CABIOS COMMUNICATION

CABIOS COMMUNICATION

Vol 7 no 4 1991 Pages 5 2 5 - 5 2 9

An improved microcomputer program for finding gene- or gene family-specific oligonucleotides suitable as primers for polymerase chain reactions or as probes K.Lucas, M.Busch1, S.Mossinger and J.A.Thompson* important new feature allows the alignment of several DNA sequences in a single search, whereby gene family-specific, We present here an easy-to-use computer program which finds group-specific, or individual gene-specific oligonucleotides can oligonucleotides suitable as primers in polymerase chain be identified. reactions (PCR) or as probes for hybridization. In contrast to The development of this program was initiated by problems other programs used for this purpose, the additional advantage arising in our own group. Our work is concerned with research of this one is the possibility of directly detecting gene- as well on the carcino-embryonic antigen (CEA) gene family as gene family-specific oligonucleotides. For this purpose, up (Thompson and Zimmermann, 1988), which consists of at least to 200 different DNA sequences, of maximally 65 000 19 members (Barnett and Zimmermann, 1990), and can be nucleotides each, can be scanned in a single search to ensure subdivided into two subgroups. One subgroup encodes the either single or multiple gene binding of the PCR primers or classical CEA cross-reacting antigens and a second encodes the probes. Specific oligonucleotides for genes carrying internal pregnancy-specific glycoproteins (PSG). The sequence repetitions and for single genes belonging to a set of highly similarity between the individual genes is extremely high conserved genes can also be detected. Many parameters such [>90% sequence similarity at the nucleotide level for some as exclusion of simple sequences, which are known to be highly PSG genes (Streydio etal, 1990)]. Furthermore, highly repeated throughout various genomes or regions of stable conserved repeating units are found within single genes e.g. secondary structures in both primer—primer and CEA (Thompson and Zimmermann, 1988). We are interested primer—template, can be taken into consideration and avoided. in designing oligonucleotide primers for amplification of specific Furthermore, the G + C content and the length of the gene products, as well as total gene-subgroup products (e.g. oligonucleotides can be changed in a broad range by the user. a primer pair for all PSG subgroup members). The program which we have developed is able to solve the Introduction problems which we are faced with, whereby up to 200 gene The polymerase chain reaction (PCR) has rapidly developed sequences can be compared in a single search. It is very versatile to become one of the most powerful techniques used for and offers special benefits to the design of oligonucleotides for amplification of DNA fragments in molecular biology today analysing gene families or genes containing internal repeated (Saiki etal., 1988). The main problems faced by PCR are: structures. (i) the selection of suitable oligonucleotides as primers (Rychlik etal., 1989; Pallansch etal., 1990; Lowe etal., 1990); (ii) System and methods choosing the optimal reaction conditions (Innis et al., 1990) and; Hardware requirements (iii) the possibility of unspecificity due to the enormous sensitivity of PCR (Kitchin et al., 1990). To solve the problems The program runs on IBM PC/XT/AT and PS/2 or 100%" concerned with the selection of suitable oligonucleotides, we compatible computers under MS-DOS 3.0 or higher. A present here a computer program that includes different rules minimum of 512 kbyte RAM and a floppy disk drive is required, for detecting primers suitable for PCR. A partially similar but for reasons of performance a 640 kbyte RAM and a hard computer program has recently been described (Lowe et al., disk drive is recommended. 1990). However, in contrast, the program presented here includes algorithms for exclusion of second binding of the Source of tissues and RNA isolation primers to the DNA or RNA sequence to be amplified, which Total RNA was isolated using the acid phenol extraction would lead to secondary bands. Furthermore, the most procedure, described by Chomczynski and Sacchi (1987) from a human colon tumor xenograft grown in a nude mouse, a InstitutfQr Invnunbiologie der University Freiburg. D-7800 Freiburg im Brsg.. Stefan-Meier-Strasse 8 and 'Max Planck Institul fiir Immunbiologie. SlUbeweg mouse liver and rat placenta from the 20th day of pregnancy. Poly(A) RNA from the rat placenta was purified once over an 51, D-7800 Freiburg, FRG oligo(dT) column according to Aviv and Leder (1972). *To whom reprint requests should be sent Abstract

525

K.Lucas et at.

