| Nucleic Acids Research | Pages |
Tolerance of 5-fluorodeoxyuridine resistant human thymidylate synthases to alterations in active site residues
Introduction
Materials And Methods
Cell lines and materials
Plasmids
Construction of the TS random library
Replacement of wild-type TS sequence with the random library
Genetic selection in E.coli
Crude extracts of wild-type and mutant TS
Determination of TS protein concentration
TS enzyme assays
Computational analysis
Results
Construction of the random sequence library
Sequencing of non-selected clones to determine diversity of library
Selection and sequencing of E.coli expressing active mutant TS
Selection and sequencing of E.coli expressing TS resistant to 5-fluorodeoxyuridine
Discussion
Substitutions that retain activity differed in many respects compared to the non-selected random library
The 5-FdUR selected library is further restricted
Catalytic activity of mutant thymidylate synthases
Conclusions
Acknowledgements
References
Tolerance of 5-fluorodeoxyuridine resistant human thymidylate synthases to alterations in active site residues
Received May 26, 1999; Revised and Accepted July 28, 1999
ABSTRACT Fluoropyrimidines, such as 5-fluorouracil (5-FU), are used extensively in cancer therapy. In the cell, 5-FU is metabolized to 5-fluorodeoxyuridylate (5-FdUMP), a tight binding covalent inhibitor of thymidylate synthase (TS). In order to create 5-FdUMP resistant enzymes to protect chemosensitive normal cells and further understand mechanisms of 5-FdUMP resistance, we have randomized four residues within the active site of TS. Our previous studies identified alterations in residues which produce active TS with enhanced resistance to 5-fluorouridine (5-FdUR). By remutagenizing a subset of the 13 previously targeted residues (A197, L198, C199 and V204), an unbiased random library can be created allowing for extensive testing of all possible amino acid substitutions at each of the sites. Using genetic complementation and selection in Escherichia coli, we identified the spectrum of substitutions that yield active TS as well as those that resulted in 5-FdUR resistant mutants of TS. The 5-FdUR resistant TS were found to share several structural features including hydrophobic substitutions at residue 197, retention of the wild-type leucine 198, the alteration C199L (present in 64% of the drug-resistant library), and polar alterations of valine 204. The catalytic activity of mutants with these features was approximately equal to that of the wild-type TS.
INTRODUCTION
Fluoropyrimidine based analogs, such as 5-fluorouracil (5-FU) and 5-fluorodeoxyuridine (FdUR) are used extensivelyfor treatment of colonic, breast and ovarian carcinomas. Intracellularly, the fluoropyrimidines are metabolized to 5-fluorodeoxyuridylate (5-FdUMP), which forms a stable inhibitory complex with thymidylate synthase (TS, EC 2.1.1.45) and the co-substrate CH2H4-folate. Inactivation of TS results in decreased production ofTMP, cessation of DNA synthesis and ultimately thymineless death (1). Additionally, cytotoxic effects may be mediatedby incorporation of the metabolic products FdUTP and FUTP into DNA and RNA, respectively. However, it is generally accepted that the primary cytotoxic mechanism of systemic 5-FU therapy is the covalent inhibition of TS. As such, TS presents an attractive target for drug design, and many novel inhibitors of this enzyme have already become promising chemotherapeutics (2-5).
For many years, the only known drug-resistant human TS enzyme was the mutant Y33H. This mutant was discovered to be heterozygously expressed in a HCT116 cell line that was resistant to the cytotoxic effects of 5-FdUR (6-8). The serendipitous discovery of this enzyme was followed by detailed kinetic studies relating the structural changes with the altered functions (9,10). However, the last 2 years have seen an explosion of new studies identifying drug-resistant TS thatcontain amino acid alterations in different regions. Treatment of human HT1080 cells with ethyl-methanesulfonate followed by selection with the antifolate Thymitaq (AG337) resulted in the discovery of two novelsubstitutions, D49G and G52S, that confer resistance to both FdUR and Thymitaq (11). These residues are located in the highly conserved Arg50 loop involved in dUMP binding. By subjecting highly conserved residues important in cofactor binding to site directed mutagenesis, I108A was found to be resistant to the antifolates Raltitrexed (Tomudex, ZD1694) and Thymitaq, and F225W was found to be resistant to BW1843U89 and 5-FdUR (12). Recently, a 5-FdUR resistant cell line was determined to encode a P303L mutant, which although metabolically unstable was nonetheless able to confer resistance to transfected cells against FdUR, Raltitrexed, Thymitaq and BW1843U89. An engineered P303D mutant was resistant to both 5-FdUR and BW1843U89 (T.Spencer and F.Berger, personal communications).
Wehave previously identified several 5-FdUR resistant TS by using a library that contained 14% random substitutions at13 residues near the active site (13). These residues, A197-C199 and V204-L212, are near to the active site cysteine (195) responsible for nucleophilic attack at the C-6 position of dUMP. Because these residues are not absolutely conserved and are adjacent to, but not in contact with, the active site cavity (14), we hypothesized that substitutions at one or several of these residues could allow subtle changes that maintain TS activity and yet promote specific resistance to 5-FdUMP. The most frequent change seen in the library was C199L. A triple mutant A197V; L198I; C199F exhibitednear wild-type catalytic activity with Km values for dUMP and CH2H4-folate, but demonstrated a Kd for dUMP 20-fold higher than that of the wild-type enzyme presumably accounting for the 5-FdUR resistance. Due to the number of random substitutions and nucleotide bias of the library, all 20 possible residues may not have been tested at each position. In order to analyze the frequency of substitutions at different positions within the active site of TS we assembled a library with 100% random substitutions at key positions and assessed the effects of different substitutions on the production of active and 5-FdUR resistant TS.
