Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

The construction and use of bacterial DNA microarrays based on an optimized two-stage PCR strategy

The construction and use of bacterial DNA microarrays based on an optimized two-stage PCR strategy Background: DNA microarrays are a powerful tool with important applications such as global gene expression profiling. Construction of bacterial DNA microarrays from genomic sequence data using a two-stage PCR amplification approach for the production of arrayed DNA is attractive because it allows, in principal, the continued re-amplification of DNA fragments and facilitates further utilization of the DNA fragments for additional uses (e.g. over-expression of protein). We describe the successful construction and use of DNA microarrays by the two-stage amplification approach and discuss the technical challenges that were met and resolved during the project. Results: Chimeric primers that contained both gene-specific and shared, universal sequence allowed the two-stage amplification of the 3,168 genes identified on the genome of Synechocystis sp. PCC6803, an important prokaryotic model organism for the study of oxygenic photosynthesis. The gene-specific component of the primer was of variable length to maintain uniform annealing st temperatures during the 1 round of PCR synthesis, and situated to preserve full-length ORFs. Genes were truncated at 2 kb for efficient amplification, so that about 92% of the PCR fragments were full-length genes. The two-stage amplification had the additional advantage of normalizing the yield of PCR products and this improved the uniformity of DNA features robotically deposited onto the microarray surface. We also describe the techniques utilized to optimize hybridization conditions and signal-to-noise ratio of the transcription profile. The inter-lab transportability was demonstrated by the virtual error-free amplification of the entire genome complement of 3,168 genes using the universal primers in partner labs. The printed slides have been successfully used to identify differentially expressed genes in response to a number of environmental conditions, including salt stress. Conclusions: The technique detailed here minimizes the cost and effort to replicate a PCR- generated DNA gene fragment library and facilitates several downstream processes (e.g. directional cloning of fragments and gene expression as affinity-tagged fusion proteins) beyond the primary objective of producing DNA microarrays for global gene expression profiling. Page 1 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 PCC6803 was predicted to contain 3,168 open reading Background DNA microarrays have proven to be a particularly effective frames (prior to May 2002) http://www.kazusa.or.jp/ cyano/index.html[8]. Of these, over 50% (>1700 ORFs) tool for genome-wide transcript profiling, and they can generate information on the accumulation and disappear- have been characterized as hypothetical or having no sim- ance of thousands of specific transcripts in a single exper- ilarity to known sequences. Thus, differential gene expres- iment [1–5]. Because such an array generates vast sion via microarray analysis will represent an important quantities of data, it is desirable to improve the statistical functional genomics approach and we describe the two- reliability of the data through experimental repetitions. stage PCR strategy that led to the successful production of Additionally, it is useful to perform time course experi- full-genome microarray for Synechocystis . ments involving multiple time points to enable the appli- cation of cluster analysis to classify groups of genes Methods exhibiting similar temporal patterns of regulation in Growth conditions and RNA isolation response to shifts in environmental conditions. Therefore, Synechocystis PCC6803 cultures were grown autotrophi- a robust experimental design may involve the consump- cally in BG-11 medium essentially as described by Wil- tion of many individual DNA microarrays. Given these liams [9]. Specifically, 500 mL cultures were grown in flat considerations, it is important to have methods for (4 cm across the light path) Bellco tissue culture flasks at improving the efficiency and economy of microarray 30°C under white light fluorescent illumination (Cool production. White , General Electric) with an incident intensity of 80 -2 -1 µmoles photons m s . Aeration was provided by bub- Since the bacterial genes are typically not cloned at a bling with air enriched with 3% CO at a rate of 300 mL -1 genome-wide scale as non-redundant, ordered sets, as min . RNA was isolated using hot phenol conjunction with eukaryotic cDNA libraries, it is not possible to use with vigorous agitation with glass beads as described by 'universal primers' (e.g. T3, T7 sequences) and therefore Bhaya et al. [10]. PCR amplification requires design and synthesis of PCR primers for all target genes. In this context, the construc- Chimeric oligonucleotide design and PCR amplification tion of bacterial DNA microarrays from genomic Figure 1 provides a flow chart of the preparation of the full-genome Synechocystis microarray. Oligonucleotides, sequence data using a 'two-stage' approach, is attractive because it allows the continued re-amplification of DNA synthesized by Sigma Genosys http://www.sigmageno- fragments and facilitates further utilization of the DNA sys.com/, were designed to amplify the complete coding fragments for additional uses (e.g. over-expression of pro- region of each open reading frame from the sequenced tein) [6]. This approach involves the use of chimeric prim- genome of Synechocystis PCC6803. The only exceptions ers containing both gene-specific and engineered were instances for genes with a predicted length extending sequences (adaptamers). These primers are used to syn- beyond 2 kb. Of the 3,168 open reading frames, fewer thesize PCR products directly from genomic DNA, which than 10% (261) of the genes were truncated at the 3'-end serves as template during the first stage of PCR amplifica- to the 2 kb limit. Each of the oligos is a chimeric sequence tion. This amplification can be performed for thousands designed for first round synthesis and, as such, contained of genes during a given production run using well-estab- two components: a 5' universal sequence and a 3' gene- lished procedures that include microtiter plates and mul- specific sequence as shown in Figure 2. The gene-specific tichannel pipetting techniques to afford efficient parallel region of the oligonucleotides begins at the first or last processing of the samples. The products of this first stage base of the open reading frame and continues inside until of amplification are gene sequences flanked by the arbi- the T of the oligo was in the range of 58–62°C (25–45 trary adaptamer sequences engineered into the first stage nt, including the universal sequence). An algorithm PCR primers. The adaptamer sequences become the basis (Sigma Genosys) based upon Primer3 (Whitehead Insti- for re-amplification of the gene sequences in a second tute) was used for this purpose and resulted in a high fre- stage amplification. During the second stage of amplifica- quency of successful amplification as discussed below. tion, the first stage PCR products serve as template, Importantly, the universal sequences were designed to whereas a single pair of primers recognizing the flanking include SapI restriction sites in order to facilitate down- adaptamer sequence can be used to re-amplify all prod- stream applications, including cloning into an expression ucts of the first stage PCR. vector. Aliquots of the oligonucleotides were transferred and mixed as primer pairs to separate locations of 96-well In this report, we describe the procedures for the produc- plates prior to delivery from the commercial source, tion and use of DNA microarrays for the cyanobacterium which facilitated amplification in the 96-well format. The Synechocystis PCC6803. This organism is a widely used remaining products from the synthesis reaction were experimental model for the study of gene expression and stored in individual tubes at -20°C. plant-like oxygenic photosynthesis [7]. Synechocystis Page 2 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 Figure 1 Overall strategy for the production of bacterial DNA microarrays using the two-stage PCR approach for the production DNA microarray elements. Page 3 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 First round amplifications were performed in 96-well microtiter dish format using PTC-100 Thermocyclers from MJ Research Inc. Five µ of 20 µM premixed primer pairs were transferred to 0.2 ml Thermo-Fast 96-well plates (Marsh cat. No. AB-0600) and mixed with 95 µl of Reac- tion mix 1 (For 110 reactions; 55 µl of 500 µg/ml chromo- somal DNA, 1.1 ml Pfx Buffer , 1.1 ml Pfx Enhancer reagent , 220 µl 50 mM MgSO , 165 µl 10 mM premixed dNTPs [Amersham Pharmacia cat. No. 93-77212], 88 µl Pfx DNA Polymerase [Invitrogen cat. No. 11708-039] and 7.722 ml sterile, deionized water). Plates were sealed using Microseal™ 'A' film (MJ Research cat. No. MSA- 50001) during thermal cycling. Step-down thermo- cycling conditions consisted of 2 min 94°C initial dena- turation followed by ten rounds of 94°C for 30 s, 62°C(- 1°C/cycle) for 45 s, and 68°C for 3 min. The step down cycling was followed immediately by 24 cycles of 94°C for 30 s, 52°C for 45 s and 68°C for 3 min. Five µl of each sample was mixed with an equal volume of 20% glycerol and loaded on 0.7% agarose/TAE gels stained with ethid- ium bromide and separated by electrophoresis in Bio-Rad Sub-cell model 192 units. Gels were analyzed using an Alpha Imager™ 2000 workstation and software for image capture and Total Lab v. 1.10 (Phoretix) software for image analysis. Second round products were generated with universal primers and products of the first round of amplification. Specifically, 5 µl of the original 100 µl reac- tion was transferred to new plates and mixed with 95 µl reaction mix 2 (For 110 reactions; 1.1 ml 10X Buffer, 220 µl 50 mM MgSO , 165 µl premixed 10 mM dNTP's, 550 µl 20 µM universal primer pair, 1.1 ml Pfx Enhancer rea- gent, 88 µl Pfx enzyme, and 7.227 ml sterile, deionized water). Amplifications were carried out by denaturation for 2 