Reverse transcription and polyme rase chain reactions Reverse transcription and polymerase chain reactions were slightly modified after Sambrook et al. (1989). The modifications were as follows: All reactions were carried out in 50 or 100 p\ total volume. For the reverse transcription step, the specific 3' oligonucleotide selected by the program was used to prime DNA synthesis. After denaturation of the reverse transcriptase at 93 °C, the corresponding 5' oligonucleotide was added, and PCR was performed over 300 cycles using a Biomed Thermocycler 60 (Bachhofer, FRG). The melting temperature was 93°C throughout, and annealing times were 15 s. Reverse transcriptase, RNasin. Taq polymerase and the amplification buffer were from Promega, and the dNTPs were supplied by Pharmacia. Algorithms and program description The program was written in Turbo Pascal 5.5. (Borland Inc.). To improve execution speed, time critical routines were implemented in Assembler language (TASM 1.0, Borland Inc.). For designing oligonucleotides specific for individual genes or groups of homologous genes, up to 200 different DNA sequences of maximally 65 000 nucleotides each can be compared. The sequence file format must be one of the following: EMBL (according to UPAC), PC-GENE (Genofit, Geneva, Switzerland), FlexP (developed by M.Kroger, Freiburg University) or ASCII (both capital and lower case letters for bases, as proposed in UPAC; no comments allowed). The program automatically recognizes the actual file format. After a template DNA sequence is selected and all parameters are adjusted, the search can be started. The program generates a list of all possible sequences according to the given range, chosen within the template sequence and the oligonucleotide length limits. Each oligonucleotide is successively tested with regard to various parameters, as listed below. Any oligonucleotide that fails one of the tests is removed from the original list. Checking the remaining oligonucleotides for their binding or non-binding to selected DNA sequences (see parameters 2 and 3) is the most time-consuming step and is, therefore, carried out at the end. In this manner, each remaining oligonucleotide is tested with all selected DNA sequences. After all these tests a compatibility report is generated. This report contains, for each oligonucleotide in one list, the numbers of all compatible partners in the other list. This compatibility considers the positions of the primers within the template DNA sequence, their annealing temperatures (the difference may not exceed 5°C), primer—primer homology, and a possible 3' end annealing between both primers. The following parameters can be adjusted to obtain suitable oligonucleotides: 1.

526

The ranges within which the strand and the complementary strand primers should bind to the target sequence must be determined. This is especially

important when analysing genomic sequences for mRNA amplification purposes, to avoid primers located in introns. 2. A list of DNA sequence files can be selected, to which the oligonucleotides must bind uniquely in strand or complementary strand, respectively. The degree of binding can be varied through the parameter of maximal allowed mismatches. In this way, primers that are common to different sequences, but bind only once within one sequence may be obtained. 3. To obtain primers which are monospecific to the sequence they are derived from, a list of DNA sequence files can be selected to which the primers should not bind. The user can adjust the minimal number of required mismatches. 4. The length of oligonucleotides can be varied in a range of 1 - 5 0 bases. 5. The G + C quotient can be varied in a range from 0-100%. 6. To avoid simple sequences within the oligonucleotides the minimum percentage content of each base can be determined. 7. To exclude long single bases repetitions (e.g. AAAAA) a parameter designed as 'maximum single base repetition' can be selected. 8. To exclude those primers, which might bind more than once to the target DNA sequence, the required minimum number of mismatches to avoid second binding to strand and complementary strand can be varied. This important parameter must be as high as possible so as to ensure specificity of the oligonucleotide. 9. All oligonucleotides which do not have C or G in the two last two positions at their 3' end can be excluded. It has been reported elsewhere that G — C bonds at this position in an oligonucleotide will facilitate the initiation of complementary strand formation by reverse transcriptase and Taq polymerase (Lowe et al., 1990). However, this parameter is optional and can be switched off. 10,11. Two parameters are included to avoid choice of primers which can form internal, stable, secondary structures (primer foldback). The maximal allowed internal sequence similarity as a percentage and the maxima] allowed length of an internal stem can be selected. 12. To avoid self-annealing between sister oligonucleotides, the maximum number of continuous matches between the 3' end of a primer and any internal position can be adjusted. 13. The maximum similarity between the strand and the complementary strand primer can be selected to avoid non-sister annealing. This parameter is of importance for the generation of compatibility lists.

