 108 and 6.2 � 107 copies of PCR amplicons �l�1, re-spectively. However, the overall sensitivity of the approach ismuch higher, since the method includes preceding PCR am-plification of the target sequence. Compared to ethidium bro-mide-stained agarose gels (separating approximately 1.5-kbpPCR amplicons), where bands containing as little as approxi-mately 10 ng of DNA can be detected (29), a detection sensi-tivity that is 2 to 3 orders of magnitude higher can be expected.The increased sensitivity of SNuPE compared to gel-baseddetection is even more pronounced when a shorter target se-quence is amplified, since primer extension depends only onthe copy number of the target, whereas nucleic acid stainingalso depends on the length of the sequence. This finding im-plies that lower cycle numbers can be used for preamplification

FIG. 5. Primer extension assays for discrimination of P. putida and P. stutzeri 16S rRNA gene sequences using either primer Pseud_SNP (A) ora mixture of primers Pp_SNP and Ps_SNP (B). (Panels a to c) SNuPE patterns obtained with 65 ng of P. stutzeri PCR product (a), 65 ng of P. putidaPCR product (b), or mixed PCR products from P. stutzeri and P. putida (16 and 48 ng per reaction, respectively) (c), using primer Pseud_SNP.(Panels e to g) SNuPE patterns obtained with 65 ng of P. stutzeri PCR product (e), 65 ng of P. putida PCR product (f), or mixed PCR productsfrom P. stutzeri and P. putida (48 and 16 ng per reaction, respectively) (g), using a mixture of primers Pp_SNP and Ps_SNP. Negative controls(panels d and h) contained sterile distilled water instead of a PCR product. The vertical axis indicates fluorescence intensity (in relativefluorescence units); the horizontal axis indicates the sizes of the extended products. The numbers indicate the lengths (in bases) of the internalstandards (orange peaks).



http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/

for SNuPE in order to avoid biases associated with high cyclenumbers (24). With more than 40 cycles of SNuPE, formationof side products and nonspecific extension products was ob-served in some cases (data not shown); therefore, the use ofhigher cycle numbers is not recommended. The sensitivity ofSNuPE is expected to increase with the development of moresensitive detection sensors, as well as with the improvement offluorophores and discovery of new fluorophores. In addition,primer extension was demonstrated to detect less than 0.1% ofthe target template in the presence of nontarget PCR ampli-cons (41) and was able to detect a minor target population inthe presence of dominant population at a ratio of 1:585 (19),which corresponds to good sensitivity and a good dynamicrange.

Single and multiplex SNuPE reactions with mixed targettemplates. Experiments were designed to evaluate SNuPE formultiplex detection of members of an artificial communityestablished by mixing PCR products of 16S rRNA genes de-rived from target strains. Figure 1 shows the target sequencesof the microorganisms (T. aromatica, P. putida, and B. subtilis)used in this study aligned with the specific primers designed fordetection. In the presence of four differently labeled ddNTPs,the type of the labeling provides information about the baselocated downstream of the hybridization site (Fig. 1.). This mayprovide extra information for additional phylogenetic typing ofthe detected sequence (19). In the first step, single reactionsusing only one type of purified PCR product of a target mi-croorganism were performed. No apparent cross-reactivity wasobserved for the target-specific primers, and primer extensionevents were confirmed by detection of incorporated fluores-cent dye (data not shown). In each case, the expected nucleo-tide was incorporated, which indicates that there was a highlevel of specificity for each target sequence tested. The threedifferent taxon-specific primers also varied in length due toattachment of a noncomplementary tail (polythymidine) to the5� end of the primers. Polythymidine tails were used as non-acid-labile mobility modifiers, since a previous study showedthat these tails were more stable than, e.g., polyadenine tails(37). The extended products were subsequently separated bycapillary electrophoresis, and the incorporated label was de-tected by laser-induced fluorescence. Including internal stan-dards labeled with a fifth color allowed further verification of

the specificity by very accurate determination of the sizes of theextended products.