min at 94°C followed by 35 cycles of the following conditions: 94°C for 15 s, 55°C for 45 s, and 68°C for 2 min. After completion of all 35 cycles, an additional Figure 2 Design and use of chimeric bipartite oligonucleotides extension period at 68°C for 20 min was added to polish for PCR amplification of each open reading frame of the ends. Five µl of each sample was loaded into 0.7% aga- the genome from Synechocystis PCC6803. Bipartite rose gels and analyzed as described above. Oligonucle- oligos (a and b) were designed to include: 1) a gene specific otides for unsuccessful products after the second round sequence at their 3' end of variable length to maintain a con- reaction were redesigned, resynthesized, and used for stant t of 58–62°C; and 2) universal sequence at their 5' to reamplification. PCR fragments were compared to a introduce a common sequence to all primary products. known size marker (Cat. No. 10068-013, Invitrogen). These primary products were then re-amplified by universal Products were scored successful if their length, as defined oligonucleotides (c and d) introducing more universal by gel analysis, was +/- 15% of their predicted length, con- sequence. The re-amplification of products using universal tained only a single band, and were of adequate density. oligonucleotides allows amplification of all products under optimized conditions, thus improving the success rate and product yields. The final products of the second amplification Purification, preparation, and printing DNAs are analyzed for size, yield, and lack of multiplicity on agarose Upon verification of successful amplification by gel anal- gels, then purified, resuspended in printing buffer, and printed ysis, each plate of second round PCR products was puri- on modified surface glass slides. This strategy minimizes the fied using Multiscreen-PCR plates from Millipore cost and effort to replicate the PCR-generated DNA gene according to the manufacturers specifications, except that fragment library and facilitates several downstream processes a pre-wash was incorporated to remove potential sur- beyond the primary objective of producing DNA factant type agents present in at least some batches of the microarrays. filtration units. The resulting products were resuspended Page 4 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 in 50 µl sterile deionized water at room temperature on a (diethylpyrocarbonate) treated water, heated to 65°C for rotary shaker for 30 minutes. The yield of each product 5 minutes and then transferred to 42°C. Subsequently, 25 val- was measured by calculating concentration from A µl of a premix containing 2 µl of 100 mM DTT (dithioth- ues using Corning Incorporated's Costar UV transparent reitol), 10 µl 5X Superscript II buffer, 1 µl 50X aadNTP 96 well plates with the Molecular Devices Spectramax 384 mix (25 mM dA, C, and GTP, 8.4 mM dTTP, and 16.6 mM Plus spectrophotometer. The products were then dried in aadUTP [Sigma Cat. No.A-5660]), 10 µl DEPC treated a Savant DNA110 Speed Vac. Using a Qiagen Bio-Robot water, and 2 µl Superscript II was added to each tube. Fol- 3000™, each product was resuspended in water to 1.0 lowing two hour incubation at 42°C, 4 µl 50 mM EDTA -1 µg.µl . To each sample, an equal volume of 2X Micros- and 2 µl 10 N NaOH was added to each tube and incu- potting solution (Cat. No. MSS-1, Telechem) was added. bated at 65°C for 20 minutes to degrade RNA. The reac- Ten microliter of each product was transferred to two sets tion was neutralized by the addition of 4 µl 5 M acetic of 384 Micro Array plates (Cat. No. X7020, Genetix Lim- acid. The cDNA was purified using Millipore Microcon 30 ited). Using a Omnigrid arrayer (Gene Machines) and 16 centrifugal filter devices by diluting the reaction mix with Microquill 2000 pins (Majer Precision), arrays were sterile deionized water to a volume of 500 µl according to printed in triplicate on Superamine slides (cat. No. SMM- the manufacturers specifications. Each of the two samples 25, Telechem), using 250 micron spacing, in 18 × 15 spot was washed five times using 500 µl deionized water and subgrids. To improve spot morphology, blotting was the final retentate adjusted to approximately 50 µl to facil- increased to 30 spots per dip, and pins were re-dipped itate handling. The resultant solutions were dried at room into the sample after every 50 slides. Cross-contamination temperature in the Speed Vac, and the pellets were resus- between successively spotted samples is a potentially pended in 20 µl 0.1 M Na CO (pH 9.0) and mixed with 2 3 severe problem and was found to be minimized by 10 µl of previously prepared Cy3 or Cy5 dye. Preparation increasing the stringency of pin washing procedures of the dyes was performed in advance and involved dis- beyond the manufacturer's recommendations by increas- solving one tube of powdered Cy3 or Cy5 dye from Amer- ing sonication in the washing solution for 5 seconds at the sham Pharmacia Biotech Cat. No. PA23001, or PA 25001 highest power setting. Relative humidity was maintained respectively, in 55 µl dimethyl sulfoxide and storing at - at 50% and temperature at 23°C throughout the arraying 20°C until use. The coupling reaction was carried out at procedure. room temperature in the dark for 1 hour. The reaction was quenched by the addition of 4.5 µl 4 M hydroxylamine, Array process verification followed by incubation for an additional 5 min. The Cy3- At the end of printing, one or two slides were used to ver- or Cy5 dye-coupled cDNA samples were combined and ify the quality of printing such as spot morphology and purified using a Qiagen PCR product purification kit intensity. Slides were baked at 80°C for one hour, washed according to the manufacturer's specifications. Samples with 0.1% SDS and rinsed in deionized water. The slide were adjusted to 14.75 µl using SpeedVac and remainder was then stained with 100 nanomole aqueous Syto-61 of the hybridization components containing 2.5 µl of 10 (Molecular Probes) solution, washed twice with 0.1% µg/µl salmon sperm DNA, 8.75 µl 20X SSC, 0.25 µl 10% SDS, rinsed once in deionized water and scanned for flu- SDS, and 8.75 µl formamide were added. The mixture was orescence at 550 nm in the ScanArray 3000. Each wash or then heated for 2 minutes at 99°C, briefly centrifuged at rinse was carried out at room temperature for 5 minutes. high speed, and maintained at 42°C until the hybridiza- The scanned images were analyzed for spot morphology tion with the DNA microarray was begun. and intensity and the results of the analysis were used to optimize the printing process during each run. Hybridization and wash conditions Printed slides were baked at 80°C for 1 hour, washed Labeled cDNA production twice for 2 minutes each at room temperature in 0.1% Fluorescently labeled cDNA was produced using a two- SDS and once in deionized water to remove unbound step procedure involving cDNA production from target material. The slides were boiled in deionized water for 3 RNA using a reverse transcriptase reaction incorporating minutes to denature the printed DNA, dried using low aminoallyl-modified deoxynucleotide (aadUTP), fol- speed centrifugation in a specialized microscope slide- lowed by the second step involving chemical coupling of accommodating rotor (Telechem). The DNA on the fluorescent dye (either Cy3 or Cy5) to the introduced microarray was then subjected to a UV-cross-linking at a amino moieties of the newly synthesized cDNA. The dose of 150 mJ/cm . Prehybridization was performed by cDNA was synthesized from 16 µg total RNA using 5 µg incubating the slide in a fresh mixture of 100 ml of 25% random 8-base oligonucleotides (Sigma GenoSys) and formamide, 5X SSC, 0.1% SDS, and 1% BSA for 45 min. Superscript II™ reverse transcriptase (Invitrogen). RNA at 42°C in a Coplin jar. The slide was then rinsed under samples (control and experimental) were each mixed with distilled water and dried using low speed centrifugation. random octamers in a total volume of 25 µl DEPC Page 5 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 Table 1: Construction of DNA microarray Microarrays are constructed by synthesizing oligonucleotides on glass slides, by spotting PCR-amplified cDNA clones from EST data (eukaryotes), or by spotting PCR products from genomic DNA template (prokaryotes) using a set of gene-specific primers. The method described in this paper has the advantage of being both more cost effective and less labor intensive for the construction of microarrays from prokaryotic organisms. The procedure is most useful for those laboratories working with organisms for which commercial microarrays are unavailable. The salient features of our approach include: • Bipartite Primer: The use of a bipartite primer for the amplification of genes has several advantages (see text). Importantly, it permits the use of a common primer for the amplification of all genes in the second and subsequent amplification steps. This simple modification in primers allevi- ates the problem of primer-limitation and therefore, is particularly valuable in constructing many arrays for a multi-investigative, collaborative effort. • Two stage amplification: The use of the two-stage amplification process increases the signal-to-noise ratio by diminishing the hybridization sig- nal resulting from contaminating genomic DNA used for amplification of PCR fragments. Additionally, it also yields relatively uniform quantities of PCR products. • Purification and Quantitation: For efficient binding of the amplified products and to reduce the background signal, it is essential to purify the products. We found ethanol precipitation unsatisfactory, whereas the use of Multiscreen filter plates gave excellent results. One drawback with this technique is the relatively lower percentage recovery of smaller fragments. Therefore, following purification, fragments were quantified and normalized to an equal concentration. • Printing: We checked the suitability of various slide surfaces (poly-Lysine and amine) and the spotting buffers from a number of commercial vendors to ensure reproducible and uniform spot morphology and maximal retention of DNA. We found that Superamine slides and MSP spot- ting buffer from Telechem gave satisfactory results. The slide was preheated at 42°C in a Telechem