• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of aemPermissionsJournals.ASM.orgJournalAEM ArticleJournal InfoAuthorsReviewers
Appl Environ Microbiol. Jan 2007; 73(2): 380–389.
Published online Nov 17, 2006. doi:  10.1128/AEM.01785-06
PMCID: PMC1796975

Influence of Dangling Ends and Surface-Proximal Tails of Targets on Probe-Target Duplex Formation in 16S rRNA Gene-Based Diagnostic Arrays[down-pointing small open triangle]

Abstract

Dangling ends and surface-proximal tails of gene targets influence probe-target duplex formation and affect the signal intensity of probes on diagnostic microarrays. This phenomenon was evaluated using an oligonucleotide microarray containing 18-mer probes corresponding to the 16S rRNA genes of 10 waterborne pathogens and a number of synthetic and PCR-amplified gene targets. Signal intensities for Klenow/random primer-labeled 16S rRNA gene targets were dissimilar from those for 45-mer synthetic targets for nearly 73% of the probes tested. Klenow/random primer-labeled targets resulted in an interaction with a complex mixture of 16S rRNA genes (used as the background) 3.7 times higher than the interaction of 45-mer targets with the same mixture. A 7-base-long dangling end sequence with perfect homology to another single-stranded background DNA sequence was sufficient to produce a cross-hybridization signal that was as strong as the signal obtained by the probe-target duplex itself. Gibbs free energy between the target and a well-defined background was found to be a better indicator of hybridization signal intensity than the sequence or length of the dangling end alone. The dangling end (Gibbs free energy of −7.6 kcal/mol) was found to be significantly more prone to target-background interaction than the surface-proximal tail (Gibbs free energy of −64.5 kcal/mol). This study underlines the need for careful target preparation and evaluation of signal intensities for diagnostic arrays using 16S rRNA and other gene targets due to the potential for target interaction with a complex background.

There are many factors that influence the hybridization signal intensity between long strands of DNA (referred to as targets) and oligonucleotide probes in microarray-based diagnostics and gene expression studies (14). Some of these factors include the characteristics of the fluorescent molecules used to label the target, the labeling efficiency, the length of the probe, the target secondary structure, and hybridization conditions (3, 4, 6, 9, 15, 22, 28, 33-35, 37). However, factors such as the interaction of dangling ends and surface-proximal tails of targets with background DNA have not yet been systematically characterized. In this study, the term “dangling end” refers to the sequence of target that extends beyond the distal end of the probe, and “surface-proximal tail” refers to the sequence of target that extends beyond the anchored end of the probe.

Target preparation and the resulting lengths, sequences, double- versus single-stranded compositions, positions of fluorescent labels, and secondary structures of the dangling ends and surface-proximal tails of targets influence the stability of duplex formation and the resulting signal intensity (25, 31, 44). The targets may also interact with other nontarget DNA sequences present in the background, previously referred to as “hitchhiking” (23). This is especially true for 16S rRNA gene-based diagnostic microarrays, where dangling ends and surface-proximal tails will generally represent conserved regions and may have greater potential to interact with the background DNA or other targets. Because the background will be sample specific, comparative evaluation of two samples on the same microarray using target mixtures that interact with each other may pose problems. Low reproducibility and false-positive and false-negative signals with 16S rRNA gene-based arrays are partly accredited to bias caused by the characteristics of target molecules (8, 17), and target-background interaction may be one of the main reasons for these phenomena. Poor statistical relationships between experimental and predicted signal intensities (from Gibbs free energy calculations) suggest that thermodynamic parameters between target-and-probe duplexes are not fully understood (26). Thus, an evaluation of the influence of dangling ends and surface-proximal tails on signal intensity and target-background interaction is needed in order to interpret the signal intensities from 16S rRNA gene-based diagnostic and microbial community analysis arrays. The interaction among probe, target, and background can be represented as Pn [left and right double arrow ] Tn [left and right double arrow ] Bn, where P represents probe, T represents target, B represents background, and n represents the length of the corresponding DNA sequence. A given background DNA sequence may also interact with the probe sequence (Pn [left and right double arrow ] Bn). However, this interaction was not considered significant in this study because signal due to background DNA when hybridized alone was absent for the targeted probes.

The objective of this study was to systematically examine the impact of dangling ends and surface-proximal tails of targets on hybridization signal intensity and to evaluate target-background interaction by (i) comparing products of unknown length generated by random primer labeling with targets of known length and (ii) examining interactions between well-defined targets and backgrounds. This study underlines the need for careful target preparation and evaluation of signal intensities for diagnostic arrays using 16S rRNA and other gene targets due to the potential for target interaction with a complex background.

MATERIALS AND METHODS

The influence of background on dangling ends and surface-proximal tails was evaluated using an oligonucleotide microarray containing 18-mer probes corresponding to the 16S rRNA genes of 10 waterborne pathogens and a number of synthetic and PCR-amplified gene targets (see the Supplemental material). Using a set of 45-mer synthetic targets and Klenow/random primer-labeled targets generated by PCR amplification (assumed to be approximately 200 to 1,500 bp long), the signal intensities were compared for a total of 95 probes targeting the 16S rRNA genes of 10 waterborne pathogens. Both types of targets were then spiked individually into a complex mixture of 16S rRNA genes to study the influence of the type of background DNA that will be present as a result of target amplification using universal primers and random primer labeling. Using a Cy3 end-labeled 59-mer sequence as a model background DNA sequence designed to complement the dangling ends of five different targets of variable length (and the associated Gibbs free energies), it was demonstrated that 7 to 12 bases of contiguous homology with the dangling end is sufficient to cause cross-hybridization. A second 59-mer sequence served as a model background designed to complement the surface-proximal tails of 11 different targets of various sequence similarities and Gibbs free energies. Sequences used for the model backgrounds were from the dangling ends of nontargeted 16S rRNA genes for which probes were also present on the array. Synthetic sequences used for the well-defined model background complemented the dangling ends of nontarget probes.

Preparation of target DNA.

The following fivedifferent types of targets were prepared: (i) a mixture of Klenow/random primer-labeled (with Cy5) targets prepared from 16S rRNA genes of 10 pathogens (TK-m); (ii) a mixture of Cy3 end-labeled 45-mer synthetic targets corresponding to 95 probes for the 10 pathogens (T45-m); (iii) 1 individual Cy3 end-labeled synthetic target of a length of 102 bases (TD) to examine dangling end length; (iv) 4 additional Cy3 end-labeled synthetic targets of lengths of 47, 60, 80, and 106 bases (Tn, where n is the length of the target) to study the effect of increasing the length of the dangling end; and (v) 11 individual Cy3 end-labeled synthetic targets of lengths of 89 to 102 bases (TS), which were used to study the effect of sequence similarity in the surface-proximal tails of targets. The sequences of dangling ends and surface-proximal tails for all the synthetic targets (T45-m, TD, Tn, and TS) matched the 16S rRNA gene sequences of the corresponding pathogens. The approach used to prepare these targets is summarized below.

