© 2003 by The Society for Integrative and Comparative Biology
| ||||||||||||||||||||||||||||||||||||||||||||||||||||
The "ups" and "downs" in Using Subtractive Cloning Techniques to Isolate Regulated Genes in Fish1
1 Marine Resources Center, Marine Biological Laboratory, Woods Hole, Massachusetts 02543
 |
SYNOPSIS |
|---|
TOP
 SYNOPSIS INTRODUCTIONCONVENTIONAL CDNA SUBTRACTIONDIFFERENTIAL DISPLAY PCRSUPPRESSION SUBTRACTION...ITERATIVE PCR SUBTRACTIONReferences |
|---|
Over the last decade, subtractive cloning approaches have been used extensively to isolate genes that are up- or down-regulated under various conditions. These techniques have provided the foundation for many subsequent studies concerning gene function and regulation and, as such, have been valuable tools for many biological fields. Over the past 10 years, we have used different subtractive cloning approaches to isolate genes in fish that are regulated in relation to hormonal stimulation or the stage of ovarian maturation. These include conventional cDNA subtraction followed by library screening, differential display PCR, suppression subtraction hybridization, and more recently, iterative PCR subtraction. We continue to use these techniques for the isolation of new genes involved in physiological processes in fish and bivalve molluscs. Examples that illustrate the use of these different subtractive cloning techniques are described, including where possible the advantages and disadvantages of each. In addition, the use of ancillary methods (e.g., "Reverse Northerns") to facilitate the use of these subtractive approaches are discussed.
|
INTRODUCTION |
|---|
|
TOP
SYNOPSISINTRODUCTIONCONVENTIONAL CDNA SUBTRACTIONDIFFERENTIAL DISPLAY PCRSUPPRESSION SUBTRACTION...ITERATIVE PCR SUBTRACTIONReferences |
|---|
Over the past decade, a number of methodological approaches have been developed to isolate regulated genes including conventional cDNA subtraction, differential display PCR (DDPCR), suppression subtraction hybridization (SSH), chemical cross-linking hybridization, and more recently DNA arrays. All of these techniques have advantages and disadvantages that must be considered when applying them to a given biological problem. In the past, there have been attempts to objectively compare some of these methods (Wan et al., 1996
) but there are many variables to consider. We have used conventional cDNA subtraction, DDPCR, SSH and more recently a process that we describe as "iterative PCR subtraction" to isolate genes from fish and bivalves. This paper describes these specific methods and some of the results that were obtained using them. Where possible, some discussion of the advantages and disadvantages are provided. |
CONVENTIONAL CDNA SUBTRACTION |
|---|
|
TOP
SYNOPSISINTRODUCTIONCONVENTIONAL CDNA SUBTRACTIONDIFFERENTIAL DISPLAY PCRSUPPRESSION SUBTRACTION...ITERATIVE PCR SUBTRACTIONReferences |
|---|
The basic goal behind subtractive cloning is to be able to isolate mRNA transcripts that are either novel or significantly up- or down-regulated between samples taken under different conditions (e.g., hormonal stimulation or time of the year). For most tissues, the majority of the mRNAs are not regulated and are, therefore, not of interest. Ideally, in the process of subtractive cloning these commonly expressed mRNAs are subtracted between the two stages so that the only genes remaining are either significantly regulated (up or down) or are novel to one sample or the other.