hybridiza- tions that are well-described in the literature. Overall, the tion chamber by placing in a static 42°C incubator. The rate-limiting production steps following the reaction opti- pre-warmed sample was pipetted and spread uniformly mization were those involving the analysis of the prod- onto a 24 × 60 mm glass cover-slip (Fisher Scientific, Cat. ucts. Figure 1 outlines the main features of this successful No., 12-548-5P) and the pre-warmed slide was inverted construction project and Table 1 highlights some of the and placed with the arrayed surface contacting the sample key parameters that we tested during the array on the cover-slip surface. Deionized water (10 µl) was construction. added to each of the reservoirs of a Telechem Hybridiza- tion Chamber, and the slide was transferred to and Critical design features of the chimeric first round primers enclosed within the chamber. The slide was incubated in were the inclusion of the directional cloning and expres- a static incubator at 42°C for 12–16 h, and washed by sion features, plus the inclusion of a good common tem- placing in a 250 ml solution of 2X SSC and 0.1% SDS at plate sequence for the subsequent second round 42°C for 5 minutes with gentle agitation provided by amplification. The adaptamers included the introduction rotation of a magnetic stir bar. The slide was transferred of G+C rich sequences at the 5' termini to stabilize the end quickly to a solution of 0.1X SSC, and 0.1% SDS, incu- structure of the first round PCR products. Another impor- bated for 10 minutes at room temperature with gentle agi- tant design aspect that proved critical, according to our tation, and washed 5 additional times in 0.1X SSC for 1 preliminary tests, was incorporation of an additional five min. at room temperature. The slide was then rinsed base G+C rich overhang in the universal primers used for briefly with deionized water and dried by low speed cen- the second round PCR (Figure 2). The absence of the addi- trifugation. The slides were stored in the dark until tional sequences resulted in very poor PCR performance scanned. compared to the excellent performance observed with their inclusion. Results and Discussion The goal of this project was to develop a DNA microarray Importantly, the two-stage amplification had the unantic- as part of an overall strategy for functional genomics in ipated advantage of normalizing the yield of PCR prod- the unicellular model cyanobacterium, Synechocystis ucts across the entire set of genes, such that lower than st PCC6803. The two-stage amplification strategy (6) was average yielding products from the 1 round PCR (using adopted as a means to facilitate replication of the arrays as the gene specific sequence and chromosomal DNA as nd needs dictated (Figure 1). In general terms, the optimiza- template) often had a robust yield during the 2 round tion of this technical approach required particular atten- of amplification using the common adaptamer sequences. tion to the design of the adaptamers, the choice and use of Although the primers were deliberately designed in such a the thermostable DNA polymerase, and the thermocy- way that oligo T 's fell in narrow range, it was determined cling conditions. Other technical factors, such as the array after a number of trials that it was necessary to use printing conditions, were found to conform to observa- 'touchdown' PCR conditions for the first round amplifica- Page 6 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 Figure 3 Thermal cycling conditions and resultant products from first and second rounds of amplification. Panel A illus- trates the thermal cycling conditions for the primary and secondary amplifications. During the first round amplification (Panel A, left side) with gene-specific chimeric primers, step-down thermal cycling PCR conditions involve decreasing the annealing temperature by 1°C per cycle beginning at 62°C for 10 cycles and the subsequent 24 cycles have a set annealing temperature of 52°C. The graph on the right depicts the conditions for constant annealing temperature of 55°C for the second round ampli- fication using a common set of adaptamers primers complementary to the entire set of first round products (Panel A, right side). Panel B illustrates one half (48 reactions) of the results from the primary (above) and secondary (below) amplification of one 96-well plate. Lanes 2 and 6 (arrows 1 and 2) illustrate the product 'leveling effect' on the yields of product due to the sec- ondary amplification with adaptamer primers. Lanes 9, 14, and 35 (arrows 3, 4, and 7) illustrate missing or double products which are scored as bad and have new primers designed for subsequent use in another PCR run. Lanes 17 and 34 (under arrows 5 and 6, respectively) contain molecular weight markers. Page 7 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 Table 2: Synthesis and hybridization of fluorescently labeled cDNA derived from sample RNA Hybridization of the microarray with sample representing the mRNA population is accomplished by synthesizing complementary DNA (cDNA) using reverse transcriptase. Microarray experiments in prokaryotic organisms cannot utilize the poly(A) tails found in eukaryotic mRNA; thus, alter- natives to the highly successful methods employed for generating fluorescent cDNA using oligo-dT to prime the reverse transcription labeling reac- tion have to be employed. • cDNA synthesis Tests comparing random hexamers and random octamers gave essentially indistinguishable results. On the other hand, critical parameters for maximizing signal-to-noise were: 1.) high level of incorporation of fluorescent dye by using a high ratio of amino allyl dUTP in the reverse transcription reaction and saturating amount of reactive dye during the coupling of activated dye; and 2.) the use of two purification steps relying on independent physical separation principals applied to the test sample: first after cDNA synthesis and second, after the coupling of fluorescent dye. Direct chemical labeling of RNA, while attractive, did not allow as high signal-to-noise during our tests. Similarly, pilot exper- iments using pooled 3' gene-specific oligos as cDNA primers did not give better results. • Hybridization: While increases in signal-to-noise were associated with optimization of fluorescent sample production (above), a fairly wide range of standard microarray pre-hybridization and hybridization protocols were found to give comparable results. However, critical aspects of the described method were keeping the microarray slide warm during the addition of sample to the surface following pre-hybridization and ensuring that the slide remains hydrated prior to the final drying immediately before scanning. tion. Nevertheless, the percentage of successful products the PCR amplification are available at: http://microbiol- after the first round synthesis appeared quite low, based ogy.okstate.edu/faculty/burnap/index.html. on yield as visualized in the ethidium stained electro- phoretic gels. Some lanes showed little to no product or Organization and quality control were emphasized multiple products after the first round amplification with throughout production of the microarrays to ensure uni- the chimeric primers. However, these products generally formity and accuracy. A semi-automated gel analysis pro- resulted in successful amplification after their use as tem- cedure scored products on the basis of size and absence of plate in the second round amplification when the univer- multiplicity. All PCR products were documented by dig- sal sequences introduced as overhangs of the chimeric ital photography and the images quantitatively analyzed primers during the first round amplification were used. using Total Lab™ Phoretics gel analysis software (Nonlin- Thus, the universal primers in the second round amplifi- ear Dynamics. Durham, North Carolina). This application cation provided the additional benefit of generating con- was found to be most accurate, user-friendly, and effi- sistent yields across most of the genes (see Figure 3). ciently handled dual tier gels. Furthermore, as with other Apparently, the amount of correct template produced by packages, it allowed simple export of the numerical results st the 1 round PCR, albeit often invisible during gel analy- to Excel spreadsheets. A database was developed to inte- nd sis, was usually sufficient to give high yields during the 2 grate information, including gene sequences, oligonucle- round of amplification using the common adaptamer otide design, PCR product length, etc. In addition, Excel sequences. Thus 86% of the products were successfully macros were developed to transform the aforementioned amplified during our initial round of two-stage PCR. information to produce lists for each plate, which were Primers for the 14% unsuccessful products were rede- then used to track and identify the location of the arrayed signed and synthesized, so that essentially all of the anno- elements under any condition, including the use of differ- tated genes were successfully amplified. Therefore, we ent array configurations. The scoring system, described in conclude that this two-stage amplification strategy for the the Materials and Methods section, was set up to identify purpose of PCR amplifying multiple products for genome successful PCR amplified products. scale applications is highly successful, robust and efficient. DNA purification was simply and efficiently performed using filtration-type purification plates. Early issues with The use of adaptamer sequences with identical T 's for the variation in printed spot morphology were resolved by nd 2 round of amplification optimized the yields, despite pre-rinsing the filters: apparently this removes material different product lengths and different template concen- acting as a surfactant and effects the spot morphology trations (which were dependent on the yield from the first and/or drying characteristics. Alternatives such as ethanol round reactions). The enzyme and thermocycling condi- or isopropanol precipitations were explored, but for the tions were also optimized. The choice of Platinum Pfx chosen PCR conditions, produced less reliable results and (Invitrogen) as the polymerase was determined to be cru- appeared to cause the co-precipitation