Klenow/random primer-labeled target mixture of 16S rRNA genes of pathogens (TK-m).

Genomic DNA from 10 bacterial pathogens (Table (Table1)1) was used as the source of 16S rRNA genes. Escherichia coli, Legionella pneumophila, Pseudomonas aeruginosa, Salmonella enterica subsp. arizonae, S. enterica serovar Typhimurium, and Yersinia enterocolitica type strains were obtained from the American Type Culture Collection (ATCC, Manassas, VA) and grown according to the protocol provided. For Helicobacter pylori and Clostridium perfringens, only genomic DNA was obtained from the ATCC. Enterococcus faecalis and Campylobacter jejuni were kindly provided by Joan B. Rose and Vincent Young (both at Michigan State University), respectively. DNA from pure cultures and the environmental sample was extracted using a Promega Wizard DNA extraction kit (Promega, Madison, WI).

TABLE 1.
Identities of pathogens, numbers of probes, and Gibbs free energy ranges for the set of probes

The 16S rRNA gene was amplified from the respective genomic DNA using 27F (5′-AGAGTTTGATCMTGGCTCAG-3′) and 1525R (5′-AAGGAGGTGWTCCARCC-3′) primer pairs (18) and Platinum Taq DNA polymerase (Invitrogen, Carlsbad, CA). Thirty cycles of the following temperature program were used for amplification: denaturation at 94°C for 30 s, annealing at 52°C for 45 s, and elongation at 72°C for 90 s. All PCR amplicons were cleaned using a QIAGEN PCR clean-up kit (QIAGEN Inc., Valencia, CA).

The amplified 16S rRNA gene (250 ng) from each pathogen was individually labeled with Cy5 by use of a Bioprime DNA labeling kit (Invitrogen, San Diego, CA). Briefly, the protocol included a 90-min incubation of the amplicon with Klenow polymerase and 5:1 amino-allyl-dUTP:dTTP (Ambion, Austin, TX) followed by Cy5 labeling. All amino-allyl-dUTP-labeled products were cleaned using a QIAGEN PCR clean-up kit with modified phosphate wash buffer (5 mM K2HPO4, pH 8.0, 80% ethyl alcohol) and phosphate elution buffer (4 mM K2HPO4, pH 8.5). Cyanine dye was attached by incubating 3 to 5 μg of amino-allyl-dUTP-labeled DNA for 1 h in a 50:50 mixture of 0.1 M sodium carbonate buffer (pH 9.3) and N-hydroxysuccinimide ester Cy dye (prepared in fresh dimethyl sulfoxide). DNA product from dye coupling was cleaned using a QIAGEN PCR clean-up kit. A mixture of Cy5-labeled 16S rRNA genes was then prepared by mixing 10 pmol (157 to 357 ng DNA) of labeled product from each of the 10 pathogens. Klenow/random primer-labeled targets were estimated to be between 200 to 1,500 bp (as assessed by gel electrophoresis). All Klenow-labeled targets are expected to have multiple Cy5 labels. The specific activity of Cy dye was calculated by dividing the measured pmol of nucleotides by the measured pmol of Cy3/Cy5 dye for each target.

Mixture of 45-mer synthetic targets (T45-m).

A mixture of synthetic targets was designed and synthesized to complement a set of 95 probes (18-mer) in the middle region of the 45-mer targets (Fig. (Fig.1a).1a). Each 45-mer synthetic target had a 14-nucleotide surface-proximal tail at the 5′ terminus and a 13-nucleotide dangling end at the 3′ terminus. The sequences of both the overhanging ends matched the 16S rRNA gene sequence of the corresponding pathogen. The single-stranded mixture of sequences was synthesized at the University of Michigan by use of an in situ oligonucleotide synthesis technology (11, 12), end labeled with Cy3 at the 5′ terminus, and harvested from the solid substrate.

FIG. 1.
Klenow/random primer-labeled target mixture versus end-labeled 45-mer synthetic target mixture. (a) Experimental strategy. (b) Signal-to-noise ratio of synthetic targets T45-m ([open triangle]) and Klenow/random primer-labeled targets TK-m ([filled square]) at 26°C. ...

Synthetic target with dangling ends of variable lengths (TD).

A 102-mer synthetic target was designed to hybridize to nine different probes on the array (TD). After hybridization, TD was expected to have a dangling end of 8 to 16 bases at the 3′ terminus and a surface-proximal tail of 68 to 76 bases at the 5′ terminus, depending upon the probe to which it was hybridized (Fig. (Fig.2a).2a). TD was synthesized and singly end labeled with Cy3 at the 5′ terminus by IDT (Coralville, IA). A 59-mer defined background DNA (BD [described below]) was designed to complement the dangling end sequence from position 86 to 98 with 4- to 12-base stretches of continuous homology. The sequence similarity between the surface-proximal tail of TD and the BD was less than 25%.

FIG. 2.
Synthetic targets of increasing dangling end length. (a) Experimental strategy. (b) The signal-to-noise ratios of BD and TD at 26°C (y axis) are shown for each of nine separate probes offset by one nucleotide relative to the target, in order to ...

The dangling end interaction due to the length of overhang was studied further by use of four synthetic targets of various lengths, i.e., T47, T60, T80, and T106 (in Tn formulations, n refers to the length of the synthetic target; for example, T60 is 60 bases long). Each Tn hybridized to 10 probes that had been designed from the same 16S rRNA sequence from position 242 to 272 (starting at the 3′ end). Each Tn was synthesized and singly end labeled with Cy3 at the 5′ terminus by IDT (Coralville, IA), and resulted in 3′ dangling ends of 21, 34, 54, and 80 bases for the probe designed at start position 242 and of 8, 21, 41, and 67 bases for the probe designed at start position 254. At their 5′ termini, all four synthetic targets had surface-proximal tails of 8 to 21 bases, depending on the probes to which they were hybridized. BD complemented the dangling end of Tn by 0, 13, 33, and 59 bases.

Synthetic targets (TS) with 10 to 100% sequence similarity at the surface-proximal tail.