One of the earliest forms of subtractive cloning was to simply make a subtracted cDNA probe and to use that to screen a cDNA library. We used this approach to isolate genes that were upregulated at the time of ovulation in the brook trout ovary (Hsu and Goetz, 1995; Garczynski and Goetz, 1997). Messenger RNA was isolated from ovaries sampled prior to undergoing oocyte final maturation (prior to the resumption of meiosis—"premeiotic"), and just prior to the time of ovulation ("preOV"; Fig. 1). Therefore, we were interested in subtracting the mRNA between these two ovarian stages. The mRNA from the preOV sample was reverse transcribed to cDNA and was hybridized with a 10 fold excess of mRNA from ovaries taken prior to the resumption of meiosis. An excess of premeiotic mRNA was used to ensure that all of the mRNAs present in the premeiotic tissue would be removed from the preOV sample. The mRNA/DNA duplexes were then separated from the remaining single stranded cDNA (representing upregulated messages), labeled with [32P]dATP, and used to screen a cDNA library made from the same preOV mRNA as used in the subtraction. Positives from this library screen would presumably represent cDNAs that had not been subtracted. It should be pointed out that this screen could theoretically be conducted in reverse if one was interested in looking at the genes that were high prior to the resumption of meiosis and were then significantly downregulated at ovulation. In that case, all of the components in Figure 1 would be reversed so that an excess of mRNA from preOV tissue would be hybridized with cDNA from premeiotic tissue. The remaining cDNA would be used to screen a cDNA library made from premeiotic mRNA.
|
In the study on brook trout, screening of the preOV cDNA library with the subtracted probe yielded several positive clones, however, they were all of the same cDNA type. When cloned and sequenced we initially obtained a small cDNA of 732 bp that we called TOP-1 (trout ovulatory protein-1; (Hsu, 1995
). This cDNA had homology with a group of mammalian proteinase inhibitors called antileukoproteinases. When we used this cDNA to probe Northern blots we found that indeed this mRNA was very highly upregulated at the time of ovulation (Fig. 2). Later it was found that this mRNA is just one of a family of at least 4 TOP transcripts produced in the trout ovary at the time of ovulation that remain elevated for 2 days and then decrease (Garczynski and Goetz, 1997). We found that the transcripts are also translated into proteins that are released from the ovary at the time of ovulation and are found in high concentration in the ovarian/coelomic fluid that bathes the ovulated eggs (Garczynski and Goetz, 1997; Coffman et al., 2000). It was also shown that these proteins could act as serine protease inhibitors in the fluid and might have antimicrobial activity (Coffman and Goetz, 1998; Coffman et al., 2000). Certainly, given the very high upregulation observed in the TOP transcripts at ovulation, it is understandable why this was a primary cDNA obtained from the screen. However, isolation of only one type of cDNA also illustrates a major disadvantage of this type of subtraction; it is not very sensitive and most likely will isolate only high copy, highly regulated transcripts. In addition, compared to some other subtraction methods, this requires a large amount of mRNA from each sample.
|
12 hr GVBD); within 24 hr following GVBD (
24 hr GVBD); at ovulation (OV); within 12 hr of ovulation ( |
DIFFERENTIAL DISPLAY PCR |
|---|
|
TOP
SYNOPSISINTRODUCTIONCONVENTIONAL CDNA SUBTRACTIONDIFFERENTIAL DISPLAY PCRSUPPRESSION SUBTRACTION...ITERATIVE PCR SUBTRACTIONReferences |
|---|
Differential display PCR (DDPCR) is actually not a subtractive cloning approach as described above for cDNA probe subtraction. Rather, the relative expression of mRNAs from different samples are compared using random PCR. Thus, in DDPCR both up and down regulation can be observed simultaneously. In DDPCR, mRNA is obtained from samples that are being compared, and first-strand cDNA is synthesized from this mRNA (Fig. 3). The cDNA is then subjected to a number of PCR reactions with relatively small primers that are meant to amplify the 3' region of the transcripts. For these reactions, an anchored primer that will bind to the poly A tail on the 5' end is used with another primer of arbitrary sequence that will bind further upstream. These reactions are usually performed in the presence of a label such as [33P]dATP for visualization, and the reactions are separated on large polyacrylamide gels. After separation, the gels are exposed to X-ray film and then realigned with the original dried acrylamide gel. There are generally many bands that are observed on the gel with a given primer pair, but only a small fraction of the bands will be differentially expressed between samples. Using the aligned X-ray film, these areas of the dried gel are cut and then the DNA is eluted, PCR amplified using the same primer pairs, cloned and sequenced. Because the original bands are anchored to the poly A tail and are generally small 100–600 bp, the full-length cDNAs must be obtained by library screening or RACE (Fig. 3). In addition, it is critical to determine if the band actually represents a mRNA that is truly regulated. Thus, some verification such as Northern or dot blots must be performed.