Furthermore, the quantitative power of the SNuPE assaywas investigated by mixing template PCR products at definedratios. Because the single reactions had revealed differencesin the extension efficiencies of the different template-primerpairs, calibration of the SNuPE was performed. Dilution seriesof the PCR products from the different strains (T. aromatica, P.putida, and B. subtilis) were used to obtain calibration curvesfor the extension efficiencies (Fig. 6). Subsequently, definedtemplate mixtures containing PCR amplicons of the threestrains at various ratios were used for multiplex SNuPE. Theresults of a quantitative analysis of the resulting peak areaswere compared to the relative abundance of the target tem-plates in the mixture by taking into account the different cali-bration curves (Table 2). The �2 test revealed no statisticallysignificant differences between the frequency distribution ofratios corresponding to the added PCR products and the dis-tribution of ratios retrieved for the mixed target templates inthe corresponding multiplex SNuPE assays (P 0.95). Thisemphasizes the similarity between the observed and expectedratio patterns at a high level of confidence. To investigate indetail the relationships between the theoretical and retrievedratios, seven different template mixtures of T. aromatica and B.subtilis PCR products (ratios of T. aromatica products to B.subtilis products of 10, 5, 3, 1, 0.3, 0.2, and 0.1) were usedseparately in duplicate in multiplex assays (Fig. 7). No statis-tically significant differences were found between the distribu-tion of theoretical ratios and the distribution of retrieved ratiosfor each of the two replicate experiments (P 0.7). Further-more, the fit of the linear regression model with the data washighly significant (P � 0.001), with a coefficient of determina-tion (R2) of 0.99, whereas the slope of the regression line wasaround 0.7. Overall, this indicates that there was high goodnessof the fit between the theoretical ratio values and the expectedratio values retrieved for the multiplex assay reactions up to aT. aromatica/B. subtilis threshold ratio of 5. For higher T.aromatica/B. subtilis ratios the experimental ratio values tendto underestimate the expected ratios. Rehybridization of theproducts (single-nucleotide extended primers) to the targetmay interfere with the primer binding and extension. Thiscompetition may particularly affect the predominant target

FIG. 6. Calibration of the extension efficiency for different primer-template pairs. The fluorescent signal generated by primer extension(normalized peak area) is expressed as a function of the amount of template PCR products (copy number) obtained from (A) B. subtilis DNA,(B) T. aromatica DNA, and (C) P. putida DNA. The dashed lines indicate the ower 95% confidence intervals of the regression curves.



http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/

sequence in the later cycles, as the detection of the less abun-dant template is not significantly influenced. This phenomenonmight be similar to the “C0t effect” described for PCR, result-ing in overrepresentation of the rare template types in the finalPCR products (18).

The results presented here suggest that SNuPE can be usedas a tool for multiplex quantification of different PCR ampli-cons in a certain range. However, due to the well-known biasinherent in PCR (33), current SNuPE protocols cannot beused for absolute quantification of microorganisms. Neverthe-less, our results suggest that SNuPE can be used as a semi-quantitative tool for following spatial and temporal changes inthe relative abundance of target microorganisms in a given setof samples. In a previous study, Wu and Liu used a differentstrategy for normalization of primer extension reactions (41).

Calibration factors for individual primers were obtained bycomparison of the extension efficiency to the results obtainedwith a primer at a higher hierarchical level (e.g., domain spe-cific). This primer was included in all reaction mixtures andwas used as an internal standard for comparison of Bacteroidespopulations in a wastewater treatment plant, in stools (10), andin wastewater samples (41).

Use of reverse-transcribed 16S rRNA as a template in aPCR-independent manner. Until now, SNuPE had to be cou-pled with PCR preamplification of the target gene for screen-ing environmental samples due to limitations in sensitivity. Ingeneral, quantitative biases are associated with the PCR ap-proach due to preferential amplification (33, 39) and differ-ences in the genomic properties, such as genome size, the copynumber of 16S rRNA genes, and the G�C content of the

FIG. 7. Experimental template mixture ratios (observed) expressed as a function of the defined template mixtures ratios (theoretical) of T.aromatica (Ta) and B. subtilis (Bs) PCR products (at defined ratios of T. aromatica to B. subtilis of 0.1, 0.2, 0.3, 1, 3, and 5). For informationconcerning the experimental setup see Table 2. The dashed lines indicate the 95% confidence intervals of the regression curves.