The effect of the sequence similarity of the surface-proximal tail of the target to a defined background DNA sequence was studied using 11 different targets (TS) of similar lengths (89 to 102 bases). Each TS was synthesized and singly end labeled with Cy3 at the 5′ terminus by IDT (Coralville, IA). All the targets had 18 bases complementary to 52 probes on the microarray (Fig. (Fig.3a),3a), a dangling end of 1 to 16 bases, and a surface-proximal tail of 67 to 76 bases depending on the probe and target. TS were designed to have 10 to 100% sequence similarity to a single background sequence (BS [described below]).

FIG. 3.
Synthetic targets with surface-proximal tail sequences complementary to background DNA. (a) Experimental strategy. (b) Mismatches, insertions, and deletions influencing Gibbs free energy between the BS and 11 different TS, resulting in a Gibbs free energy ...

Preparation of background DNA.

Two types of background DNA were used in this study: (i) a mixture of 16S rRNA genes obtained by PCR from an environmental sample referred to as complex background DNA (B16S), and (ii) two defined background DNA of known sequences (both were 59-mers and were referred to as BD and BS) with various potentials of interaction with dangling ends of TD and surface-proximal tails of TS. For TD and TS, synthetic sequences complemented the dangling ends of nontarget probes. These are described below (Table (Table2)2) .

TABLE 2.
Characteristics of various DNA sequences used to study the effect of dangling ends and surface-proximal tails on signal intensity

Complex background DNA (B16S).

A mixture of 16S rRNA genes amplified from genomic DNA from activated sludge from the wastewater treatment plant in East Lansing, MI, served as the source of complex background DNA. The genomic DNA was obtained using the protocol described above, except that a purification step was also performed using a Wizard DNA clean-up kit (Promega, Madison, WI). The PCR protocol used was similar to that used in the preparation of TK-m, but only 25 cycles were used, and the annealing temperature was set to 50°C for 45 s. The extent of sequence similarity of the complex background DNA to target sequences was unknown but was expected to be substantial (because of conserved regions of 16S rRNA genes) and a function of the lengths of background and target DNA.

Synthetic background DNA of defined length.

Two 59-mer synthetic defined background sequences (BD and BS) were designed to complement the dangling ends of TD and surface-proximal tails of TS (11 different TS sequences were used). The extent of sequence similarity for defined background DNA was carefully controlled to result in either continuous homologies of variable length at the dangling ends of TD or various sequence similarities at the surface-proximal tails of TS. For TD, the sequence of BD had 4 to 12 bases of continuous homology from position 86 to 98 considering the 5′ end as 0 (Table (Table22 and Fig. Fig.2a).2a). The Gibbs free energy between TD and BD varied between −3.2 and −14.9 kcal/mol for 4- to 12-base stretches of continuous sequence homology at the dangling end. The Gibbs free energy between BD and P18 (complementing the TD) was between −4.3 and −4.6 kcal/mol.

In comparison with Tn (T47, T60, T80, and T106), the sequence of BD had 0, 13, 33, and 59 bases matching within the dangling ends of Tn and a Gibbs free energy of −2.3 to −74.9 kcal/mol. The Gibbs free energy between BD and P18 (complementing the Tn) was between −3.2 and −5.3 kcal/mol.

The second defined background DNA sequence, BS, was designed to complement the surface-proximal tails of 11 different TS (as shown in Table Table22 and depicted in Fig. Fig.3a)3a) with various degrees of sequence similarity. The surface-proximal tails of TS had an 8- to 16-base overhang (duplex dependent) with 0% sequence homology to BS followed by a stretch of 59 bases with variable regions of contiguous perfect matches and sequence similarity with BS. This resulted in a sequence similarity of 0 to 90% on the whole surface-proximal tail and 10 to 100% sequence similarity considering only the 59 bases on the 5′ end of TS. The dangling end of TS had 1 to 16 bases depending on the probe to which it was hybridized. For all interactions between BS and TS, the Gibbs free energy was always less negative than −5.7 kcal/mol. The Gibbs free energy between BS and P18 (complementing the TS) was between −3.8 and −8.1 kcal/mol.

Probe design and microarray synthesis.

Probes were designed using a Perl script developed in-house that screened for all 18-mers of 16S rRNA genes that had at least two mismatches to every other sequence in the RDP-II (http://rdp.cme.msu.edu/). Screened probes for a given pathogen were ranked based on melting temperature and G+C content by following the parameters incorporated into OligoArray 2.0 probe design software (32). The final array included a set of 146 probes (P18) targeting 21 different 16S rRNA gene sequences and mismatch probes identical to the perfect match sequence except for 1 or 2 incorrect bases in the middle of the oligomer to measure the degree of cross-hybridization. Microarrays containing the above-described set of probes were synthesized by Xeotron Corporation, Houston, TX (now part of Invitrogen, Carlsbad, CA), using a proprietary in situ synthesis technology developed by the University of Michigan (11, 12).

Experimental approach. (i) Comparison of 45-mer synthetic target mixture to the Klenow/random primer-labeled target mixture.

Two-sample comparative hybridization, similar to that used in gene expression studies (5), was performed to evaluate the differences and similarities between synthetic targets and Klenow/random primer-labeled targets. A similar experiment was conducted to study the impact of the complex background 16S rRNA gene mixture on target signal intensity. A total of three comparative hybridizations were performed in triplicate (shown by double arrows in Fig. Fig.1a).1a). These were (i) T45-m versus TK-m, to compare the hybridization behaviors of two types of target mixtures; (ii) T45-m versus B16S, to evaluate the impact of background DNA on the 45-mer synthetic target mixture; and (iii) TK-m versus B16S, to evaluate the impact of background DNA on the Klenow/random primer-labeled target mixture. The complex background mixture B16S, BD, and BS was also hybridized alone to ensure that target signals for the probes to be tested did not cross-hybridize with various backgrounds. For the first of these hybridizations, 20 pmol (1,069 ng DNA) of T45-m was mixed with 100 pmol (2,481 ng DNA) of TK-m. For the second hybridization, 20 pmol (1,069 ng DNA) of Cy3-end-labeled T45-m was mixed with 200 pmol (6,550 ng DNA) of Cy5-labeled B16S. For the third hybridization, 100 pmol of Cy5-labeled TK-m was mixed with 200 pmol (6,550 ng DNA) of Cy3-labeled B16S. The last mixture was obtained by combining 10 pmol of labeled 16S rRNA gene from each of the 10 pathogens.

(ii) Evaluation of lengths and sequences of dangling ends of synthetic targets.