|
We have used DDPCR several times to obtain genes that are regulated in the fish ovary (Lee and Goetz, 1998
; Langenau et al., 1999; Bobe and Goetz, 2000a, b; 2001a). In one study, we were interested in the genes that might be upregulated by steroids in the yellow perch ovary. Past studies that we had completed showed that the progestational steroid, 17
, 20ß-dihydroxy-4-pregnen-3-one (DHP), stimulated both the resumption of meiosis and ovulation completely in vitro in perch ovaries (Goetz and Theofan, 1979). Later it was shown that DHP-induced ovulation required transcription (Theofan and Goetz, 1981). Thus, we were interested in looking at the genes that were activated by DHP in in vitro incubations. In the DDPCR study, 2 different poly T anchor primers were used in conjunction with 15 random primers (Langenau et al., 1999). This generated approximately 60 bands that appeared to be differentially regulated when analyzed by DDPCR. While nearly all of these bands could be amplified and cloned, only 5 cDNAs were eventually shown to be consistently regulated by DHP (Fig. 4). These cDNAs included homologs of calmodulin (accession #AF085250), neprilysin (accession #AF077612), lysyl oxidase-like protein (accession #AF143003), microtubule aggregate-like protein (accession #AF085251) and an egg protein related in structure to amphibian D7 (accession #AF085252).
|
The experimental use of DDPCR has been facilitated by the presence of several companies that supply specific reagents and technical advice. There is no question that DDPCR is a very sensitive method to investigate regulated genes. The primary limitations with this system appear to be the generation of a large number of false positives (though techniques and reagents may be available to reduce this see: http://www.see-gene.com/product/tr_deg_101.htm) and the number of primer pairs (and reactions) needed to cover all possible transcripts expressed in a tissue. Since DDPCR uses arbitrary primers, it is necessary to cover a very large number of these primer combinations so that all transcripts in a tissue would be analyzed. This number has been theoretically calculated (http://www.genhunter.com/support/) and could constitute a large number of reactions for some investigators. The generation of false positives in DDPCR may come at several levels in the process. One way we have found to reduce this number initially is to conduct the DDPCR on replicate samples rather than using RNA from pooled samples. Obviously, since DDPCR is based on PCR, one sample may greatly bias the results of a pool. However, in some cases it may be impossible to work with individual samples because of limitations in the quantity of mRNA. Finally, new techniques and reagents provided by companies such as GenHunter® and Seegene® may have circumvented some of the more onerous steps in DDPCR such as the use of radioactive labels and polyacrylamide gel separation.
|
SUPPRESSION SUBTRACTION HYBRIDIZATION |
|---|
|
TOP
SYNOPSISINTRODUCTIONCONVENTIONAL CDNA SUBTRACTIONDIFFERENTIAL DISPLAY PCRSUPPRESSION SUBTRACTION...ITERATIVE PCR SUBTRACTIONReferences |
|---|
Suppression subtraction hybridization (SSH) has been a very popular subtraction method in the last 5 years and is available as a kit (PCR Select®) from BD Biosciences–Clontech. In SSH, mRNA is obtained from two samples that are being subtracted and first and second strand cDNA is synthesized for each (Fig. 5). This double strand DNA is then cut with Rsa I to produce smaller pieces of DNA that more efficiently and consistently hybridize and amplify using PCR. Suppression subtraction hybridization can subtract in either direction (i.e., up- or down-regulation) but the steps following the cDNA synthesis will depend on which direction is being investigated. For example, we used SSH to look at genes that were up and downregulated after ovulation in the brook trout ovary (Bobe and Goetz, 2001b
). If we were interested in the genes that were down-regulated following ovulation, then we would take the cDNAs generated from the mRNA obtained at ovulation and divide them into two equal pools (Fig. 5). These pools become known as the "testers" and different adaptors are then ligated to the cDNA pools. The cDNA derived from the mRNA of the postovulatory tissue becomes the "driver" and is restriction digested but not ligated with adaptors. In separate reactions, an excess of driver is hybridized with each tester pool (Fig. 5). This hybridization removes transcripts in the tester that are found in the driver and results in the cDNA arrangement given in Figure 5. A second round of hybridization is performed adding both testers together with new driver. After filling in the overhanging ends, this is subjected to a first PCR reaction and then a second PCR reaction using nested primer sites within each adaptor. Ultimately, only DNA molecules that have both of the adaptors will be amplified by PCR in the nested reaction. If one is looking for genes that would be upregulated following ovulation, then the cDNAs derived from this tissue would be used as the testers and the ovulatory cDNA would be the driver.