Selecting oligonucleotides for PCR of gene families

Implementation and results In order to check the reliability of this program, oligonucleotides were chosen for PCR, and their specificities were determined empirically. As the studies in our laboratory are directed toward analysing different members of the highly conserved CEA gene family (Thompson and Zimmermann, 1988), it is important in selecting oligonucleotides, to compare sequences of all known members to ensure specificity. We have chosen PCR for studying the expression of different CEA family members in both humans and rodents. As a control for the integrity of the isolated RNA, oliognucleotides have beendesigned using this program, to amplify portions of the ubiquitous /3-actin mRNA. In order to ensure that the amplification products from RNA preparations were not due to contaminating DNA, the oligonucleotides for all PCR reactions were chosen so that at least one intron region would be present in between them. Design of evolutionarily conserved (3-actin oligonucleotides We first designed a set of oligonucleotides to bind to /S-actin RNA from three species, to determine the integrity of the RNAs. For this purpose, /3-actin sequences from man (Ponte etal, 1984), rat (Nudel et al., 1983) and mouse (Tokunaga et al., 1986) were compared, and the following oligonucleotides, with no mismatches to any of the three species, were chosen: Strand oligonucleotide, 5'-TCTACAATGAGCTGCGTGTGG-3'; Complementary strand oligonucleotide, 5'-AGTACTTGCGCTCAGGAGG-3'. The results of the PCR analyses using them as primers can be seen in Figure la. It is obvious that one strong amplification product is visible with the expected length of 744 bp for RNA isolated from all three

Fig. 1. Specificity tests of selected oligonucleotides primer pairs using PCR reactions. The templates used were, (a,b) human and mouse total RNA (200 ng/reaction) or poly(A)RNA from rat placenta (150 ng/reaction), (c,d) or selected cDNA probes, as listed (20 ng/reaction). The PCR products were separated according to size by electrophoresis on agarose gels, stained with ethidium bromide and photographed under UV illumination, (a) PCR reaction products using /3-actin oligonucleotide primers, whose sequences are conserved between human, rat and mouse (annealing temperature, 56°C). (b) PCR reaction products with human /3-actin specific oligonucleotide primers (annealing temperature 64°C). (c) PCR reaction products using CEA gene-specific and CEA subgroup-specific oligonucleotide primers (annealing temperature, 56°C) (d) PCR reaction products using PSG subgroup-specific oligonucleotide primers (annealing temperature, 56°C)

species, although some background smear, which could be due to weak cross-reactivity with other transcripts, is also visible. Design of species-specific fi-actin oligonucleotides In an attempt to get more specific signals, and to test the versatility of the program, we decided to search for speciesspecific oligonucleotides for the human /3-actin. Although /3-actin sequences are conserved between species, regions of divergence also exist. With the help of our program, we have also designed a set of oligonucleotides for PCR, which should only recognize and amplify human, but not mouse or rat, /3-actin mRNAs. The oligonucleotides chosen have the following sequences: Strand oligonucleotide, 5'-ACGGCTGCTTCCAGCTCCTC-3'; Complementary strand oligonucleotide, 5'-AGCCATGCCAATCTCATCTTGT-3'. These oligonucleotides have at least 5 mismatches to the rat and mouse /3-actin mRNA sequences. The results of the PCR analyses using them as primers can be seen in Figure lb. In this case, a strong amplification product with no smear and the expected size of 525 bp is only visible for human, but for neither rat nor mouse RNA. Design of CEA gene-specific and CEA subgroup-specific oligonucleotides for PCR In order to design oligonculeotides for the specific amplification of portions of the CEA mRNA, the sequences of all known members belonging to the CEA gene family have been compared using our program, and the two oligonucleotides selected have the following sequences: strand oligonucleotide, 5'-CCATGGAGTCTCCCTCG-3'; complementary strand oligonucleotide, 5 '-GGAATTCATCAGCAGGGATGCATTGG-3'. The 3'-oligonucleotide chosen is not gene-specific, but could bind with the mRNAs of all members of the CEA subgroup, with no mismatches, apart from CGM2 (Thompson etal., 1989), where two mismatches are found. This oligonucleotide was originally synthesized together with a CEA subgroup-specific strand oligonucleotide (sequence not shown) for amplification of DNA fragments directly for subcloning. For this reason, a linker containing an EcoRl cloning site (bold letters), was included. In order to test the specificity of this oligonucleotide together with the new CEA-specific, strand oligonucleotide, which reveals at least 4 mismatches to other members of the CEA gene family, they have been used for PCR with cDNA clones for individual members of this gene family. Despite the fact that one oligonucleotide is not CEA genespecific, the results of the PCR analyses reveal an amplification product with the expected size of 335 bp only with the CEA (own clone: unpublished), but not with closely related NCA (Neumaier et al., 1988), CGM6 (Berling etal., 1990), BGP (Hinoda et al., 1989) or PSG3 (Rooney et al., 1988) cDNA clones (Figure lc). This shows that it is not necessary to generate two gene-specific oligonucleotides to achieve specificity.