The effect of dangling end length was studied by varying the length of the continuous sequence homology between TD and BD from 4 to 12 bases as shown in Fig. Fig.2a.2a. The dangling end had the sequence 5′-(86 bases)-CGACTTGCATGTGTTG-3′, wherein nucleotides in boldface complement the sequence of BD. In order to generate the increasing dangling end overhang of TD from 8 to 16 bases, nine separate probes with increasing offsets relative to the target were constructed. The sequence of background DNA (BD) included the complementary sequence for the 12 bases of the dangling end to result in target-background interaction, which was expected to increase with an increase in dangling end length. The hybridization included 67 pmol of Cy3-labeled TD (3,557 ng) and 37 pmol of Cy5-labeled BD (1,540 ng).

The dangling end interaction due to Gibbs free energy was examined further using four synthetic targets with 100% sequence similarity and of various lengths (T47, T60, T80, and T106) along with a 59-mer BD that was designed to complement the dangling end of Tn by 0, 13, 33, and 59 bp (having Gibbs free energies of −8.8, −14.8, −43.8, and −74.9 kcal/mol, respectively). The BD for this examination was the same as before; however, Tn targeted 10 probes different than those targeted by TD. The hybridization included 52 to 84 pmol of Cy3-labeled Tn (1,522 to 4,277 ng) and 37 pmol of Cy5-labeled BD (1,540 ng).

(iii) Evaluation of sequence similarity in the surface-proximal tail of a synthetic target.

The dependence of sequence similarity at the surface-proximal tail of a synthetic target with background DNA was studied by synthesizing 11 different synthetic targets with TS of similar lengths (Fig. (Fig.3a)3a) and a new synthetic background DNA sequence (BS). The Gibbs free energies of TS sequences varied between −9.2 and −76.8 kcal/mol, with sequence similarity varying between 10 and 100% (Fig. (Fig.3a).3a). Each TS was end labeled with Cy3 at the 5′ end and complemented three to eight probes. The hybridization included a total of 160 pmol of Cy3-labeled TS (7,895 ng) and 30 pmol of Cy5-labeled BS (1,115 ng).

Hybridization and scanning.

Target and background mixtures were prepared in 100 μl of hybridization solution containing 35% formamide, 0.4% Triton X-100, and 6× SSPE. SSPE buffers were made from a stock of 18× SSPE, which is 2.7 M NaCl, 180 mM Na2PO4, 18 mM Na2EDTA (adjusted to pH 6.6 with HCl). The hybridization solution was heated at 95°C for 3 min, cooled on ice for 1 min, and passed through a 0.22-μm filter. All hybridizations were carried out in triplicate for 16 to 18 h at 20°C using an M-2 microfluidic station (Xeotron Corporation). A flow rate of 500 μl per min was used for the recirculation of hybridization solution through the microfluidic array during hybridization (42). After overnight hybridization at 20°C, the microarray was washed using wash buffer 2 (6× SSPE, 0.2% Triton X-100), wash buffer 4 (1× SSPE, 0.2% Triton X-100), and wash buffer 2 with no Triton X-100 in series for 2.2 min each (500 μl per min, 20°C). A nonequilibrium thermal dissociation approach adapted to the Xeotron platform was used in all hybridization experiments (8, 20, 39, 42). The protocol was based on earlier studies utilizing a dissociation curve approach for diagnostic arrays (8, 20, 39, 42). The microarrays were washed with a high-stringency wash buffer (10 mM Na2HPO4, 5 mM EDTA, pH 6.6; flow rate, 500 μl per min) for 2.2 min at increasing temperatures from 20 to 70°C at 2°C intervals. Experimental conditions have been examined previously for optimal specificity (42). Signal intensities were quantified after each wash by use of a GenePix 4000B non-confocal laser scanner (Axon Instruments, Inc, Foster City, CA) at a photomultiplier tube voltage setting of 650 V for Cy5 (635-nm laser) and 600 V for Cy3 (532-nm laser). For complex background DNA (B16S), these settings were 750 V (for Cy5) and 500 V (for Cy3). With these settings, photobleaching caused an average decrease in signal intensity of 0.85% for Cy5-labeled targets and 0.54% for Cy3-labeled targets between each of the 20 scans at 20°C.

Data analysis.

Data were analyzed using an XL script that imported raw GenePix dissociation curve data between 20 and 70°C into Excel and generated a sigmoid curve for each microarray feature. The signal-to-noise ratio (S/N) at a given temperature was calculated from three replicate hybridization experiments. Noise was defined as the average signal of twenty empty spots (containing linker chemistry and no probes). The average signal of 20 randomly selected nontarget probes (80.3 ± 10.6 arbitrary units [a.u.] with a 635-nm laser or 148.3 ± 16.7 a.u. with a 532-nm laser) was similar to the average signal of 20 empty spots (83.6 ± 6.4 a.u. with a 635-nm laser or 155.9 ± 10.6 a.u. with a 532-nm laser) when the microarray was hybridized with the complex background. Compared to results obtained for no background, the average noise increased slightly when a complex background DNA (B16S) was hybridized (1.19-fold for Cy3 and 1.54-fold for Cy5). For all experiments, an average S/N greater than 3.0 was considered positive to balance the maximum percentages of true-positive calls of targeted probes and true-negative calls of nontargeted probes. Figure Figure1c1c is presented as the S/N for the area under the dissociation curve to include variation in both dissociation and signal intensity. For statistical comparison of probe signal intensities when hybridization took place with Klenow/random primer fragmented and 45-mer targets, a two-tailed inference about differences in population means for independent samples was performed with a 95% confidence interval. The free energy of hybridization was computed using the two-state hybridization server (www.bioinfo.rpi.edu/) developed by Dimitrov and Zuker (7) with a value of 43°C instead of the actual 20°C. This value was chosen because the hybridization buffer contained 35% formamide, which is expected to destabilize duplexes equivalently to increasing the temperature by 21 to 25°C (2, 39).

RESULTS AND DISCUSSION

Klenow/random primer-labeled target mixture versus end-labeled 45-mer synthetic target mixture.

The hybridization behavior of Klenow/random primer-labeled target mixture compared to that of the 45-mer synthesized target mixture, plotted as S/N of T45-m and TK-m, is shown in Fig. Fig.1b.1b. Tested probes in Fig. Fig.1b1b were sorted based on the difference in S/N between T45-m and TK-m. Considering an S/N greater than 3 as a positive signal, 87 out of 95 tested probes yielded positive signals for T45-m while 91 probes gave positive signals for TK-m. Approximately 5.4% of nontargeted probes gave false-positive signals with T45-m, and 10.9% gave false-positive signals with TK-m. An amount of nonspecific hybridization with TK-m higher than that with single-stranded DNA targets has been reported previously (10). Most of the probes having similar S/N with both the targets generally had S/N of less than 10 (probes 21 to 46). The S/N for 73 out of the 95 probes differed significantly between the two target mixtures (95% confidence interval considering mean difference). There was no systematic pattern when signal intensities of all probes for a single 16S rRNA gene were compared for the two targets. Often the difference in S/N was as large as 33-fold (e.g., for probe numbers 80 to 95).