|
We used SSH as a subtractive technique after we had already investigated the regulation of genes using conventional subtraction and DDPCR. Thus, several genes that we had already observed to be regulated following ovulation in trout were also obtained using SSH (e.g., TOPs downregulated postOV), confirming that it was isolating regulated genes. In addition, we obtained another gene that we had not previously isolated (Bobe and Goetz, 2001b
). This cDNA was a homolog of mammalian osteopontin (accession #AF204760), a gene that has recently been shown to be produced by white blood cells and be involved in inflammation.
An important consideration with SSH is what to do with the cDNAs that are produced after subtraction and PCR. It is possible to simply clone them all and then to start randomly sequencing them, and this approach has been reported in the literature. However, there may be two problems in taking this path. The first is gene redundancy. If there are a few genes that are highly differentially expressed, then they will appear in the library a large number of times and my mask the isolation of transcripts that are still differentially expressed but at a lower overall level. The second is whether all of these genes are truly regulated between the two samples. We found that even with the subtraction process, there were still a number of false positives in the PCR mixture at the end. Thus, we developed a "Northern Capture" technique to confirm which products were truly regulated and to clone those products at the same time (Bobe and Goetz, 2001b). To perform this capture, we ran Northern gels of mRNA obtained from individual fish at ovulation and postovulation (3 individuals/stage). These gels were run for longer times to increase the overall separation of transcripts. We then hybridized the blot with a probe that we made during the second (nested) PCR reaction of the SSH. This reaction was run in the presence of [32P]dATP. After exposure, the X-ray film was realigned with the Northern and areas on the blot that corresponded to bands that were consistently regulated across all samples were cut and the DNA eluted. This DNA was then amplified by PCR with the nested primers used to make the original probe and was then cloned and sequenced. We found that while there were a number of bands on the Northern that were regulated, a number were not and presumably would have been false positives. Dot blot analysis (as described by Clontech) can also be used to verify whether clones from the subtracted library are truly regulated. The important point is that some verification/sorting of the subtracted DNA that was amplified by PCR is necessary since it cannot be assumed that all of it is regulated.
|
ITERATIVE PCR SUBTRACTION |
|---|
|
TOP
SYNOPSISINTRODUCTIONCONVENTIONAL CDNA SUBTRACTIONDIFFERENTIAL DISPLAY PCRSUPPRESSION SUBTRACTION...ITERATIVE PCR SUBTRACTIONReferences |
|---|
A "gene expression screen" to uncover genes regulated during amphibian metamorphosis, was reported by Wang and Brown in 1991 (Wang and Brown, 1991
). The technique used for this screen was a subtractive hybridization procedure that involved sequential rounds of hybridization and PCR. The specific subtraction technique was not given a name by these investigators, but in view of the number of rounds of hybridization and PCR, I have termed this "iterative PCR subtraction." In this technique, mRNA is isolated and first and second strand DNA is synthesized as with SSH and then cut with a restriction enzyme such as Rsa I or Alu I (Fig. 6). However, unlike SSH, an adaptor is ligated to all of the cut DNA This adaptor is made from two complimentary oligos; [5'CTCTTGCTTGAATTCGGACTA3'] and [3'ACACGAGAACGAACTTAAGCCTGAT5'] and contains an EcoRI site close to the flush end that ligates to the cDNA. After ligation, the DNA is amplified using the adaptor primers to produce ug quantities for subtraction. For each sample, the amplified DNA is divided into two pools: a smaller "trace" and a larger quantity that will be used as the "driver." The drivers are then cut with EcoRI to remove most of the adaptor so that they will not be amplified following each round of hybridization with the tracer, and then biotinylated (Vector-Photoprobe Biotin) so that they can be selectively removed by incubating with streptavidin.