TABLE 2. Quantitative analysis of multiplex SNuPE reactions for simple model communities consisting of mixtures of PCR amplicons fromstrains of P. putida, T. aromatica, and B. subtilis

Overall amt of templateper 10-�l reaction

mixture (ng)

No. of measurementsper assay

Theoretical ratios of the different membersof the artificial community Ratio observed by SNuPEa

T. aromatica/B. subtilis

T. aromatica/P. putida

B. subtilis/P. putida

T. aromatica/B. subtilis T. aromatica/P. putida B. subtilis/P. putida

7.78 4 1.17 0.79 0.035.96 4 2.38 1.70 0.128.09 4 3.45 2.30 0.096.26 4 6.10 7.80 0.696.93 4 1.17 4.76 4.06 0.78 0.03 3.91 0.32 5.06 0.567.48 4 2.93 11.90 4.06 2.00 0.04 10.24 0.96 5.13 0.567.02 4 0.47 4.76 10.16 0.31 0.02 4.14 0.09 13.32 0.98

29.57 8 10.00 4.97 0.0916.13 6 5.00 3.55 0.0310.75 4 3.00 1.78 0.035.38 8 1.00 0.58 0.0010.75 8 0.33 0.22 0.0116.13 8 0.20 0.16 0.0129.57 8 0.10 0.14 0.00

a The values are the means standard deviations for the numbers of measurements indicated.



http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/

FIG. 8. Use of reverse-transcribed rRNA as a template for SNuPE. The RNA used for the reverse transcription reactions were as follows:(A) total RNA from B. subtilis (60 ng), (B) total RNA from P. putida (60 ng), (C) a 2:1 mixture of total RNA from P. putida and B. subtilis (120ng and 60 ng, respectively), (D) a 1:1 mixture of total RNA from P. putida and B. subtilis (60 ng and 60 ng, respectively), and (E) a 1:5 mixtureof total RNA from P. putida and B. subtilis (12 ng and 60 ng, respectively). (F) Negative control containing sterile distilled water instead of cDNA.

2858


http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/

target microorganisms (6). As an alternative approach, rRNAwas considered as a potential naturally amplified (i.e., abun-dant) target. Bacterial cells harbor around 103 to 105 ribo-somes per cell (1), which may suggest that the predominantmembers of a microbial community can be targeted by SNuPEin a PCR-independent manner. Since rRNA is not a templatefor the enzymatic system used for SNuPE reactions, RNA wasfirst reverse transcribed, and the cDNA was used for furtherreactions. Different ratios of total RNA were used to obtainmixtures of P. putida and B. subtilis RNA, as shown in Fig. 8.A decrease in the P. putida-specific signal corresponded pro-portionally to a decrease in the P. putida RNA concentration inpreparations with constant amounts of B. subtilis RNA. How-ever, the P. putida-specific peak area was approximately twicethe B. subtilis-specific product area when the RNA templateswere used at a P. putida/B. subtilis ratio of 1:5.

Although this result demonstrates the potential of SNuPE totarget reverse-transcribed rRNA, it also illustrates that thePCR-independent procedure does not necessarily result in asolution without bias. The Pseudomonas rRNA was preferen-tially detected, and use of this rRNA resulted in signal strengththat was significantly higher than the signal strengths obtainedwith the other templates (Fig. 7). When P. putida RNA wasmixed with T. aromatica RNA, even more pronounced prefer-ential detection of the P. putida target was observed (data notshown). The secondary structure of the cDNA or primer an-nealing of the reverse transcript may significantly influence theresults of PCR-independent RNA SNuPE, which should beinvestigated further.

