RAMHA(tm): Monte Carlo simulation of the random mutagenesis of synthetic cDNAs David P. Siderovski Department of Medical Biophysics, University of Toronto, Toronto, Ontario M4X 1K9 CANADA INTERNET: Siderovski@Galen.OCI.UToronto.CA (This text is paraphrased from the article Siderovski, D.P. (1992) Computer Applications in the Biosciences [submitted].) Random mutagenesis is a powerful tool in protein structure-function analysis. One approach is the synthesis of cDNA sequences using doped nucleoside phosphoramidites. A Turbo PASCAL program, RAMHA(tm), was written for modelling this method of random cDNA mutagenesis on PC-compatible computers. With RAMHA(tm), one can enter cDNA sequences and various parameters of the mutagenesis (e.g., nucleotide doping level, span of DNA sequence to be mutagenized, position of codons known to be sensitive to mutation). RAMHA(tm) will then simulate the synthesis of a user-specified number of oligos, translate each into polypeptides, and accumulate various statistics, including average percent similarity of mutant proteins to wildtype sequence, frequency of resultant proteins retaining wildtype sequence, and percent of premature open-reading frame terminations. Tests on a cDNA containing all possible codons have revealed the requirement for two distinct "anti-termination" strategies, schemes circumventing the random introduction of stop codons (TAG, TGA, TAA), which have been added to the algorithm. cDNA entry Editing facilities allow the target cDNA to be entered directly as nucleotide sequence or by back-translation of the encoded polypeptide, as entered using one-letter amino acid abbreviations; back-translation is based on pre-defined codon frequency tables supplied for several common research organisms or on user-defined tables as created within RAMHA(tm). Sequences can also be imported from and exported to GCG sequence file format. The nucleotide sequence is displayed along with its translation product; the user can specify the reading frame and one- vs three-letter amino acid abbreviations. In addition to the insertion, deletion, and substitution of bases in the sequence, several features and ranges within the sequence can be defined. The user can specify the first and last nucleotide of the sequence subset destined for mutagenesis; for example, the PCR primer binding sites at the 5'- and 3'-ends of the synthetic oligo are invariant within the resultant cDNA and can therefore be excluded from the mutagenized sequence. Specific codons within the cDNA sequence can be marked as "sensitive to mutation" when prior knowledge exists regarding domains and/or amino acids critical for protein function. In addition, a "truncation flag" can be set on any codon which marks the C-terminal boundary of indispensible primary sequence; this flag is used to compile statistics on premature truncations. Simulation of random mutagenesis After a target cDNA has been entered or imported into RAMHA(tm), the user can directly simulate the random mutagenesis of this sequence using a single set of parameters. Several parameters controlling the mutagenesis simulation can be specified, including nucleotide doping level, number of iterations of the simulation algorithm, and sequence(s) of enzyme site(s) to be detected within "synthesized" oligos. The user can also test a range of nucleotide doping levels using batch execution; an option is provided for writing keyboard "scripts" to automate multiple RAMHA(tm) sessions, using the STUFFIT keyboard control program written by Terje Mathisen and provided with RAMHA(tm). During simulated mutagenesis, RAMHA(tm) will display a scrolling list of translation products generated from "synthesized" oligonucleotides. Upon completion of each round of simulation, a full record of all summary statistics and histograms can be saved to disk, displayed on the screen, and/or printed to an attached ASCII printer. In addition, individual histograms can be output to disk as plain or delimited ASCII; these files can then be directly imported into user-supplied statistical and graphics software packages for graphical output of data resulting from the mutagenesis simulation. As the cloning of the synthetic cDNA pool is dependent on unique restriction enzyme sites encoded by the 5' and 3' PCR primers, any cDNAs with internal copies of these same sites will be fragmented upon restriction enzyme digestion and, hence, rendered unclonable. However, RAMHA(tm) can scan "synthesized" genes for new internal restriction sites arising from nucleotide substitution and report on "hotspots" within the wildtype cDNA sequence. Anti-termination strategies By accumulating data on the number and position of stop codons (TAA, TAG, TGA) arising in "synthesized" oligonucleotides, RAMHA(tm) can report on the frequency of premature open-reading frame terminations and the number of mutant proteins truncated before the end of a critical domain. To examine the contribution of each possible codon to the random appearance of stop codons, simulated random mutagenesis at a 4.5% doping level was performed on an imaginary cDNA possessing every possible coding codon exactly once. Two types of codons contribute exclusively to the appearance of stop codons at this frequency of random substitution (i.e., predominantly single base changes per codon). Codons of the form xAA, xAG, and xGA (where x = [A, C, G]) can easily be converted into stop codons with the substitution of thymine into the first position of the codon (e.g., Glu codons GAA and GAG). In addition, codons of the form Txx (where at least one x= [A, G]) can be converted into stop codons with adenine and/or guanine substitutions into the second and/or third positions of the codon (e.g., Cys codons TGC and TGT). One strategy to avoid random appearance of nonsense mutations is simply to substitute alternative codons without the potential to form stop codons (e.g., Gly codon GGA to GGC, GGG, or GGT; Leu codons TTA or TTG to CTn). However, several amino acids can only be specified with codons exhibiting "stop-codon potential" (Cys, Glu, Lys, Gln, Trp, Tyr). Therefore, two additional sets of doped nucleotide precursors ("anti-termination mixes") were designed to circumvent the random placement of nucleotides producing stop codons. For codons of the form xAA, xAG, and xGA (where x = [A, C, G]), the nucleotide mixture used for the first codon position is not doped with thymidine phosphoramidite, such that the levels of the other two contaminants are increased to maintain the total level of doping. This anti-termination strategy therefore entails a set of three additional nucleotide mixes, one each for adenine, cytosine, and guanine in the first position of xAA, xAG, and xGA codons. For codons of the form Txz, where (x = [A, G] and z = [C, T]) or (x = [C, T] and z = [A, G]), the position encoded by a pyrimidine (C or T) in the wild-type sequence is synthesized using a doped mixture lacking purine nucleotides (A and G); for example, the third position of the tyrosine codon TAC would be synthesized using a cytidine phosphoramidite stock doped solely with thymidine phosphoramidite. Of the four remaining Txz codons, where both x and z = [A, G], three are stop codons (TAA, TAG, TGA), whereas only one encodes an amino acid (Trp codon TGG). In this case, the nucleotide mixture used for both the second and third codon positions is not doped with adenosine phosphoramidite. Hence, to avoid the random generation of nonsense mutations at Txz codons, a second set of three "anti-termination" nucleotide mixes is required. The basic algorithm of RAMHA(tm) therefore accomodates the use of either one or both anti-termination strategies during simulated random mutagenesis. To aid the user in identifying those codons which would benefit from either anti-termination strategy, RAMHA(tm)'s cDNA editor automatically displays a single-letter code immediately underneath nucleotides within such codons and tags with carats (^) those codons which have alternates without "stop-codon potential"; a help window describing the anti-termination mix associated with each single-letter code is available during cDNA entry.