|
Tracer and driver preparations are constructed separately as described above for each sample. Pairs of tracers and drivers from 2 treatments are then used for subtraction as outlined in Figure 7. The basic premise of the subtractive procedure is that hybridization between a driver of one treatment and a tracer of another treatment will remove genes that are common to both driver and tracer. These genes can be completely removed from the hybridization mixture because the driver is biotinylated. Streptavidin is added to the hybridization and it binds to the biotinylated driver. This is then removed by phenol/chloroform extraction. In addition, any nonhybridized driver that is left behind will not PCR because most of the adaptor linker was removed by EcoRI digestion prior to biotinylation.
|
As indicated in Figure 7 subtraction between treatments can procede in both directions since individual drivers and tracers are made for each treatment. This allows for the isolation of genes that are both up- and down-regulated between treatments. Following streptavidin treatment and phenol/chloroform extraction, the mixture is then subjected again to PCR amplification using the adapter oligo as a primer. Because the driver has been removed, this primarily amplifies the remaining tracer. In reality, the hybridizations used originally by Wang and Brown (1991)
are more complex than indicated in Figure 7. Each hybridization actually consists of a long (20 hours) and a short (2 hours) hybridization. The driver used in each short hybridization is always the first driver produced from the original cDNA (as shown in Fig. 6), while the drivers used for the long hybridization are newly produced at each subsequent round of subtraction. The reason for using different drivers/hybridization times, is that the short hybridization is more efficient in subtracting out the more abundant and common transcripts, while the long hybridization is necessary to subtract out the less abundant but commonly expressed transcripts following each round of subtraction (Wang and Brown, 1991).
The entire process of driver and tracer hybridization (long and short), driver removal, PCR amplification and new driver construction, is repeated through 3–6 rounds of subtraction. At each round, the PCR products obtained at the end can be observed on a gel (e.g., Fig. 8) to look at which products are enriched in each treatment. These bands will become more distinct with each round and can be directly cut from the gel and cloned, and the entire PCR can also be cloned as with SSH. An advantage of this subtraction procedure is that any cDNAs can be subtracted from the tracer by simply biotinylating them and using them as a driver for hybridization. Thus, once highly regulated messages are cloned, they can be biotinylated and subtracted from a given tracer. This has been demonstrated to uncover and enhance the isolation of regulated messages of much lower abundance (Wang and Brown, 1991; Kanamori, 2000). In addition, this reduces the redundancy for genes that are isolated following subtractions with other activating agents or activation times.
|
There are a number of techniques available to isolate regulated genes. The choice of technique is somewhat dependent on the desired result and how much RNA can initially be obtained for subtraction. However, the more recent approaches such as SSH and DDPCR are clearly more efficient and sensitive than the original techniques involving cDNA subtraction and library screening. Iterative PCR subtraction is probably the most inclusive subtraction technique, and because of repeated hybridizations, may result in the lowest overall number of false positives. The major disadvantage for the routine use of this technique is that reagents and protocols are not available as a kit as they are for SSH, and it may be the most technically challenging subtraction approach.
|
FOOTNOTES |
|---|
1 From the Symposium Contemporary Approaches to Endocrine Signaling presented at the Annual Meeting of the Society for Integrative and Comparative Biology, 4–8 January 2003, at Toronto, Canada.
|
References |
|---|
|
TOP
SYNOPSISINTRODUCTIONCONVENTIONAL CDNA SUBTRACTIONDIFFERENTIAL DISPLAY PCRSUPPRESSION SUBTRACTION...ITERATIVE PCR SUBTRACTIONReferences |
|---|
Bobe, J., and F. W. Goetz. 2000a. A S100 homologue mRNA isolated by differential display PCR is down-regulated in the brook trout (Salvelinus fontinalis) post-ovulatory ovary. Gene, 257:187-94.[Medline]
Bobe, J., and F. W. Goetz. 2000b. A tumor necrosis factor decoy receptor homologue is up-regulated in the brook trout (Salvelinus fontinalis) ovary at the completion of ovulation. Biol. Reprod, 62:420-6.