In conclusion, SNuPE has excellent mismatch discriminationpotential, and the linear signal amplification contributes to theincreased sensitivity of detection. Since SNuPE reactions takeplace in solution, they are easy to optimize by altering theannealing temperature, cycle number, and template concen-tration using widely used thermocyclers. Not only PCR prod-ucts but also reverse-transcribed RNA are suitable templatesfor primer extension, as long as a sufficient number of variablesites are present in the phylogenetic or functional marker geneanalyzed. The separation and detection of the products usingcapillary electrophoresis and laser-induced fluorescence arevery fast and can be highly automated without a need forantigen-antibody reactions and lengthy posthybridizationsteps. Although there is a limit to the number of primers usedper reaction (38), the primers can be grouped according to theoptimal annealing temperature, and several SNuPE reactionscan be run in parallel using a gradient thermocycler. There-fore, SNuPE has considerable potential in applied and en-vironmental microbiology for filling the gap between theperformance of FISH and the performance of oligonucleo-tide microarrays.

ACKNOWLEDGMENTS

This project was financially supported by the Helmholtz Centre forEnvironmental Research—UFZ. G.I. was supported by the MarieCurie Early Stage Training Project of the EU (AXIOM, contractMEST-CT-2004-8332).

We thank Ute Lohse for technical assistance with the capillaryelectrophoresis.

REFERENCES

1. Amann, R. I., W. Ludwig, and K.-H. Schleifer. 1995. Phylogenetic identifi-cation and in situ detection of individual microbial cells without cultivation.Microbiol. Rev. 59:143–169.

2. Ayyadevara, S., J. J. Thaden, and R. J. Shmookler Reis. 2000. Discriminationof primer 3�-nucleotide mismatch by Taq DNA polymerase during polymer-ase chain reaction. Anal. Biochem. 284:11–18.

3. Behrens, S., B. M. Fuchs, F. Mueller, and R. Amann. 2003. Is the in situaccessibility of the 16S rRNA of Escherichia coli for Cy3-labeled oligonucle-otide probes predicted by a three-dimensional structure model of the 30Sribosomal subunit? Appl. Environ. Microbiol. 69:4935–4941.

4. Bru, D., F. Martin-Laurent, and L. Philippot. 2008. Quantification of thedetrimental effect of a single primer-template mismatch by real-time PCRusing the 16S rRNA gene as an example. Appl. Environ. Microbiol. 74:1660–1663.

5. Cole, J. R., B. Chai, R. J. Farris, Q. Wang, A. S. Kulam-Syed-Mohideen,D. M. McGarrell, A. M. Bandela, E. Cardenas, G. M. Garrity, and J. M.Tiedje. 2007. The Ribosomal Database Project (RDP-II): introducingmyRDP space and quality controlled public data. Nucleic Acids Res. 35:D169–D172.

6. Farrelly, V., F. A. Rainey, and E. Stackebrandt. 1995. Effect of genome sizeand rrn gene copy number on PCR amplification of 16S rRNA genes from amixture of bacterial species. Appl. Environ. Microbiol. 61:2798–2801.

7. Fuchs, B. M., G. Wallner, W. Beisker, I. Schwippl, W. Ludwig, and R.Amann. 1998. Flow cytometric analysis of the in situ accessibility of Esche-richia coli 16S rRNA for fluorescently labeled oligonucleotide probes. Appl.Environ. Microbiol. 64:4973–4982.

8. Heuer, H., M. Krsek, P. Baker, K. Smalla, and E. M. H. Wellington. 1997.Analysis of actinomycete communities by specific amplification of genesencoding 16S rRNA and gel-electrophoretic separation in denaturing gradi-ents. Appl. Environ. Microbiol. 63:3233–3241.

9. Hommais, F., S. Pereira, C. Acquaviva, P. Escobar-Paramo, and E. De-namur. 2005. Single-nucleotide polymorphism phylotyping of Escherichiacoli. Appl. Environ. Microbiol. 71:4784–4792.

10. Hong, P. Y., J. H. Wu, and W. T. Liu. 2008. Relative abundance of Bacte-roides spp. in stools and wastewaters as determined by hierarchical oligonu-cleotide primer extension. Appl. Environ. Microbiol. 74:2882–2893.