527

K.Lucas et al.

Design of PSG subgroup-specific oligonucleotides As a final test for the reliability of this program, sequences common to all of the 11 known PSG genes (Barnett and Zimmermann, 1990), but divergent to members of the CEA subgroup were searched for, in order to design PSG subgroupspecific oliogonucleotides. The two oligonucleotides chosen have the following sequences: strand oligonucleotide, 5'-AGTTTCTGGATCCTAGGCT-3'; complementary strand oligonucleotide, 5'-GATGTAATGTAATGGTAGAGG-3'. These oligonucleotides reveal no mismatches to all known PSG genes, apart from PSG11 (Arakawa et al., 1990), where one mismatch each is found. However, they reveal minimally one (strand oligonucleotide) and seven (complementary strand oligonucleotide) mismatches, respectively, to members of the CEA subgroup. The results of the PCR are shown in Figure Id, and reveal an amplification product with the expected size of 295 bp only for the PSG3 cDNA clone (Rooney et al., 1988) tested, but were negative using cDNA clones from the CEA subgroup, i.e. CEA (own clone; unpublished) and NCA (Neumaier etai., 1988), as templates. As there are no nucleotide differences for all other PSG genes known, apart from PSG 11 (Arakawa et al., 1990), PCR with other PSG cDNA clones as templates was unnecessary.

Discussion Some other computer programs to detect oligonucleotides suitable as primers in PCR have already been presented (Rychlik et al., 1989; Pallansch et al., 1990; Lowe et al, 1990). The added advantages of the program discussed here are concerned with the problems arising from the work on many evolutionarily closely-related genes with high degrees of sequence similarity between the family members. Gene families are widespread, e.g. genes encoding CEA-related molecules (Barnett and Zimmermann, 1990), globins (Efstratiadis etai, 1980), immunoglobulin heavy and light chains (Tonegawa, 1983), homeotic proteins (Gehring, 1987), MHC class I and class II molecules (Steinmetz and Hood, 1983) etc. and genes coding for proteins with internal repetitions are not uncommon, e.g. CEA (Thompson and Zimmermann, 1988),fibronectin(Hynes, 1985), apolipoprotein (a) (McClean et al., 1987) etc. The most important advantage of this program is the included possibility of checking up to 200 different DNA sequences for binding or nonbinding. In this way, gene (Figure lb,c), or genefamily specific (Figure Id) oligonucleotides can be selected. A second advantage is the elimination of all primers which would bind more than once to the DNA they are derived from. This is especially important for genes which encode proteins with internal repeats such as CEA. If, however, oligonucleotides are needed to bind with all repeating units, then these must be entered as separate sequence files. With the algorithms realised in the program many problems are solved. Long palindromic sequences are excluded because each oligonucleotide is tested 528