While a difference in signal intensity with various target mixtures is expected, higher signals with single-stranded targets than with Klenow/random primer-labeled targets have not been reported previously. More specifically, probes 1 through 30 had much higher S/N with T45-m than with TK-m, while the opposite was observed for probes 60 to 95. For 32 out of 95 probes, the S/N was significantly higher with T45-m than with TK-m (P ≤ 0.05). Dissimilar results were observed in a study by Franke-Whittle et al. (10), in which S/N was higher for a majority of duplexes targeted with Klenow fragmented targets than with those targeted with single-stranded targets. Higher S/N with Klenow fragmented targets is expected because longer dangling ends in TK-m provide the opportunity for the attachment of many more Cy5 dye molecules than what was obtained for T45-m, i.e., only one Cy3 dye molecule. The average specific activity for TK-m was 72 ± 20 nucleotides per Cy5 dye molecule (calculated using the arithmetic average of the specific activities of each of 10 targets that were labeled separately). This implies that the average number of Cy5 dye molecules was between 2 and 30 (assuming a target length of between 200 and 1,500).

It is evident from Fig. Fig.1b1b that normalization of S/N for the higher number of dye molecules in TK-m or dye bias would not result in a profile that overlaps completely with the S/N profile of T45-m. This implies that factors in addition to differences in label abundance and dye bias are responsible for the differences in S/N. These factors include (i) the secondary structure of longer dangling ends of TK-m making certain regions inaccessible for hybridization with probes (4, 19, 28); (ii) TK-m and T45-m displacement from probes due to complementing strands of double-stranded TK-m (25), known as a zipper effect (30); (iii) bias in Klenow labeling, as previously suggested by Tiquia et al. (38); (iv) probe sequence variability and linear probe influences on thermodynamic stability (31); (v) physical quenching of Cy dyes (6); and (vi) conserved regions on overhangs of hybridized TK-m interacting with complementary strands from one or more of the 10 organisms also targeted in the mixture (23), creating chains of targets. Interactions occurring in solution between 16S rRNA gene targets from various organisms may influence the spot signal intensity by competing with the targets bound to the probe.

Further experiments were performed to examine and quantify the extent of target-background interaction on the two dissimilar targets by spiking both TK-m and T45-m into background DNA. When the two target mixtures were spiked (separately) into a complex mixture of nontarget 16S rRNA genes (B16S), the S/N of TK-m changed significantly compared to the S/N of T45-m (Fig. (Fig.1c).1c). The x axis in Fig. Fig.1c1c represents the factor by which the S/N of the target mixture (T45-m or TK-m) is influenced by the background, B16S, and the y axis is the frequency of probes with a given influence factor. For a given probe, the influence factor was computed as the S/N in the absence of B16S divided by the S/N in the presence of B16S. For targets not impacted by the background DNA, this influence factor should be close to unity (implying no effect), and the spread of the bell-shaped curve should be minimal. For T45-m, the influence factor was close to 1, while for TK-m, it was approximately 4. Similarly, the spread of the T45-m influence curve was very small compared to that obtained for TK-m. A simple comparison of the areas under each influence factor curve indicated that the interaction of TK-m with the background was 3.7 times higher than that of T45-m (obtained by comparing the areas under the curve of TK-m and T45-m, which were 69.5 and 19 arbitrary units, respectively). A Wilcoxon signed-rank test was used to statistically conclude that the distribution of influence factor was greater for TK-m than for T45-m with a 95% confidence interval. Therefore, background significantly influences the signal intensity of TK-m and not that of T45-m. Since the same probes were targeted with both target mixtures, the influence of background on signal intensity is dependent on the characteristics of the target molecules and independent of the spatial region on the array. The B16S was also hybridized alone to ensure that targeted probes did not produce any signal with the background 16S rRNA gene mixture.

The influence of background on the signal intensity of double-stranded TK-m may be the result of target interactions influenced by complementary and nontarget strands in background DNA. In one study, greater hybridization efficiency was observed with single-stranded targets than with double-stranded targets (13). This is due to competition between complementary strands and probes with double-stranded DNA. Related studies have reported increased competition between the complementing strand and probe as the length of the dangling end increased, while increasing the length of the surface-proximal tail did not influence hybridization (25). Interactions between the conserved regions of target dangling ends and sequences from a second target have also been observed (23). These interactions were minimized by using single-stranded RNA instead of double-stranded DNA. If this observation can be extended to DNA sequences, the use of asymmetric PCR for target preparation may be advantageous in minimizing target-background interaction.

Target-background interaction on the dangling ends of the target may also influence signal intensity by affecting the secondary structure. As studied by Chandler et al. (4), signal intensity was substantially increased as the target secondary structure was relieved using a second probe in solution that annealed to the target dangling end. Interaction between the target dangling end and a background DNA sequence may have a similar effect. Peplies et al. (24) observed similar results for a majority of tested probes. However, the signal of some probes decreased with a second probe in solution, suggesting that the opening of a selected binding site may lead to a reorganization of one or more secondary structures in other target regions.

There are additional reasons for using shorter targets with diagnostic arrays. Shorter targets reduce intermolecular structures that occur more frequently with single-stranded targets than with stiffer double-stranded products (36). Lane et al. (19) used amplicons of various lengths and concluded that shorter fragments reduced false negatives due to the high level of intermolecular secondary structure in longer targets. Fluorophore position on a target DNA sequence can also influence signal intensity (44). Signal intensity decreases as the proximity of the fluorescent molecule from a probe target duplex increases. Thus, target preparation strategies should include amplifying shorter targets with specific asymmetric PCR or fragmenting longer amplicons to produce a range of desired lengths.

Wilson et al. (43) suggested using fragments under 100 bp rather than full-length amplicons to obtain a stronger hybridization signal. A number of fragmentation strategies have been described for obtaining a size range between 35 and 200 bases with pre- and postlabeled targets and asymmetric PCR-amplified products (1, 21, 27, 41). Subsequent studies comparing costs, efficiencies, and influences on signal intensity for some of these methods are currently under way in our laboratory. Target-target and target-background interactions are expected to be less for targets other than 16S rRNA genes, especially if the targets have less-conservedregions.

Synthetic targets of increasing dangling end length.