11. Ikuta, S., K. Takagi, R. B. Wallace, and K. Itakura. 1987. Dissociationkinetics of 19 base paired oligonucleotide-DNA duplexes containing differ-ent single mismatched base pairs. Nucleic Acids Res. 15:797–811.

12. Kwok, S., D. E. Kellogg, N. Mckinney, D. Spasic, L. Goda, C. Levenson, andJ. J. Sninsky. 1990. Effects of primer template mismatches on the polymer-ase chain reaction: human immunodeficiency virus type-1 model studies.Nucleic Acids Res. 18:999–1005.

13. Lane, D. J. 1991. 16S/23S rRNA sequencing, p. 115–147. In E. Stackebrandtand M. Goodfellow (ed.), Nucleic acid techniques in bacterial systematics.Wiley, New York, NY.

14. Leontis, N. B., J. Stombaugh, and E. Westhof. 2002. The non-Watson-Crickbase pairs and their associated isostericity matrices. Nucleic Acids Res.30:3497–3531.

15. Li, Q., Z. Liu, H. Monroe, and C. T. Culiat. 2002. Integrated platform fordetection of DNA sequence variants using capillary array electrophoresis.Electrophoresis 23:1499–1511.

16. Loy, A., F. Maixner, M. Wagner, and M. Horn. 2007. probeBase—an onlineresource for rRNA-targeted oligonucleotide probes: new features 2007. Nu-cleic Acids Res. 35:D800–804.

17. Mackay, I. M. 2004. Real-time PCR in the microbiology laboratory. Clin.Microbiol. Infect. 10:190–212.

18. Mathieu-Daude, F., J. Welsh, T. Vogt, and M. McClelland. 1996. DNArehybridization during PCR: the ‘Cot effect’ and its consequences. NucleicAcids Res. 24:2080–2086.

19. Nikolausz, M., A. Chatzinotas, M. Palatinszky, G. Imfeld, P. Martinez, andM. Kastner. 2008. Single-nucleotide primer extension assay for detectionand sequence typing of “Dehalococcoides” spp. Appl. Environ. Microbiol.74:300–304.

20. Nikolausz, M., U. Kappelmeyer, A. Szekely, A. Rusznyak, K. Marialigeti,and M. Kastner. 2008. Diurnal redox fluctuation and microbial activity in therhizosphere of wetland plants. Eur. J. Soil Biol. 44:324–333.

21. Okayama, H., D. T. Curiel, M. L. Brantly, M. D. Holmes, and R. G. Crystal.1989. Rapid nonradioactive detection of mutations in the human genome byallele specific amplification. J. Lab. Clin. Med. 114:105–113.

22. Petruska, J., M. F. Goodman, M. S. Boosalis, L. C. Sowers, C. Cheong, andI. Tinoco. 1988. Comparison between DNA melting thermodynamics andDNA-polymerase fidelity. Proc. Natl. Acad. Sci. USA 85:6252–6256.

23. Piggee, C. A., J. Muth, E. Carrilho, and B. L. Karger. 1997. Capillaryelectrophoresis for the detection of known point mutations by single-nucle-otide primer extension and laser-induced fluorescence detection. J. Chro-matogr. A 781:367–375.

24. Qiu, X. Y., L. Y. Wu, H. S. Huang, P. E. McDonel, A. V. Palumbo, J. M.Tiedje, and J. Z. Zhou. 2001. Evaluation of PCR-generated chimeras, mu-



http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/

tations, and heteroduplexes with 16S rRNA gene-based cloning. Appl. En-viron. Microbiol. 67:880–887.

25. Rudi, K., S. L. Flateland, J. F. Hanssen, G. Bengtsson, and H. Nissen. 2002.Development and evaluation of a 16S ribosomal DNA array-based approachfor describing complex microbial communities in ready-to-eat vegetablesalads packed in a modified atmosphere. Appl. Environ. Microbiol. 68:1146–1156.