of non-DNA mate- cial as was its use in conjunction with the proprietary reac- rial that interfered with downstream operations. The tion enhancer solution supplied with the buffer. Tests nature of this material remains undefined, although it with alternatives, including mixes of Taq and Pfu , yielded may be due to constituents of the PCR reaction proprie- less satisfactory results. All the images for the analysis of tary enhancer that is supplied with the Pfx polymerase. Page 8 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 Figure 4 Scatter-plot of differential hybridization of fluorescently labeled cDNA from a 3-hour salt shock treated cell culture versus an untreated sample. Log phase cultures were subjected to an upshift in NaCl concentration from ~25 mM to 650 mM. Syne- chocystis is halotolerant and grows at concentrations up to approximately 1.2 M NaCl. Fluorescently labeled (Cy-3) cDNA derived from total RNA extracted from cells exposed 3 hours to the higher salt concentration was co-hybridized with Cy-5- labelled cDNA derived from the culture immediately prior to the upshift (control cells). The full genome Synechocystis sp. PCC 6803 high-density microarrays contain DNA features printed in triplicate on glass slides. The gene names according to original annotation are behind the arrows. As indicated with arrows, signals from the replicated elements exhibited similarity in the estimated expression ratio. Each data point corresponds to a different gene address on the microarray and each gene is repli- cated at three separate and spatially distant addresses. The X and Y axes correspond to the normalized fluorescence intensity of fluorescence-labeled cDNA of the control sample (Y-axis, Cy-5, 532 nm fluorescence) and the 650 mM NaCl three hour time point (X-axis, Cy-3, 635 nm fluorescence). Page 9 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 The concentrations of the purified DNA fragments were not provided as high a hybridization signal-to-noise ratio determined (Molecular Devices Spectramax Plus) and a compared to the chemical coupling to cDNA approach. Qiagen Bio-Robot 3000 was used to dilute the DNA sam- ples to 500 ng/µl in 2X spotting solution. Tests showed Substantial effort was devoted to ensure spot uniformity, that the combination of Superamine slides and MSP high signal-to-noise ratio, and good statistical reproduci- printing buffer solution from Telechem provided the best bility. We believe that we have met these objectives, and combination of spot morphology, retention of DNA after these arrays relate favorably to the best of current prokary- processing of slides and signal-to-noise ratio after hybrid- otic and eukaryotic arrays. The strengths of these Syne- ization. Alternately, Clontech type II slides in chocystis arrays include: 1.) consistent and high DNA combination with the Telechem spotting buffer also gave concentrations in each of over 11,000 spots per array; 2.) good results in terms of printing and hybridization 3 replicates per probe per slide; and 3.) high signal-to- results. Tests with other buffer systems, e.g. 6X SSC and noise ratio. This last parameter is due both to the quality 50% DMSO, gave less satisfactory results under the speci- of the slides, the hybridization procedures that we utilized fied printing conditions. The synthesis and hybridization and the quality of the products printed. We have used of the fluorescently labeled cDNA derived from sample these arrays to develop robust statistical tools for the anal- RNA is outlined in Table 2. ysis of the microarray data. The analysis involves an ANOVA model that makes excellent use of the technical In order to generate uniform spot morphology and to replicates designed into each slide and a loop design maximize DNA concentration and consistency for each (Singh, McIntyre and Sherman, unpublished results). The spot across the array, all products were printed at the same availability of large quantities of DNA, the ability to print concentration. This was accomplished using the Qiagen many slides and the statistical tools will enable us to ana- Bio-robot 3000™ to resuspend each of the purified PCR lyze transcriptional changes under many environmental products to the same concentration, based on the results conditions and for many mutants. The two-stage amplifi- of the Spectramax UV analysis. These products were then cation procedure also provides full gene copies for tech- mixed with an equal volume of 2X MSP printing buffer niques like PCR-fusion mutagensis. This will permit many from Telechem and two aliquots were transferred via the comparisons among different conditions and should Bio-robot to separate printing plates. The use of this greatly facilitate functional genomics. In conclusion, the robotic transfer reduced the chance for human error, two-stage PCR amplification strategy detailed here is increased the rate of transfer (thus reducing the exposure demonstrated to be highly useful and undoubtedly can be of the products to desiccation), and increased the repro- applied to other bacterial species. ducibility of the transfer. Syto-61 staining was regularly used to maintain a high level of quality control during the Competing interests printing process. None declared. Several internal and external comparisons were analyzed Authors Contributions to demonstrate the effective use of the arrays. Internally, BLP, H-LW and RLB developed and optimized the PCR the use of triplicate spotting of all array elements and dis- primer design strategy, BLP, H-LW, AS, HW, and LI per- tribution of the triplicate elements in different regions of formed the PCR amplifications; HLA, JK, and GR per- the array allows the comparison of hybridization formed gel analysis and quality control, DP developed uniformity across the array, as depicted in Figure 4. Ini- informatics, BLP, AS, MD, and DWG conducted the tially, we determined the effectiveness of printed slides by microarray printing, BP, LAS, and RLB organized the hybridization with total RNA from control and salt- project and prepared the manuscript. stressed cells. The results were compared with that of the published report from Kanesaki et al. [11]. Similar pat- Acknowledgements The authors would like to acknowledge Dr. Prade, Dr. Cushman, Dr. Bon- terns of regulation of specific transcripts were observed hert, and Dr. Stintzi for their advice and input. This work was supported by and several examples of the differentially regulated genes the National Science Foundation (to RLB and LAS) and the Department of are indicated in Figure 4. Qualitatively similar results Energy (to LAS) and by the Grula Fellowship to BP as part of his thesis work using a direct RNA-dye coupling procedure can be through the OSU Foundation. achieved using the Micromax™ ASAP RNA Labeling Kit (Perkin Elmer Life Sciences, Inc. cat. No. MPS544) follow- References ing the manufacturer's suggested protocol. While it is sig- 1. Schena M, Shalon D, Davis RW and Brown PO: Quantitative mon- nificantly quicker and requires less manipulation, the itoring of gene expression patterns with a complementary DNA microarray Science 1995, 270:467-470. direct chemical labeling kit is more expensive, and pro- 2. Spellman PT, Sherlock G, Zhang MQ, Iyer VR, Anders K, Eisen MB, duces a final RNA product which is less stable than the Brown PO, Botstein D and Futcher B: Comprehensive identification of cell cycle-regulated genes of the yeast cDNA produced. Significantly direct labeling has thus far Page 10 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 Saccharomyces cerevisiae by microarray hybridization Mol Biol Cell 1998, 9:3273-3297. 3. Iyer VR, Eisen MB, Ross DT, Schuler G, Moore T, Lee JCF, Trent JM, Staudt LM, Hudson J., Jr., Boguski MS, Lashkari D, Shalon D, Botstein D and Brown PO: The transcriptional program in the response of human fibroblasts to serum Science 1999, 283:83-87. 4. White KP, Rifkin SA, Hurban P and Hogness DS: Microarray anal- ysis of Drosophila development during metamorphosis Sci- ence 1999, 286:2179-2184. 5. Hihara Y, Kamei A, Kanehisa M, Kaplan A and Ikeuchi M: DNA microarray analysis of cyanobacterial gene expression dur- ing acclimation to high light Plant Cell 2001, 13:793-806. 6. Richmond CS, Glasner JD, Mau R, Jin H and Blattner FR: Genome- wide expression profiling in Escherichia coli K-12 Nucleic Acids Res 1999, 27:3821-3835. 7. Bryant DA: The Molecular Biology of Cyanobacteria The Neth- erlands, Kluwer Academic Publishers; 1994. 8. Kaneko T, Sato S, Kotani H, Tanaka A, Asamizu E, Nakmura Y, Miya- jima N, Hirosawa M, Sugiura M, Sasamoto S, Kimura T, Hosouchi T, Matsuno A, Muraki A, Nakzaki N, Naruo K, Okumura S, Shimpo S, Takeuchi C, Wada T, Watanabe A, Yamada M, Yasuda M and Tabata S: Sequence analysis of the genome of the unicellular cyano- bacterium Synechocystis sp. strain 6803. II. Sequence deter- mination of the entire genome and assignment of potential protein-coding regions. DNA Research 1996, 3:109-136. 9. Williams JGK: Construction of specific mutations in Photosys- tem II photosynthetic reaction center by genetic engineer- ing methods in Synechocystis 6803. Methods in Enzymology 1988, 167:766-778. 10. He Q, Dolganov N, Bjorkman O and Grossman AR: The high light- inducible polypeptides in Synechocystis PCC6803. Expres- sion and function in high light J Biol Chem 2001, 276:306-314. 11. Kanesaki Y, Suzuki I, Allakhverdiev SI, Mikami K and Murata N: Salt stress and hyperosmotic stress regulate the expression of different sets of genes in Synechocystis sp. PCC 6803 Biochem Biophys Res Commun 2002, 290:339-348. Publish with Bio Med Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical researc h in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright BioMedcentral Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp Page 11 of 11 (page number not for citation purposes) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png BMC Genomics Springer Journals

The construction and use of bacterial DNA microarrays based on an optimized two-stage PCR strategy

Download PDF
11 pages

Loading next page...
 
/lp/springer-journals/the-construction-and-use-of-bacterial-dna-microarrays-based-on-an-mQe89t85Ar

References (13)

Publisher
Springer Journals
Copyright
Copyright © 2003 by Postier et al; licensee BioMed Central Ltd.