The effect of dangling end length was evaluated using a target with an increasing stretch of continuous homology. A Cy3 end-labeled synthetic 102-mer target (TD) was designed to hybridize to nine different probes on the array. After hybridization, TD was expected to have a dangling end of 8 to 16 bases (Fig. (Fig.2a).2a). A 59-mer synthetic background DNA sequence (BD) end labeled with Cy5 was designed to complement the dangling end sequence from position 86 to 98 with 4- to 12-base stretches of continuous homology. The signal-to-noise ratio of BD and TD at 26°C is shown for each of the nine separate probes (Fig. (Fig.2b).2b). In order to generate the increasing dangling end overhang on the TD, probes were offset by one nucleotide relative to the target. Considering an S/N greater than 3 as a positive signal, BD hybridized to the dangling end of TD only when the dangling end stretch of continuous homology was 7 bases or longer. Signal from BD was not observed when hybridized alone (i.e., without TD), and mismatch probes did not display signal for BD or TD, directly implicating that BD did not cross-hybridize with the probe. As evident by the lack of signal for probes with target dangling ends of 6 bases or less, interaction of the surface-proximal end of TD with BD was insignificant. The S/N of BD for 11- and 12-base overhangs was higher than the S/N given by TD. This may be due to dye bias between Cy5 and Cy3 and a higher specific activity of Cy5-labeled BD. The 59-mer BD may have resulted in a great percentage of dye-labeled product compared to that due to the 102-mer TD, as the purity of oligonucleotide synthesis decreases with increasing length.

The dangling end interaction due to the length of overhang was reexamined using four synthetic targets of various lengths (T47, T60, T80, and T106, together represented by Tn) along with the same 59-mer BD that was designed to complement the dangling end of Tn by 0, 13, 33, and 59 bp. Dangling end interaction occurred between T60, T80, and T106 and BD but not between T47 and BD (data not shown). The zero sequence homology between T47 and BD was the obvious reason for this lack of interaction.

Gibbs free energy calculations incorporate influences such as length, sequence similarity, secondary structures, and hairpin loops caused by near-perfect match sequences with insertions or deletions (16). It was used to examine interaction between TD, Tn and BD. For TD, an overhang with 4 sequential perfect match bases had a Gibbs free energy of −3.2 kcal/mol, while an overhang with 12 sequential perfect match bases had a Gibbs free energy of −14.9 kcal/mol. The S/N of BD became greater than 3 when the stretch of continuous bases on the dangling end was 7 or higher and Gibbs free energy was −7.6 kcal/mol or more negative. With the experimental conditions used in this study, target-background interaction on dangling end can occur with Gibbs free energy that is more negative than −7.6 kcal/mol. However, due to phenomena such as differences in intermolecular structures and competitive influences of complementary strands (28), the interaction observed with this free energy may not be extensible to double-stranded DNA targets. For calculating the Gibbs free energy between BD and TD, only the 3′ end of TD was considered to circumvent the influence of probes on dangling end interaction. Neglecting the presence of the probe, the Gibbs free energy of the duplex formed between TD and BD was −15.4 kcal/mol. Because the 5′ end of BD preferentially hybridizes to the 3′ dangling end of TD, the interaction occurs solely on the dangling end rather than on the surface-proximal end of TD or the probe.

The S/N of BD increased as the Gibbs free energy became more negative and the length of continuous homology on the TD dangling end increased. Gibbs free energies between BD and T47, T60, T80, and T106 were −2.3, −14.2, −43.8, and −74.9 kcal/mol, respectively. Previous studies have also reported successful hybridization to 9-mer probes, with a Gibbs free energy of between −6 and −10 kcal/mol producing weak signals and one between −8 and −12 kcal/mol producing strong signals (29). It should be noted that binding free energy has been shown to be a function of the surface material, the surface charge density, the length of the linker molecule, and other experimental conditions (40).

Synthetic targets with surface-proximal tail sequences complementary to background DNA.

The effect of the similarity of the sequence of the surface-proximal tail of the target to the sequence of defined background DNA was studied using 11 different targets (TS) of similar lengths (89 to 102 bases). All the targets had 18 bases complementary to 52 probes on the microarray (Fig. (Fig.3a),3a), a dangling end of 1 to 16 bases, and a surface-proximal tail of 67 to 76 bases, depending on the probe and target. Targets were designed to have 10 to 100% sequence similarity with BS. Probes were not offset (as with dangling end experiments); however, 11 targets were used, resulting in a Gibbs free energy of between −9.2 and −76.8 kcal/mol for the TS [left and right double arrow ] BS duplex (Fig. (Fig.3b).3b). The Gibbs free energy for P18 [left and right double arrow ] TS varied between −17.5 and −24.1 kcal/mol. Continuous stretches of 38 and 45 perfect matches at the surface-proximal tails of TS (with free energies of −64.5 to −65.9 kcal/mol) displayed target-background interaction (Fig. (Fig.3c),3c), while continuous stretches of 19 or fewer perfectly matching bases (with free energy of −42.1 kcal/mol or more positive) did not result in signals attributable to BS. Targeted probes did not display signal when BS was hybridized alone (i.e., without TS).

These results suggest that the Gibbs free energy requirement for the target-background interaction at the surface-proximal tail is different from that for a similar interaction at the dangling end. While a Gibbs free energy of −7.6 kcal/mol (for a stretch of 7 continuous perfect matches) was sufficient for the dangling end interaction, a Gibbs free energy more positive than −64.5 kcal/mol (including a stretch of 19 continuous perfect matches) on the surface-proximal tail did not display interaction. Further examination of the TS with a Gibbs free energy of −36.8 kcal/mol when hybridized with BS (19-base stretch of continuous homology) showed no secondary structure on the surface-proximal tail. Thus, external influences that are not incorporated into Gibbs free energy calculations may be causing greater destabilization on the surface-proximal end. A study examining the influence of target overhang length on hybridization efficiency showed that the dangling end of a target interacted with its complementary strand more than with the surface-proximal tail (25). It was suggested that this is because the dangling end is exposed to the liquid phase, providing better access to complementary strands in solution. On the other hand, the surface-proximal tail is closer to the surface, thus limiting the availability to other strands in solution and reducing the association with complementary strands. It is obvious that the difference between target-background interactions on surface-proximal tails and those on dangling ends requires further investigation.

Conclusions.

The lengths and sequences of surface-proximal tails and dangling ends of targets and the resulting target-background interactions influence signal intensity and decrease the specificity of oligonucleotide probe-target hybridizations. The dangling end was found to be significantly more prone to target-background interaction than the surface-proximal tail. Gibbs free energy between the target and background was found to be a better indicator of hybridization signal intensity than the sequence or length of the dangling end alone. This study underlines the need for careful target preparation (fragmentation and asymmetric PCR) and evaluation of signal intensities for diagnostic arrays by use of 16S rRNA and other gene targets due to the potential for target interaction with a complex background.