26. Rudi, K., T. Katla, and K. Naterstad. 2003. Multi locus fingerprinting ofListeria monocytogenes by sequence-specific labeling of DNA probes com-bined with array hybridization. FEMS Microbiol. Lett. 220:9–14.

27. Rudi, K., O. M. Skulberg, F. Larsen, and K. S. Jakobsen. 1998. Quantifica-tion of toxic cyanobacteria in water by use of competitive PCR followed bysequence-specific labeling of oligonucleotide probes. Appl. Environ. Micro-biol. 64:2639–2643.

28. Rudi, K., O. M. Skulberg, R. Skulberg, and K. S. Jakobsen. 2000. Applica-tion of sequence-specific labeled 16S rRNA gene oligonucleotide probes forgenetic profiling of cyanobacterial abundance and diversity by array hybrid-ization. Appl. Environ. Microbiol. 66:4004–4011.

29. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: alaboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, ColdSpring Harbor, NY.

30. Schonhuber, W., B. Fuchs, S. Juretschko, and R. Amann. 1997. Improvedsensitivity of whole-cell hybridization by the combination of horseradishperoxidase-labeled oligonucleotides and tyramide signal amplification. Appl.Environ. Microbiol. 63:3268–3273.

31. Scott, J. C., M. S. Koylass, M. R. Stubberfield, and A. M. Whatmore. 2007.Multiplex assay based on single-nucleotide polymorphisms for rapid identi-fication of Brucella isolates at the species level. Appl. Environ. Microbiol.73:7331–7337.

32. Settanni, L., and A. Corsetti. 2007. The use of multiplex PCR to detect and

differentiate food- and beverage-associated microorganisms: a review. J.Microbiol. Methods 69:1–22.

33. Sipos, R., A. J. Szekely, M. Palatinszky, S. Revesz, K. Marialigeti, and M.Nikolausz. 2007. Effect of primer mismatch, annealing temperature andPCR cycle number on 16S rRNA gene-targeting bacterial community anal-ysis. FEMS Microbiol. Ecol. 60:341–350.

34. Sokolov, B. P. 1990. Primer extension technique for the detection of singlenucleotide in genomic DNA. Nucleic Acids Res. 18:3671.

35. Syvanen, A. C., K. Aalto-Setala, L. Harju, K. Kontula, and H. Soderlund.1990. A primer-guided nucleotide incorporation assay in the genotyping ofapolipoprotein E. Genomics 8:684–692.

36. Trebesius, K., R. Amann, W. Ludwig, K. Muhlegger, and K. H. Schleifer.1994. Identification of whole fixed bacterial cells with nonradioactive 23SrRNA-targeted polynucleotide probes. Appl. Environ. Microbiol. 60:3228–3235.

37. Tully, G., K. M. Sullivan, P. Nixon, R. E. Stones, and P. Gill. 1996. Rapiddetection of mitochondrial sequence polymorphisms using multiplex solid-phase fluorescent minisequencing. Genomics 34:107–113.

38. Vallone, P. M., R. S. Just, M. D. Coble, J. M. Butler, and T. J. Parsons. 2004.A multiplex allele-specific primer extension assay for forensically informativeSNPs distributed throughout the mitochondrial genome. Int. J. Leg. Med.118:147–157.

39. von Wintzingerode, F., U. B. Gobel, and E. Stackebrandt. 1997. Determina-tion of microbial diversity in environmental samples: pitfalls of PCR-basedrRNA analysis. FEMS Microbiol. Rev. 21:213–229.

40. Wu, D. Y., L. Ugozzoli, B. K. Pal, and R. B. Wallace. 1989. Allele-specificenzymatic amplification of beta-globin genomic DNA for diagnosis of sicklecell anemia. Proc. Natl. Acad. Sci. USA 86:2757–2760.

41. Wu, J. H., and W. T. Liu. 2007. Quantitative multiplexing analysis of PCR-amplified ribosomal RNA genes by hierarchical oligonucleotide primer ex-tension reaction. Nucleic Acids Res. 35:e82.



http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/

evaluation of single-nucleotide primer extension for ... · potential of snupe for simultaneous...

Documents