Subject
Life Sciences; Life Sciences, general; Microarrays; Proteomics; Animal Genetics and Genomics; Microbial Genetics and Genomics; Plant Genetics & Genomics
eISSN
1471-2164
DOI
10.1186/1471-2164-4-23
pmid
12803655
Publisher site
See Article on Publisher Site

Abstract

Background: DNA microarrays are a powerful tool with important applications such as global gene expression profiling. Construction of bacterial DNA microarrays from genomic sequence data using a two-stage PCR amplification approach for the production of arrayed DNA is attractive because it allows, in principal, the continued re-amplification of DNA fragments and facilitates further utilization of the DNA fragments for additional uses (e.g. over-expression of protein). We describe the successful construction and use of DNA microarrays by the two-stage amplification approach and discuss the technical challenges that were met and resolved during the project. Results: Chimeric primers that contained both gene-specific and shared, universal sequence allowed the two-stage amplification of the 3,168 genes identified on the genome of Synechocystis sp. PCC6803, an important prokaryotic model organism for the study of oxygenic photosynthesis. The gene-specific component of the primer was of variable length to maintain uniform annealing st temperatures during the 1 round of PCR synthesis, and situated to preserve full-length ORFs. Genes were truncated at 2 kb for efficient amplification, so that about 92% of the PCR fragments were full-length genes. The two-stage amplification had the additional advantage of normalizing the yield of PCR products and this improved the uniformity of DNA features robotically deposited onto the microarray surface. We also describe the techniques utilized to optimize hybridization conditions and signal-to-noise ratio of the transcription profile. The inter-lab transportability was demonstrated by the virtual error-free amplification of the entire genome complement of 3,168 genes using the universal primers in partner labs. The printed slides have been successfully used to identify differentially expressed genes in response to a number of environmental conditions, including salt stress. Conclusions: The technique detailed here minimizes the cost and effort to replicate a PCR- generated DNA gene fragment library and facilitates several downstream processes (e.g. directional cloning of fragments and gene expression as affinity-tagged fusion proteins) beyond the primary objective of producing DNA microarrays for global gene expression profiling. Page 1 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 PCC6803 was predicted to contain 3,168 open reading Background DNA microarrays have proven to be a particularly effective frames (prior to May 2002) http://www.kazusa.or.jp/ cyano/index.html[8]. Of these, over 50% (>1700 ORFs) tool for genome-wide transcript profiling, and they can generate information on the accumulation and disappear- have been characterized as hypothetical or having no sim- ance of thousands of specific transcripts in a single exper- ilarity to known sequences. Thus, differential gene expres- iment [1–5]. Because such an array generates vast sion via microarray analysis will represent an important quantities of data, it is desirable to improve the statistical functional genomics approach and we describe the two- reliability of the data through experimental repetitions. stage PCR strategy that led to the successful production of Additionally, it is useful to perform time course experi- full-genome microarray for Synechocystis . ments involving multiple time points to enable the appli- cation of cluster analysis to classify groups of genes Methods exhibiting similar temporal patterns of regulation in Growth conditions and RNA isolation response to shifts in environmental conditions. Therefore, Synechocystis PCC6803 cultures were grown autotrophi- a robust experimental design may involve the consump- cally in BG-11 medium essentially as described by Wil- tion of many individual DNA microarrays. Given these liams [9]. Specifically, 500 mL cultures were grown in flat considerations, it is important to have methods for (4 cm across the light path) Bellco tissue culture flasks at improving the efficiency and economy of microarray 30°C under white light fluorescent illumination (Cool production. White , General Electric) with an incident intensity of 80 -2 -1 µmoles photons m s . Aeration was provided by bub- Since the bacterial genes are typically not cloned at a bling with air enriched with 3% CO at a rate of 300 mL -1 genome-wide scale as non-redundant, ordered sets, as min . RNA was isolated using hot phenol conjunction with eukaryotic cDNA libraries, it is not possible to use with vigorous agitation with glass beads as described by 'universal primers' (e.g. T3, T7 sequences) and therefore Bhaya et al. [10]. PCR amplification requires design and synthesis of PCR primers for all target genes. In this context, the construc- Chimeric oligonucleotide design and PCR amplification tion of bacterial DNA microarrays from genomic Figure 1 provides a flow chart of the preparation of the full-genome Synechocystis microarray. Oligonucleotides, sequence data using a 'two-stage' approach, is attractive because it allows the continued re-amplification of DNA synthesized by Sigma Genosys http://www.sigmageno- fragments and facilitates further utilization of the DNA sys.com/, were designed to amplify the complete coding fragments for additional uses (e.g. over-expression of pro- region of each open reading frame from the sequenced tein) [6]. This approach involves the use of chimeric prim- genome of Synechocystis PCC6803. The only exceptions ers containing both gene-specific and engineered were instances for genes with a predicted length extending sequences (adaptamers). These primers are used to syn- beyond 2 kb. Of the 3,168 open reading frames, fewer thesize PCR products directly from genomic DNA, which than 10% (261) of the genes were truncated at the 3'-end serves as template during the first stage of PCR amplifica- to the 2 kb limit. Each of the oligos is a chimeric sequence tion. This amplification can be performed for thousands designed for first round synthesis and, as such, contained of genes during a given production run using well-estab- two components: a 5' universal sequence and a 3' gene- lished procedures that include microtiter plates and mul- specific sequence as shown in Figure 2. The gene-specific tichannel pipetting techniques to afford efficient parallel region of the oligonucleotides begins at the first or last processing of the samples. The products of this first stage base of the open reading frame and continues inside until of amplification are gene sequences flanked by the arbi- the T of the oligo was in the range of 58–62°C (25–45 trary adaptamer sequences engineered into the first stage nt, including the universal sequence). An algorithm PCR primers. The adaptamer sequences become the basis (Sigma Genosys) based upon Primer3 (Whitehead Insti- for re-amplification of the gene sequences in a second tute) was used for this purpose and resulted in a high fre- stage amplification. During the second stage of amplifica- quency of successful amplification as discussed below. tion, the first stage PCR products serve as template, Importantly, the universal sequences were designed to whereas a single pair of primers recognizing the flanking include SapI restriction sites in order to facilitate down- adaptamer sequence can be used to re-amplify all prod- stream applications, including cloning into an expression ucts of the first stage PCR. vector. Aliquots of the oligonucleotides were transferred and mixed as primer pairs to separate locations of 96-well In this report, we describe the procedures for the produc- plates prior to delivery from the commercial source, tion and use of DNA microarrays for the cyanobacterium which facilitated amplification in the 96-well format. The Synechocystis PCC6803. This organism is a widely used remaining products from the synthesis reaction were experimental model for the study of gene expression and stored in individual tubes at -20°C. plant-like oxygenic photosynthesis [7]. Synechocystis Page 2 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 Figure 1 Overall strategy for the production of bacterial DNA microarrays using the two-stage PCR approach for the production DNA microarray elements. Page 3 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 First round amplifications were performed in 96-well microtiter dish format using PTC-100 Thermocyclers from MJ Research Inc. Five µ of 20 µM premixed primer pairs were transferred to 0.2 ml Thermo-Fast 96-well plates (Marsh cat. No. AB-0600) and mixed with 95 µl of Reac- tion mix 1 (For 110 reactions; 55 µl of 500 µg/ml chromo- somal DNA, 1.1 ml Pfx Buffer , 1.1 ml Pfx Enhancer reagent , 220 µl 50 mM MgSO , 165 µl 10 mM premixed dNTPs [Amersham Pharmacia cat. No. 93-77212], 88 µl Pfx DNA Polymerase [Invitrogen cat. No. 11708-039] and 7.722 ml sterile, deionized water). Plates were sealed using Microseal™ 'A' film (MJ Research cat. No. MSA- 50001) during thermal cycling. Step-down thermo- cycling conditions consisted of 2 min 94°C initial dena- turation followed by ten rounds of 94°C for 30 s, 62°C(- 1°C/cycle) for 45 s, and 68°C for 3 min. The step down cycling was followed immediately by 24 cycles of 94°C for 30 s, 52°C for 45 s and 68°C for 3 min. Five µl of each sample was mixed with an equal volume of 20% glycerol and loaded on 0.7% agarose/TAE gels stained with ethid- ium bromide and separated by electrophoresis in Bio-Rad Sub-cell model 192 units. Gels were analyzed using an Alpha Imager™ 2000 workstation and software for image capture and Total Lab v. 1.10 (Phoretix) software for image analysis. Second round products were generated with universal primers and products of the first round of amplification. Specifically, 5 µl of the original 100 µl reac- tion was transferred to new plates and mixed with 95 µl reaction mix 2 (For 110 reactions; 1.1 ml 10X Buffer, 220 µl 50 mM MgSO , 165 µl premixed 10 mM dNTP's, 550 µl 20 µM universal primer pair, 1.1 ml Pfx Enhancer rea- gent, 88 µl Pfx enzyme, and 7.227 