for binding in the strand and complementary strand. When choosing a high value for the parameter of minimum mismatches to avoid second binding, specificity can be achieved. The problem of stable secondary structures in the template discussed by Pallansch et al. (1990) can also be avoided. This is because secondary structure is related to longer stretches of complementary sequences within one strand. These structures can be commonly formed in regions containing inverted repeats, simple sequences etc. Such regions are excluded by testing the complementary strand. Further, the higher the value for exclusion of second binding, the lower the possibility for PCR primer pairs to build stable dimers. Two other parameters, i.e. the internal homology and the longest acceptable stem length within the oligonucleotide, help to exclude those primers which would build stable internal secondary structures. A guideline for the optimal starting parameters and how to modify them is included with the program. Prior to oligonucleotide synthesis, additional controls may be made, if desired, by checking against complete databank sequences using other available programs. The calculation of the best annealing temperature for a given set of primers (Suggs et al., 1981) is only a theoretical one. However, it gives a good starting temperature which must be empirically modified for individual oligonucleotides, where the optima may be slightly higher or lower. In contrast to the program created by Lowe et al. (1990), this program does not offer pairs of primers, but a compatibility list of all possible partners for both strand and complementary strand oligonucleotides. This program works very quickly, due to the programming of the alignment routines in assembler language. It is not only useful for the detection of oligonucleotides for PCR, but also in the search for oligonucleotides to be used as specific probes in hybridization analyses, e.g. for identifying individual transcripts from highly conserved gene families. Alternatively, oligonucleotide probes could be designed to recognise and bind to common repeating sequences, such as Alu-elements, which could be useful in PCR amplification of genomic fragments, together with specific primers. This program 'Primer' is available, including a printed manual, for distribution for a fee of fifty US dollars. Acknowledgements We wish to acknowledge the help of Volker Reichardt with some of the PCR reactions and Dr Wolfgang Zimmermann for critical reading of the manuscript. A special thank you to Drs R.Paxton and N.Hardman for providing us with the BGP and PSG3 cDNA clones, respectively. This work was supported in part by the Dr Mildred Scheel Stiftung fOr Krebsforschung.

References Arakawa,F., Kuroki.M., Misumi.Y., Matsuo.Y. and Matsuoka.Y. (1990) The nucleotide and deduced amino acid sequences of a cDNA encoding a new species of pregnancy-specific 01-glycoprotein. Biochim. Biophys. Ada.. 1(M8. 303-305. Aviv.H and Leder.P. (1972) Purification of biologically active globin mRNA by chromatography on oligo thymidylic acid-cellulose. Proc. Nail. Acad. Sd. USA, 69, 1408-1412.