Supplementary Material

[Supplemental material]

Acknowledgments

We are grateful for the help from Jimmy Johnson and Weihong Qi with scripts for sequence analysis and to Alok Dhawan for critically reading the manuscript.

This work was supported by the National Institutes of Health (grant R01 RR018625-01) and Michigan Economic Development Corporation (GR-476 PO 085P3000517).

Footnotes

[down-pointing small open triangle]Published ahead of print on 17 November 2006.

Supplemental material for this article may be found at http://aem.asm.org/.

REFERENCES

1. Bekal, S., R. Brousseau, L. Masson, G. Prefontaine, J. Fairbrother, and J. Harel. 2003. Rapid identification of Escherichia coli pathotypes by virulence gene detection with DNA microarrays. J. Clin. Microbiol. 41:2113-2125. [PMC free article] [PubMed]
2. Blake, R. D., and S. G. Delcourt. 1996. Thermodynamic effects of formamide on DNA stability. Nucleic Acids Res. 24:2095-2103. [PMC free article] [PubMed]
3. Bodrossy, L., N. Stralis-Pavese, J. C. Murrell, S. Radajewski, A. Weilharter, and A. Sessitsch. 2003. Development and validation of a diagnostic microbial microarray for methanotrophs. Environ. Microbiol. 5:566-582. [PubMed]
4. Chandler, D. P., G. J. Newton, J. A. Small, and D. S. Daly. 2003. Sequence versus structure for the direct detection of 16S rRNA on planar oligonucleotide microarrays. Appl. Environ. Microbiol. 69:2950-2958. [PMC free article] [PubMed]
5. Churchill, G. A. 2002. Fundamentals of experimental design for cDNA microarrays. Nat. Genet. 32:490-495. [PubMed]
6. Cox, W. G., M. P. Beaudet, J. Y. Agnew, and J. L. Ruth. 2004. Possible sources of dye-related signal correlation bias in two-color DNA microarray assays. Anal. Biochem. 331:243-254. [PubMed]
7. Dimitrov, R., and M. Zuker. 2004. Prediction of hybridization and melting for double-stranded nucleic acids. Biophys. J. 87:215-216. [PMC free article] [PubMed]
8. El Fantroussi, S., H. Urakawa, A. E. Bernhard, J. J. Kelly, P. A. Noble, H. Smidt, G. M. Yershov, and D. A. Stahl. 2003. Direct profiling of environmental microbial populations by thermal dissociation analysis of native rRNAs hybridized to oligonucleotide microarrays. Appl. Environ. Microbiol. 69:2377-2382. [PMC free article] [PubMed]
9. Franke-Whittle, I. H., S. H. Klammer, and H. Insam. 2005. Design and application of an oligonucleotide microarray for the investigation of compost microbial communities. J. Microbiol. Methods 62:37-56. [PubMed]
10. Franke-Whittle, I. H., S. H. Klammer, S. Mayrhofer, and H. Insam. 2006. Comparison of different labeling methods for the production of labeled target DNA for microarray hybridization. J. Microbiol. Methods 65:117-126. [PubMed]
11. Gao, X. L., E. Gulari, and X. C. Zhou. 2004. In situ synthesis of oligonucleotide microarrays. Biopolymers 73:579-596. [PubMed]
12. Gao, X. L., E. LeProust, H. Zhang, O. Srivannavit, E. Gulari, P. L. Yu, C. Nishiguchi, Q. Xiang, and X. C. Zhou. 2001. A flexible light-directed DNA chip synthesis gated by deprotection using solution photogenerated acids. Nucleic Acids Res. 29:4744-4750. [PMC free article] [PubMed]
13. Guo, Z., R. A. Guilfoyle, A. J. Thiel, R. Wang, and L. M. Smith. 1994. Direct fluorescence analysis of genetic polymorphisms by hybridization with oligonucleotide arrays on glass supports. Nucleic Acids Res. 11:5456-5465. [PMC free article] [PubMed]
14. Hashsham, S. A., L. M. Wick, J. M. Rouillard, E. Gulari, and J. M. Tiedje. 2004. Potential of DNA microarrays for developing parallel detection tools (PDTs) for microorganisms relevant to biodefense and related research needs. Biosens. Bioelectron. 20:668-683. [PubMed]
15. Kane, M. D., T. A. Jatkoe, C. R. Stumpf, J. Lu, J. D. Thomas, and S. J. Madore. 2000. Assessment of the sensitivity and specificity of oligonucleotide (50mer) microarrays. Nucleic Acids Res. 28:4552-4557. [PMC free article] [PubMed]
16. Koehler, R. T., and N. Peyret. 2005. Thermodynamic properties of DNA sequences: characteristic values for the human genome. Bioinformatics 21:3333-3339. [PubMed]
17. Koizumi, Y., J. J. Kelly, T. Nakagawa, H. Urakawa, S. El-Fantroussi, S. Al-Muzaini, M. Fukui, Y. Urushigawa, and D. A. Stahl. 2002. Parallel characterization of anaerobic toluene- and ethylbenzene-degrading microbial consortia by PCR-denaturing gradient gel electrophoresis, RNA-DNA membrane hybridization, and DNA microarray technology. Appl. Environ. Microbiol. 68:3215-3225. [PMC free article] [PubMed]
18. Lane, D. J. 1991. 16S/23S rRNA sequencing, p. 115-148. In E. Stackebrandt and M. Goodfellow (ed.), Nucleic acid techniques in bacterial systematics. John Wiley, New York, NY.
19. Lane, S., J. Evermann, F. Loge, and D. R. Call. 2004. Amplicon secondary structure prevents target hybridization to oligonucleotide microarrays. Biosens. Bioelectron. 20:728-735. [PubMed]
20. Li, E. S. Y., J. K. K. Ng, J. H. Wu, and W. T. Liu. 2004. Evaluating single-base-pair discriminating capability of planar oligonucleotide microchips using a non-equilibrium dissociation approach. Environ. Microbiol. 6:1197-1202. [PubMed]
21. Longo, M. C., M. S. Berninger, and J. L. Hartley. 1990. Use of uracil DNA glycosylase to control carry-over contamination in polymerase chain reactions. Gene 93:125-128. [PubMed]
22. Loy, A., C. Schulz, S. Lucker, A. Schopfer-Wendels, K. Stoecker, C. Baranyi, A. Lehner, and M. Wagner. 2005. 16S rRNA gene-based oligonucleotide microarray for environmental monitoring of the betaproteobacterial order “Rhodocyclales.” Appl. Environ. Microbiol. 71:1373-1386. [PMC free article] [PubMed]