ml sterile, deionized water). Amplifications were carried out by denaturation for 2 min at 94°C followed by 35 cycles of the following conditions: 94°C for 15 s, 55°C for 45 s, and 68°C for 2 min. After completion of all 35 cycles, an additional Figure 2 Design and use of chimeric bipartite oligonucleotides extension period at 68°C for 20 min was added to polish for PCR amplification of each open reading frame of the ends. Five µl of each sample was loaded into 0.7% aga- the genome from Synechocystis PCC6803. Bipartite rose gels and analyzed as described above. Oligonucle- oligos (a and b) were designed to include: 1) a gene specific otides for unsuccessful products after the second round sequence at their 3' end of variable length to maintain a con- reaction were redesigned, resynthesized, and used for stant t of 58–62°C; and 2) universal sequence at their 5' to reamplification. PCR fragments were compared to a introduce a common sequence to all primary products. known size marker (Cat. No. 10068-013, Invitrogen). These primary products were then re-amplified by universal Products were scored successful if their length, as defined oligonucleotides (c and d) introducing more universal by gel analysis, was +/- 15% of their predicted length, con- sequence. The re-amplification of products using universal tained only a single band, and were of adequate density. oligonucleotides allows amplification of all products under optimized conditions, thus improving the success rate and product yields. The final products of the second amplification Purification, preparation, and printing DNAs are analyzed for size, yield, and lack of multiplicity on agarose Upon verification of successful amplification by gel anal- gels, then purified, resuspended in printing buffer, and printed ysis, each plate of second round PCR products was puri- on modified surface glass slides. This strategy minimizes the fied using Multiscreen-PCR plates from Millipore cost and effort to replicate the PCR-generated DNA gene according to the manufacturers specifications, except that fragment library and facilitates several downstream processes a pre-wash was incorporated to remove potential sur- beyond the primary objective of producing DNA factant type agents present in at least some batches of the microarrays. filtration units. The resulting products were resuspended Page 4 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 in 50 µl sterile deionized water at room temperature on a (diethylpyrocarbonate) treated water, heated to 65°C for rotary shaker for 30 minutes. The yield of each product 5 minutes and then transferred to 42°C. Subsequently, 25 val- was measured by calculating concentration from A µl of a premix containing 2 µl of 100 mM DTT (dithioth- ues using Corning Incorporated's Costar UV transparent reitol), 10 µl 5X Superscript II buffer, 1 µl 50X aadNTP 96 well plates with the Molecular Devices Spectramax 384 mix (25 mM dA, C, and GTP, 8.4 mM dTTP, and 16.6 mM Plus spectrophotometer. The products were then dried in aadUTP [Sigma Cat. No.A-5660]), 10 µl DEPC treated a Savant DNA110 Speed Vac. Using a Qiagen Bio-Robot water, and 2 µl Superscript II was added to each tube. Fol- 3000™, each product was resuspended in water to 1.0 lowing two hour incubation at 42°C, 4 µl 50 mM EDTA -1 µg.µl . To each sample, an equal volume of 2X Micros- and 2 µl 10 N NaOH was added to each tube and incu- potting solution (Cat. No. MSS-1, Telechem) was added. bated at 65°C for 20 minutes to degrade RNA. The reac- Ten microliter of each product was transferred to two sets tion was neutralized by the addition of 4 µl 5 M acetic of 384 Micro Array plates (Cat. No. X7020, Genetix Lim- acid. The cDNA was purified using Millipore Microcon 30 ited). Using a Omnigrid arrayer (Gene Machines) and 16 centrifugal filter devices by diluting the reaction mix with Microquill 2000 pins (Majer Precision), arrays were sterile deionized water to a volume of 500 µl according to printed in triplicate on Superamine slides (cat. No. SMM- the manufacturers specifications. Each of the two samples 25, Telechem), using 250 micron spacing, in 18 × 15 spot was washed five times using 500 µl deionized water and subgrids. To improve spot morphology, blotting was the final retentate adjusted to approximately 50 µl to facil- increased to 30 spots per dip, and pins were re-dipped itate handling. The resultant solutions were dried at room into the sample after every 50 slides. Cross-contamination temperature in the Speed Vac, and the pellets were resus- between successively spotted samples is a potentially pended in 20 µl 0.1 M Na CO (pH 9.0) and mixed with 2 3 severe problem and was found to be minimized by 10 µl of previously prepared Cy3 or Cy5 dye. Preparation increasing the stringency of pin washing procedures of the dyes was performed in advance and involved dis- beyond the manufacturer's recommendations by increas- solving one tube of powdered Cy3 or Cy5 dye from Amer- ing sonication in the washing solution for 5 seconds at the sham Pharmacia Biotech Cat. No. PA23001, or PA 25001 highest power setting. Relative humidity was maintained respectively, in 55 µl dimethyl sulfoxide and storing at - at 50% and temperature at 23°C throughout the arraying 20°C until use. The coupling reaction was carried out at procedure. room temperature in the dark for 1 hour. The reaction was quenched by the addition of 4.5 µl 4 M hydroxylamine, Array process verification followed by incubation for an additional 5 min. The Cy3- At the end of printing, one or two slides were used to ver- or Cy5 dye-coupled cDNA samples were combined and ify the quality of printing such as spot morphology and purified using a Qiagen PCR product purification kit intensity. Slides were baked at 80°C for one hour, washed according to the manufacturer's specifications. Samples with 0.1% SDS and rinsed in deionized water. The slide were adjusted to 14.75 µl using SpeedVac and remainder was then stained with 100 nanomole aqueous Syto-61 of the hybridization components containing 2.5 µl of 10 (Molecular Probes) solution, washed twice with 0.1% µg/µl salmon sperm DNA, 8.75 µl 20X SSC, 0.25 µl 10% SDS, rinsed once in deionized water and scanned for flu- SDS, and 8.75 µl formamide were added. The mixture was orescence at 550 nm in the ScanArray 3000. Each wash or then heated for 2 minutes at 99°C, briefly centrifuged at rinse was carried out at room temperature for 5 minutes. high speed, and maintained at 42°C until the hybridiza- The scanned images were analyzed for spot morphology tion with the DNA microarray was begun. and intensity and the results of the analysis were used to optimize the printing process during each run. Hybridization and wash conditions Printed slides were baked at 80°C for 1 hour, washed Labeled cDNA production twice for 2 minutes each at room temperature in 0.1% Fluorescently labeled cDNA was produced using a two- SDS and once in deionized water to remove unbound step procedure involving cDNA production from target material. The slides were boiled in deionized water for 3 RNA using a reverse transcriptase reaction incorporating minutes to denature the printed DNA, dried using low aminoallyl-modified deoxynucleotide (aadUTP), fol- speed centrifugation in a specialized microscope slide- lowed by the second step involving chemical coupling of accommodating rotor (Telechem). The DNA on the fluorescent dye (either Cy3 or Cy5) to the introduced microarray was then subjected to a UV-cross-linking at a amino moieties of the newly synthesized cDNA. The dose of 150 mJ/cm . Prehybridization was performed by cDNA was synthesized from 16 µg total RNA using 5 µg incubating the slide in a fresh mixture of 100 ml of 25% random 8-base oligonucleotides (Sigma GenoSys) and formamide, 5X SSC, 0.1% SDS, and 1% BSA for 45 min. Superscript II™ reverse transcriptase (Invitrogen). RNA at 42°C in a Coplin jar. The slide was then rinsed under samples (control and experimental) were each mixed with distilled water and dried using low speed centrifugation. random octamers in a total volume of 25 µl DEPC Page 5 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 Table 1: Construction of DNA microarray Microarrays are constructed by synthesizing oligonucleotides on glass slides, by spotting PCR-amplified cDNA clones from EST data (eukaryotes), or by spotting PCR products from genomic DNA template (prokaryotes) using a set of gene-specific primers. The method described in this paper has the advantage of being both more cost effective and less labor intensive for the construction of microarrays from prokaryotic organisms. The procedure is most useful for those laboratories working with organisms for which commercial microarrays are unavailable. The salient features of our approach include: • Bipartite Primer: The use of a bipartite primer for the amplification of genes has several advantages (see text). Importantly, it permits the use of a common primer for the amplification of all genes in the second and subsequent amplification steps. This simple modification in primers allevi- ates the problem of primer-limitation and therefore, is particularly valuable in constructing many arrays for a multi-investigative, collaborative effort. • Two stage amplification: The use of the two-stage amplification process increases the signal-to-noise ratio by diminishing the hybridization sig- nal resulting from contaminating genomic DNA used for amplification of PCR fragments. Additionally, it also yields relatively uniform quantities of PCR products. • Purification and Quantitation: For efficient binding of the amplified products and to reduce the background signal, it is essential to purify the products. We found ethanol precipitation unsatisfactory, whereas the use of Multiscreen filter plates gave excellent results. One drawback with this technique is the relatively lower percentage recovery of smaller fragments. Therefore, following purification, fragments were quantified and normalized to an equal concentration. • Printing: We checked the suitability of various slide surfaces (poly-Lysine and amine) and the spotting buffers from a number of commercial vendors to ensure reproducible and uniform spot morphology and maximal retention