Bobe, J., and F. W. Goetz. 2001a. A cysteine protease inhibitor is specifically expressed in pre- and early vitellogenic oocytes of the brook trout ovary. Mol. Reprod. Dev, 60:312-318.[CrossRef][ISI][Medline]
Bobe, J., and F. W. Goetz. 2001b. A novel osteopontin-like protein is expressed in the trout ovary during ovulation. FEBS Lett, 489:119-24.[CrossRef][Medline]
Coffman, M. A., and F. W. Goetz. 1998. Trout ovulatory proteins are partially responsible for the anti-proteolytic activity found in trout coelomic fluid. Biol. Reprod, 59:497-502.
Coffman, M. A., J. H. Pinter, and F. W. Goetz. 2000. Trout ovulatory proteins: site of synthesis, regulation, and possible biological function. Biol. Reprod, 62:928-38.
Garczynski, M. A., and F. W. Goetz. 1997. Molecular characterization of a ribonucleic acid transcript that is highly up-regulated at the time of ovulation in the brook trout (Salvelinus fontinalis) ovary. Biol. Reprod, 57:856-64.[Abstract]
Goetz, F. W., and G. Theofan. 1979. In vitro stimulation of germinal vesicle breakdown and ovulation of yellow perch (Perca flavescens) oocytes. Effects of 17-hydroxy-20ß-dihydroprogesterone and prostaglandins. Gen. Comp. Endocrinol, 37:273-285.[Medline]
Hsu, T., and F. W. Goetz. 1995. Ovulation specific transcription of antileukoproteinase-like mRNAs in the fish ovary. In J. F. S. S. Fujimoto, A. J. W. Hsueh, and T. Tanaka (ed.), Frontiers in endocrinology, Vol. 13, New achievements in research of ovarian function, pp. 183–189. Ares-Serono Symposium Publications, Rome, Italy.
Kanamori, A. 2000. Systematic identification of genes expressed during early oogenesis in medaka. Mol. Reprod. Dev, 55:31-6.[CrossRef][ISI][Medline]
Langenau, D. M., F. W. Goetz, and S. B. Roberts. 1999. The upregulation of messenger ribonucleic acids during 17, 20ß-dihydroxy-4-pregnen-3-one-induced ovulation in the perch ovary. J. Mol. Endocrinol, 23:137-52.[Abstract]
Lee, P. H., and F. W. Goetz. 1998. Characterization of a novel cDNA obtained through differential-display PCR of phorbol ester-stimulated ovarian tissue from the brook trout (Salvelinus fontinalis). Mol. Reprod. Dev, 49:112-8.[CrossRef][Medline]
Theofan, G., and F. W. Goetz. 1981. The in vitro effects of transcriptional and translational protein synthesis inhibitors on final maturation and ovulation of yellow perch (Perca flavescens) oocytes. Comp. Biochem. Physiol, 69A:557-561.[CrossRef]
Wan, J. S., S. J. Sharp, G. M. Poirier, P. C. Wagaman, J. Chambers, J. Pyati, Y. L. Hom, J. E. Galindo, A. Huvar, P. A. Peterson, M. R. Jackson, and M. G. Erlander. 1996. Cloning differentially expressed mRNAs. Nat. Biotechnol, 14:1685-91.[CrossRef][ISI][Medline]
Wang, Z., and D. D. Brown. 1991. A gene expression screen. Proc. Natl. Acad. Sci. U.S.A, 88:11505-9.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||