23. Palmer, C., E. M. Bik, M. B. Eisen, P. B. Eckburg, T. R. Sana, P. K. Wolber, D. A. Relman, and P. O. Brown. 2006. Rapid quantitative profiling of complex microbial populations. Nucleic Acids Res. 34:e5. [PMC free article] [PubMed]
24. Peplies, J., F. O. Glockner, and R. Amann. 2003. Optimization strategies for DNA microarray-based detection of bacteria with 16S rRNA-targeting oligonucleotide probes. Appl. Environ. Microbiol. 69:1397-1407. [PMC free article] [PubMed]
25. Peytavi, R., L.-Y. Tang, F. R. Raymond, K. Boissinot, L. Bissonnette, M. Boissinot, F. J. Picard, A. Huletsky, M. Ouellette, and M. G. Bergeron. 2005. Correlation between microarray DNA hybridization efficiency and the position of short capture probe on the target nucleic acid. BioTechniques 39:89-96. [PubMed]
26. Pozhitkov, A., P. A. Noble, T. Domazet-Loso, A. W. Nolte, R. Sonnenberg, P. Staehler, M. Beier, and D. Tautz. 2006. Tests of rRNA hybridization to microarrays suggest that hybridization characteristics of oligonucleotide probes for species discrimination cannot be predicted. Nucleic Acids Res. 34:e66. [PMC free article] [PubMed]
27. Ramakrishnan, R., D. Dorris, A. Lublinsky, A. Nguyen, M. Domanus, A. Prokhorova, L. Gieser, E. Touma, R. Lockner, M. Tata, X. Zhu, M. Patterson, R. Shippy, T. J. Sendora, and A. Mazumder. 2002. An assessment of Motorola CodeLinkTM microarray performance for gene expression profiling applications. Nucleic Acids Res. 30:e30. [PMC free article] [PubMed]
28. Ratushna, V. G., J. W. Weller, and C. J. Gibas. 2005. Secondary structure in the target as a confounding factor in synthetic oligomer microarray design. BMC Genomics 6:31. [PMC free article] [PubMed]
29. Reyes-Lopez, M. A., A. Mendez-Tenorio, R. Maldonado-Rodriguez, M. J. Doktycz, J. T. Fleming, and K. L. Beattie. 2003. Fingerprinting of prokaryotic 16S rRNA genes using oligodeoxyribonucleotide microarrays and virtual hybridization. Nucleic Acids Res. 31:779-789. [PMC free article] [PubMed]
30. Reynaldo, L. P., A. V. Vologodskii, B. P. Neri, and V. I. Lyamichev. 2000. The kinetics of oligonucleotide replacements. J. Mol. Biol. 297:511-520. [PubMed]
31. Riccelli, P. V., K. E. Mandell, and A. S. Benight. 2002. Melting studies of dangling-ended DNA hairpins: effects of end length, loop sequence and biotinylation of loop bases. Nucleic Acids Res. 30:4088-4093. [PMC free article] [PubMed]
32. Rouillard, J.-M., M. Zuker, and E. Gulari. 2003. OligoArray 2.0: design of oligonucleotide probes for DNA microarrays using a thermodynamic approach. Nucleic Acids Res. 31:3057-3062. [PMC free article] [PubMed]
33. Sengupta, S., K. Onodera, A. Lai, and U. Melcher. 2003. Molecular detection and identification of influenza viruses by oligonucleotide microarray hybridization. J. Clin. Microbiol. 41:4542-4550. [PMC free article] [PubMed]
34. Small, J., D. R. Call, F. J. Brockman, T. M. Straub, and D. P. Chandler. 2001. Direct detection of 16S rRNA in soil extracts by using oligonucleotide microarrays. Appl. Environ. Microbiol. 67:4708-4716. [PMC free article] [PubMed]
35. Southern, E., K. Mir, and M. Shchepinov. 1999. Molecular interactions on microarrays. Nat. Genet. 21:5-9. [PubMed]
36. Southern, E. M., S. C. Case-Green, J. K. Elder, M. Johnson, K. U. Mir, L. Wang, and J. C. Williams. 1994. Arrays of complementary oligonucleotides for analysing the hybridisation behaviour of nucleic acids. Nucleic Acids Res. 22:1368-1373. [PMC free article] [PubMed]
37. Taroncher-Oldenburg, G., E. M. Griner, C. A. Francis, and B. B. Ward. 2003. Oligonucleotide microarray for the study of functional gene diversity in the nitrogen cycle in the environment. Appl. Environ. Microbiol. 69:1159-1171. [PMC free article] [PubMed]
38. Tiquia, S. M., L. Wu, S. C. Chong, S. Passovets, D. Xu, Y. Xu, and J. Zhou. 2004. Evaluation of 50-mer oligonucleotide arrays for detecting microbial populations in environmental samples. BioTechniques 36:664-675. [PubMed]
39. Urakawa, H., P. A. Noble, S. El Fantroussi, J. J. Kelly, and D. A. Stahl. 2002. Single-base-pair discrimination of terminal mismatches by using oligonucleotide microarrays and neural network analyses. Appl. Environ. Microbiol. 68:235-244. [PMC free article] [PubMed]
40. Vainrub, A., and B. M. Pettitt. 2003. Surface electrostatic effects in oligonucleotide microarrays: control and optimization of binding thermodynamics. Biopolymers 68:265-270. [PubMed]
41. Wang, H., J. Li, H. Liu, Q. Liu, Q. Mei, Y. Wang, J. Zhu, N. He, and Z. Lu. 2002. Label-free hybridization detection of a single nucleotide mismatch by immobilization of molecular beacons on an agarose film. Nucleic Acids Res. 30:e61. [PMC free article] [PubMed]
42. Wick, L. M., J. M. Rouillard, T. S. Whittam, E. Gulari, J. M. Tiedje, and S. A. Hashsham. 2006. On-chip non-equilibrium dissociation curve and dissociation rate constant as methods to assess specificity of oligonucleotide probes. Nucleic Acids Res. 34:e26. [PMC free article] [PubMed]
43. Wilson, W. J., C. L. Strout, T. Z. DeSantis, J. L. Stilwell, A. V. Carrano, and G. L. Andersen. 2002. Sequence-specific identification of 18 pathogenic microorganisms using microarray technology. Mol. Cell. Probes 16:119-127. [PubMed]
44. Zhang, L., T. Hurek, and B. Reinhold-Hurek. 2005. Position of the fluorescent label is a crucial factor determining signal intensity in microarray hybridizations. Nucleic Acids Res. 33:e166. [PMC free article] [PubMed]

Articles from Applied and Environmental Microbiology are provided here courtesy of American Society for Microbiology (ASM)

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...