of DNA. We found that Superamine slides and MSP spot- ting buffer from Telechem gave satisfactory results. The slide was preheated at 42°C in a Telechem hybridiza- tions that are well-described in the literature. Overall, the tion chamber by placing in a static 42°C incubator. The rate-limiting production steps following the reaction opti- pre-warmed sample was pipetted and spread uniformly mization were those involving the analysis of the prod- onto a 24 × 60 mm glass cover-slip (Fisher Scientific, Cat. ucts. Figure 1 outlines the main features of this successful No., 12-548-5P) and the pre-warmed slide was inverted construction project and Table 1 highlights some of the and placed with the arrayed surface contacting the sample key parameters that we tested during the array on the cover-slip surface. Deionized water (10 µl) was construction. added to each of the reservoirs of a Telechem Hybridiza- tion Chamber, and the slide was transferred to and Critical design features of the chimeric first round primers enclosed within the chamber. The slide was incubated in were the inclusion of the directional cloning and expres- a static incubator at 42°C for 12–16 h, and washed by sion features, plus the inclusion of a good common tem- placing in a 250 ml solution of 2X SSC and 0.1% SDS at plate sequence for the subsequent second round 42°C for 5 minutes with gentle agitation provided by amplification. The adaptamers included the introduction rotation of a magnetic stir bar. The slide was transferred of G+C rich sequences at the 5' termini to stabilize the end quickly to a solution of 0.1X SSC, and 0.1% SDS, incu- structure of the first round PCR products. Another impor- bated for 10 minutes at room temperature with gentle agi- tant design aspect that proved critical, according to our tation, and washed 5 additional times in 0.1X SSC for 1 preliminary tests, was incorporation of an additional five min. at room temperature. The slide was then rinsed base G+C rich overhang in the universal primers used for briefly with deionized water and dried by low speed cen- the second round PCR (Figure 2). The absence of the addi- trifugation. The slides were stored in the dark until tional sequences resulted in very poor PCR performance scanned. compared to the excellent performance observed with their inclusion. Results and Discussion The goal of this project was to develop a DNA microarray Importantly, the two-stage amplification had the unantic- as part of an overall strategy for functional genomics in ipated advantage of normalizing the yield of PCR prod- the unicellular model cyanobacterium, Synechocystis ucts across the entire set of genes, such that lower than st PCC6803. The two-stage amplification strategy (6) was average yielding products from the 1 round PCR (using adopted as a means to facilitate replication of the arrays as the gene specific sequence and chromosomal DNA as nd needs dictated (Figure 1). In general terms, the optimiza- template) often had a robust yield during the 2 round tion of this technical approach required particular atten- of amplification using the common adaptamer sequences. tion to the design of the adaptamers, the choice and use of Although the primers were deliberately designed in such a the thermostable DNA polymerase, and the thermocy- way that oligo T 's fell in narrow range, it was determined cling conditions. Other technical factors, such as the array after a number of trials that it was necessary to use printing conditions, were found to conform to observa- 'touchdown' PCR conditions for the first round amplifica- Page 6 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 Figure 3 Thermal cycling conditions and resultant products from first and second rounds of amplification. Panel A illus- trates the thermal cycling conditions for the primary and secondary amplifications. During the first round amplification (Panel A, left side) with gene-specific chimeric primers, step-down thermal cycling PCR conditions involve decreasing the annealing temperature by 1°C per cycle beginning at 62°C for 10 cycles and the subsequent 24 cycles have a set annealing temperature of 52°C. The graph on the right depicts the conditions for constant annealing temperature of 55°C for the second round ampli- fication using a common set of adaptamers primers complementary to the entire set of first round products (Panel A, right side). Panel B illustrates one half (48 reactions) of the results from the primary (above) and secondary (below) amplification of one 96-well plate. Lanes 2 and 6 (arrows 1 and 2) illustrate the product 'leveling effect' on the yields of product due to the sec- ondary amplification with adaptamer primers. Lanes 9, 14, and 35 (arrows 3, 4, and 7) illustrate missing or double products which are scored as bad and have new primers designed for subsequent use in another PCR run. Lanes 17 and 34 (under arrows 5 and 6, respectively) contain molecular weight markers. Page 7 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 Table 2: Synthesis and hybridization of fluorescently labeled cDNA derived from sample RNA Hybridization of the microarray with sample representing the mRNA population is accomplished by synthesizing complementary DNA (cDNA) using reverse transcriptase. Microarray experiments in prokaryotic organisms cannot utilize the poly(A) tails found in eukaryotic mRNA; thus, alter- natives to the highly successful methods employed for generating fluorescent cDNA using oligo-dT to prime the reverse transcription labeling reac- tion have to be employed. • cDNA synthesis Tests comparing random hexamers and random octamers gave essentially indistinguishable results. On the other hand, critical parameters for maximizing signal-to-noise were: 1.) high level of incorporation of fluorescent dye by using a high ratio of amino allyl dUTP in the reverse transcription reaction and saturating amount of reactive dye during the coupling of activated dye; and 2.) the use of two purification steps relying on independent physical separation principals applied to the test sample: first after cDNA synthesis and second, after the coupling of fluorescent dye. Direct chemical labeling of RNA, while attractive, did not allow as high signal-to-noise during our tests. Similarly, pilot exper- iments using pooled 3' gene-specific oligos as cDNA primers did not give better results. • Hybridization: While increases in signal-to-noise were associated with optimization of fluorescent sample production (above), a fairly wide range of standard microarray pre-hybridization and hybridization protocols were found to give comparable results. However, critical aspects of the described method were keeping the microarray slide warm during the addition of sample to the surface following pre-hybridization and ensuring that the slide remains hydrated prior to the final drying immediately before scanning. tion. Nevertheless, the percentage of successful products the PCR amplification are available at: http://microbiol- after the first round synthesis appeared quite low, based ogy.okstate.edu/faculty/burnap/index.html. on yield as visualized in the ethidium stained electro- phoretic gels. Some lanes showed little to no product or Organization and quality control were emphasized multiple products after the first round amplification with throughout production of the microarrays to ensure uni- the chimeric primers. However, these products generally formity and accuracy. A semi-automated gel analysis pro- resulted in successful amplification after their use as tem- cedure scored products on the basis of size and absence of plate in the second round amplification when the univer- multiplicity. All PCR products were documented by dig- sal sequences introduced as overhangs of the chimeric ital photography and the images quantitatively analyzed primers during the first round amplification were used. using Total Lab™ Phoretics gel analysis software (Nonlin- Thus, the universal primers in the second round amplifi- ear Dynamics. Durham, North Carolina). This application cation provided the additional benefit of generating con- was found to be most accurate, user-friendly, and effi- sistent yields across most of the genes (see Figure 3). ciently handled dual tier gels. Furthermore, as with other Apparently, the amount of correct template produced by packages, it allowed simple export of the numerical results st the 1 round PCR, albeit often invisible during gel analy- to Excel spreadsheets. A database was developed to inte- nd sis, was usually sufficient to give high yields during the 2 grate information, including gene sequences, oligonucle- round of amplification using the common adaptamer otide design, PCR product length, etc. In addition, Excel sequences. Thus 86% of the products were successfully macros were developed to transform the aforementioned amplified during our initial round of two-stage PCR. information to produce lists for each plate, which were Primers for the 14% unsuccessful products were rede- then used to track and identify the location of the arrayed signed and synthesized, so that essentially all of the anno- elements under any condition, including the use of differ- tated genes were successfully amplified. Therefore, we ent array configurations. The scoring system, described in conclude that this two-stage amplification strategy for the the Materials and Methods section, was set up to identify purpose of PCR amplifying multiple products for genome successful PCR amplified products. scale applications is highly successful, robust and efficient. DNA purification was simply and efficiently performed using filtration-type purification plates. Early issues with The use of adaptamer sequences with identical T 's for the variation in printed spot morphology were resolved by nd 2 round of amplification optimized the yields, despite pre-rinsing the filters: apparently this removes material different product lengths and different template concen- acting as a surfactant and effects the spot morphology trations (which were dependent on the yield from the first and/or drying characteristics. Alternatives such as ethanol round reactions). The enzyme and thermocycling condi- or isopropanol precipitations were