Selecting odgonucleoUdes for PCR of gene families

Bamett.T. and Zimmerniann.W (1990) Proposed nomenclature for the carcinoembryonic antigen (CEA) gene family: Results of the CEA workshop at the XVII'" Meeting of the International Society for Oncodevelopmental Biology and Medicine. Tumor Biol., 11, 59—63. Berling.B., Kolbinger.F., Grunert.F., ThompsonJ.A., Brombacher.F., Buchegger.F., von Kleist.S. and Zimmermann.W. (1990) Cloning of a carcinoembryonic antigen gene family member expressed in leukocytes of chronic myeloid leukaemia patients and bone marrow. Cancer Res., 50, 6534-6539. Chomczynski.P. and Sacchi.N. (1987) Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Analvi Biodiem., 162, 156-159. Efstratiadis.A., PosakonyJ.W., Maniatis.T, Lawn.R.M., O'Connell.C, Spritz.R.A., DeRiel.J.K., Forget,B.G., Weissman.S.M., Shghtom,J.L., Blechl.A.E., Smithies.O., Baralle,F.E., Shoulders.C.C. and Proudfoot.N.J. (1980) The structure and evolution of the human |3-globin gene family. Cell, 21, 653-668. Gehring.W.J. (1987) Homeoboxes in the study of development Science, 236, 1245-1252. Hinoda.Y.M., Neumaier.M., Hefta.S.A., Drzenick.Z., Wagener.C , Shively.L., Hefta.L.J., ShivelyJ.E. and Paxton.R.J (1989) Molecular cloning of a cDNA coding biliary glycoprotein I: Primary structure of a glycoprotein immunologicaUy crossreacting with carcinoembryonic antigen (CEA). Proc Nail. Acad. Sci. USA, 86, 1668. Hynes.R. (1985) Molecular biology of fibronectin. Annu. Rev. Cell Biol., 1, 67-90. Innis.M.A. and Gelfand.D.H. (1990) Optimization of PCRs In Innis.M.A., Gelfand.D.H., SninskyJJ. and White.T.J. (eds), PCR Protocols—A Guide to Methods and Applications. Academic Press, Inc pp. 3 - 1 2 . Kitchin.P.A., Szotyori.Z., Fromholc.C. and Almond,N. (1990) Avoidance of false positive. Nature, 344, 201. Lowe.T., Sharefkin J., Yang.S.Q. and Dieffenbach.C.W. (1990) A computer program for selection of oligonucleotide primers from polymerase chain reactions. Nucleic Acids Res.. 18, 1757-1761. McCleanJ.W., Tomlinson.J.E., Kuang,W.-J.. Eaton.D.L., Chen.E.Y.. Fless.G.M., Scanu.A.M. and Lawn.R.M. (1987) cDNA sequence of human apolipoprotein(a) is homologous to plasminogen Nature, 300, 132-137 Neumaier.M., Zimmermann.W., Shively.L., Hinoda.Y., Riggs.A.D. and ShivelyJ.E. (1988) Characterization of a cDNA clone for the nonspecific crossreacting antigen (NCA) and a comparison of NCA and carcinoembryoruc antigen (CEA). J. Biol. Chem., 263, 3202-3207 Nudel.U., Zakut.R , Shani.M., Neuman.S., Levy.Z. and Yaffe.D (1983) The nucleotide sequence of the rat cytoplasmic beta-actin gene. Nucleic Acids Res.. 11, 1759-1771. Pallansch.L., Beswick.H., Talian.J. and Zelenka.P. (1990) Use of an RNA folding algorithm to choose regions for amplification by the polymerase chain reaction. Analyt. Biochem., 185, 57-62. Ponte.P., Ng,S-Y., Gunning.P. and Kedes.L. (1984) Evolutionary conservation in the untranslated regions of actin mRNAs. DNA sequence of a human betaactin gene. Nucleic Acids Res., 12, 1687-1696. Rooney.B.C, Home.C.H.W. and Hardman.N. (1988) Molecular cloning of a cDNA for human pregnancy-specific |S1 -glycoprotein. Homology with carcinoembryonic antigen and related proteins. Gene, 71, 439-449. Rychlik.W. and Rhoads.R.E. (1989) A computer program for choosing optimal oligonucleotides for filter hybridization, sequencing and in vitro amplification of DNA. Nucleic Acids Res , 17, 8543-8551. Saiki.R.K., Gelfand.D.H., Stoffel.S., Scharf.SJ., Higuchi.R., Hom.G.T., Mullis.K.B. and Erlich.H.A (1988) Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science, 239, 487-491 Sambrook.J., Fritsch.E.F. and Maniatis.T. (1989) Molecular Cloning- A Laboratory Manual (2nd ed). Cold Spring Harbor Press, Cold Spring Harbor. Sleinmetz.M. and Hood.L. (1983) Genes of the major histocompatiWity complex in mouse and man. Science, 222, 727 — 733. Streydio.C, Swillens.S., Georges,M., Szpiper.C. and Wassart.G. (1990) Structure, evolution and chromosomal localization of the human pregnancyspecific /31 glycoprotein gene family. Genomics, 6, 579-592 Suggs,S.V., Hirose.T , Miyake.E.H., Kawashima.M.J , Johnson,K.I. and Wallace.R.B. (1981) In D.D Brown (ed.), ICN-UCLA Symp. Dev. Biol. Using Purified Genes. Academic Press Inc., New York, Vol. 23, pp. 683-693. ThompsonJ. and Zimmermann.W. (1988) The carcinoembryonic antigen gene family: Structure, expression and evolution. Tumor Biol., 9, 63 — 83.

ThompsonJ.A., Mauch,E-M.. Chen,F.-S., Hinoda.Y , Schrewe.H., Berling.B.. Bamert.S , von Kleist.S., ShivelyJ.E. and Zimmermann.W. (1989) Analysis of the size of the carcinoembryonic antigen (CEA) gene family: Isolation and sequencing of N-terminal domain exons. Biochem. Bioph\s. Res. Comm., 158, 996-1004. Tokunaga.K., Taniguchi.H., Yoda.K., Shimizu.M. and Sakiyama.S. (1986) Nucleotide sequence of a full-length cDNA for mouse cytoskeletal |3-actin mRNA. Nucleic Aads Res., 14, 2829-2829. Tonegawa.S. (1983) Somatic generation of antibody diversity. Nature, 302. 575-581. Received on January 14, 1991; accepted on March 14, 1991

Circle No. 12 on Reader Enquiry Card

529