explored, but for the tions were also optimized. The choice of Platinum Pfx chosen PCR conditions, produced less reliable results and (Invitrogen) as the polymerase was determined to be cru- appeared to cause the co-precipitation of non-DNA mate- cial as was its use in conjunction with the proprietary reac- rial that interfered with downstream operations. The tion enhancer solution supplied with the buffer. Tests nature of this material remains undefined, although it with alternatives, including mixes of Taq and Pfu , yielded may be due to constituents of the PCR reaction proprie- less satisfactory results. All the images for the analysis of tary enhancer that is supplied with the Pfx polymerase. Page 8 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 Figure 4 Scatter-plot of differential hybridization of fluorescently labeled cDNA from a 3-hour salt shock treated cell culture versus an untreated sample. Log phase cultures were subjected to an upshift in NaCl concentration from ~25 mM to 650 mM. Syne- chocystis is halotolerant and grows at concentrations up to approximately 1.2 M NaCl. Fluorescently labeled (Cy-3) cDNA derived from total RNA extracted from cells exposed 3 hours to the higher salt concentration was co-hybridized with Cy-5- labelled cDNA derived from the culture immediately prior to the upshift (control cells). The full genome Synechocystis sp. PCC 6803 high-density microarrays contain DNA features printed in triplicate on glass slides. The gene names according to original annotation are behind the arrows. As indicated with arrows, signals from the replicated elements exhibited similarity in the estimated expression ratio. Each data point corresponds to a different gene address on the microarray and each gene is repli- cated at three separate and spatially distant addresses. The X and Y axes correspond to the normalized fluorescence intensity of fluorescence-labeled cDNA of the control sample (Y-axis, Cy-5, 532 nm fluorescence) and the 650 mM NaCl three hour time point (X-axis, Cy-3, 635 nm fluorescence). Page 9 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 The concentrations of the purified DNA fragments were not provided as high a hybridization signal-to-noise ratio determined (Molecular Devices Spectramax Plus) and a compared to the chemical coupling to cDNA approach. Qiagen Bio-Robot 3000 was used to dilute the DNA sam- ples to 500 ng/µl in 2X spotting solution. Tests showed Substantial effort was devoted to ensure spot uniformity, that the combination of Superamine slides and MSP high signal-to-noise ratio, and good statistical reproduci- printing buffer solution from Telechem provided the best bility. We believe that we have met these objectives, and combination of spot morphology, retention of DNA after these arrays relate favorably to the best of current prokary- processing of slides and signal-to-noise ratio after hybrid- otic and eukaryotic arrays. The strengths of these Syne- ization. Alternately, Clontech type II slides in chocystis arrays include: 1.) consistent and high DNA combination with the Telechem spotting buffer also gave concentrations in each of over 11,000 spots per array; 2.) good results in terms of printing and hybridization 3 replicates per probe per slide; and 3.) high signal-to- results. Tests with other buffer systems, e.g. 6X SSC and noise ratio. This last parameter is due both to the quality 50% DMSO, gave less satisfactory results under the speci- of the slides, the hybridization procedures that we utilized fied printing conditions. The synthesis and hybridization and the quality of the products printed. We have used of the fluorescently labeled cDNA derived from sample these arrays to develop robust statistical tools for the anal- RNA is outlined in Table 2. ysis of the microarray data. The analysis involves an ANOVA model that makes excellent use of the technical In order to generate uniform spot morphology and to replicates designed into each slide and a loop design maximize DNA concentration and consistency for each (Singh, McIntyre and Sherman, unpublished results). The spot across the array, all products were printed at the same availability of large quantities of DNA, the ability to print concentration. This was accomplished using the Qiagen many slides and the statistical tools will enable us to ana- Bio-robot 3000™ to resuspend each of the purified PCR lyze transcriptional changes under many environmental products to the same concentration, based on the results conditions and for many mutants. The two-stage amplifi- of the Spectramax UV analysis. These products were then cation procedure also provides full gene copies for tech- mixed with an equal volume of 2X MSP printing buffer niques like PCR-fusion mutagensis. This will permit many from Telechem and two aliquots were transferred via the comparisons among different conditions and should Bio-robot to separate printing plates. The use of this greatly facilitate functional genomics. In conclusion, the robotic transfer reduced the chance for human error, two-stage PCR amplification strategy detailed here is increased the rate of transfer (thus reducing the exposure demonstrated to be highly useful and undoubtedly can be of the products to desiccation), and increased the repro- applied to other bacterial species. ducibility of the transfer. Syto-61 staining was regularly used to maintain a high level of quality control during the Competing interests printing process. None declared. Several internal and external comparisons were analyzed Authors Contributions to demonstrate the effective use of the arrays. Internally, BLP, H-LW and RLB developed and optimized the PCR the use of triplicate spotting of all array elements and dis- primer design strategy, BLP, H-LW, AS, HW, and LI per- tribution of the triplicate elements in different regions of formed the PCR amplifications; HLA, JK, and GR per- the array allows the comparison of hybridization formed gel analysis and quality control, DP developed uniformity across the array, as depicted in Figure 4. Ini- informatics, BLP, AS, MD, and DWG conducted the tially, we determined the effectiveness of printed slides by microarray printing, BP, LAS, and RLB organized the hybridization with total RNA from control and salt- project and prepared the manuscript. stressed cells. The results were compared with that of the published report from Kanesaki et al. [11]. Similar pat- Acknowledgements The authors would like to acknowledge Dr. Prade, Dr. Cushman, Dr. Bon- terns of regulation of specific transcripts were observed hert, and Dr. Stintzi for their advice and input. This work was supported by and several examples of the differentially regulated genes the National Science Foundation (to RLB and LAS) and the Department of are indicated in Figure 4. Qualitatively similar results Energy (to LAS) and by the Grula Fellowship to BP as part of his thesis work using a direct RNA-dye coupling procedure can be through the OSU Foundation. achieved using the Micromax™ ASAP RNA Labeling Kit (Perkin Elmer Life Sciences, Inc. cat. No. MPS544) follow- References ing the manufacturer's suggested protocol. While it is sig- 1. Schena M, Shalon D, Davis RW and Brown PO: Quantitative mon- nificantly quicker and requires less manipulation, the itoring of gene expression patterns with a complementary DNA microarray Science 1995, 270:467-470. direct chemical labeling kit is more expensive, and pro- 2. Spellman PT, Sherlock G, Zhang MQ, Iyer VR, Anders K, Eisen MB, duces a final RNA product which is less stable than the Brown PO, Botstein D and Futcher B: Comprehensive identification of cell cycle-regulated genes of the yeast cDNA produced. Significantly direct labeling has thus far Page 10 of 11 (page number not for citation purposes) BMC Genomics 2003, 4 http://www.biomedcentral.com/1471-2164/4/23 Saccharomyces cerevisiae by microarray hybridization Mol Biol Cell 1998, 9:3273-3297. 3. Iyer VR, Eisen MB, Ross DT, Schuler G, Moore T, Lee JCF, Trent JM, Staudt LM, Hudson J., Jr., Boguski MS, Lashkari D, Shalon D, Botstein D and Brown PO: The transcriptional program in the response of human fibroblasts to serum Science 1999, 283:83-87. 4. White KP, Rifkin SA, Hurban P and Hogness DS: Microarray anal- ysis of Drosophila development during metamorphosis Sci- ence 1999, 286:2179-2184. 5. Hihara Y, Kamei A, Kanehisa M, Kaplan A and Ikeuchi M: DNA microarray analysis of cyanobacterial gene expression dur- ing acclimation to high light Plant Cell 2001, 13:793-806. 6. Richmond CS, Glasner JD, Mau R, Jin H and Blattner FR: Genome- wide expression profiling in Escherichia coli K-12 Nucleic Acids Res 1999, 27:3821-3835. 7. Bryant DA: The Molecular Biology of Cyanobacteria The Neth- erlands, Kluwer Academic Publishers; 1994. 8. Kaneko T, Sato S, Kotani H, Tanaka A, Asamizu E, Nakmura Y, Miya- jima N, Hirosawa M, Sugiura M, Sasamoto S, Kimura T, Hosouchi T, Matsuno A, Muraki A, Nakzaki N, Naruo K, Okumura S, Shimpo S, Takeuchi C, Wada T, Watanabe A, Yamada M, Yasuda M and Tabata S: Sequence analysis of the genome of the unicellular cyano- bacterium Synechocystis sp. strain 6803. II. Sequence deter- mination of the entire genome and assignment of potential protein-coding regions. DNA Research 1996, 3:109-136. 9. Williams JGK: Construction of specific mutations in Photosys- tem II photosynthetic reaction center by genetic engineer- ing methods in Synechocystis 6803. Methods in Enzymology 1988, 167:766-778. 10. He Q, Dolganov N, Bjorkman O and Grossman AR: The high light- inducible polypeptides in Synechocystis PCC6803. Expres- sion and function in high light J Biol Chem 2001, 276:306-314. 11. Kanesaki Y, Suzuki I, Allakhverdiev SI, Mikami K and Murata N: Salt stress and hyperosmotic stress regulate the expression of different sets of genes in Synechocystis sp. PCC 6803 Biochem Biophys Res Commun 2002, 290:339-348. Publish with Bio Med Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical researc h in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright BioMedcentral Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp Page 11 of 11 (page number not for citation purposes)

Journal

BMC Genomics – Springer Journals

Published: Jun 12, 2003

There are no references for this article.