Since 5-FdUMP is structurally similar to the natural substrate dUMP, it is difficult to predict how single amino acid substitutions or multiple substitutions could restrict the binding of 5-FdUMP without affecting binding of dUMP. Random oligonucleotide mutagenesis provides a combinatorial alternative to creating altered enzymes without requiring detailed knowledge about amino acid interactions or effects of specific alterations. In the present study, we have generated a large library of substitutions at a small number of residues. The powerful positive genetic complementation allows facile collection of catalytically active mutants with amino acid replacements in the active site of TS. These active mutants were subsequently selected for their ability to confer growth of Escherichia coli in the presence of 5-FdUR. Several mutants were identified that demonstrate both 5-FdUR resistance and near wild-type catalytic activity.
MATERIALS AND METHODS
Cell lines and materials
(6R,S)-N5-N10-Methylenetetrahydrofolate (CH2H4-folate) was obtained as a racemic mixture from Schircks Labs (Jona, Switzerland). [6-3H]-5-FdUMP was supplied by Moravek Biochemicals (Brea, CA). ABI Prism Dye Terminator Cycle Sequencing kits for fluorescent sequencing were the products of Perkin Elmer (Branchburg, NJ). Escherichia coli DNA Pol I was from New England Biolabs (Beverly, MA). Pfu DNA Polymerase was from Stratagene (La Jolla, CA). Plasmid DNA was isolated using the Maxiprep and Miniprep kits from Qiagen (Chatsworth, CA), or the PERFECTprep Plasmid DNA kit from 5-Prime 3-Prime Inc. (Boulder, CO). 5-FdUR, 5-FdUMP, dUMP, TES and all other reagents were from Sigma. Escherichia coli NM522 (Stratagene, La Jolla, CA) was used for cloning and library construction. Escherichia coli [chi]2913recA ([Delta]thyA572, recA56), kindly provided by Dr Daniel Santi (UCSF), was used inall complementation studies and in the purification of plasmid encoded TS. Unless otherwise stated, all DNA oligomers were from Operon Technologies (Alameda, CA).
Plasmids
The construction of a TS stuffer vector, containing a 1.3-kb DNA fragment insert between coding nucleotides 555 (BglII site) and 648 (MroIsite), has been described previously (13). Briefly, the wild-type human TS expression vector pGCHTS-TAA was modified by creation of an MroI restriction endonuclease site at coding nucleotide 648. This coding region was replaced by a 1.3-kb stuffer DNA fragment derived from a modified pET3a vector (Novagen, Madison, WI).
Construction of the TS random library
The TS random library was constructed by annealing two single-stranded DNA oligodeoxyribonucleotides (Fig. 1, step 1). `Oligo 1', d(ATCATCATGTGCGCTTGGAATCCAAGAGA-
TCTTCCTCTGATGGCGCTGCCTCCATGC), is a 57mer that corresponds to the sense nucleotides 529-585 and contains a BglII site (nucleotide 555) for cloning. Oligodeoxyribonucleotide `Random-100' is a 96merthat containsrandom nucleotide substitutions at amino acid positions 198-200 and 205. Its sequence is d(GCCCATGTCTCCGGATCTCTGGTACAGCTG-
GTACAGCTGGCAGGACAGCTCACTGTTNNNTACGTA-
GAACTGNNNNNNNNNATGGCATGGAGGCAGCGC CATCAG), where N denotes an equal proportion (25%) of each deoxyribonucleotide at each of the 12 randomized positions, and an MroI site for cloning. Two silent mutations (double underline) were introduced in the non-random region of this oligodeoxyribonucleotide to create a unique SnaBI site for subsequent identification based on digestion with the restriction enzyme. This oligodeoxyribonucleotide was synthesizedby Integrated DNA Technologies (Coralville, IA). Oligo 1 and Random-100 were annealed in 50 µl of 200 mM Tris-HCl, pH 7.5, 100 mM MgCl2, 250 mM NaCl by incubation at 80°C for 5 min, followed by 55°C for 15 min, at 37°C for 15 min, and at room temperature for 15 min (Fig. 1, step 1). The partial duplex was extended in a 40 µl reaction mixture containing 10 mM Tris-HCl, pH 7.5, 5 mM MgCl2, 7.5 mM dithiothreitol, 250 µM dNTPs, and 5 U Klenow fragment of E.coli DNA Pol I for 2 h at 37°C (Fig. 1, step 2). The double-stranded oligodeoxyribonucleotides were then amplified in a 100 µl polymerase chain reaction containing 20 mM Tris-HCl, pH 8.75, 10 mM KCl, 10 mM (NH4)2SO4, 2 mM MgSO4, 0.1% Triton X-100, 100 µg/ml BSA, 1 µM primers DLTS2F and DLTS3Rand 2.5 U Pfu DNA polymerase. The reaction mixture was heated for 30 cycles in a programmable thermal controller (MJ Research, Watertown, MA) at 95°C for 1 min, 50°C for 1 min and 72°C for 1 min (Fig. 1, step 3). DLTS2F is d(ATCATCATGTGCGCTTGG), and corresponds identically tothe 5[prime] end of Oligo 1. DLTS3R, d(AAAAAAAAACCATGTCTCCGGATCTCTGGTAC), is a 32mer corresponding to antisense nucleotides 636-658 and is identical to the 3[prime]-end of Random-100 with the addition of a 9-bp oligo-A at the 5[prime] terminus to facilitate subsequent restriction digestion. The amplified DNA was purified using the Qiagen Qiaquick PCR Purification Kit and digested with BglII (New England Biolabs) and MroI (an isoschizomer of BspEI; Boehringer Mannheim, Indianapolis, IN) (Fig. 1, step 4), purified by phenol extraction and ethanol precipitated.
Figure 1. Schematic of random sequence mutagenesis of human TS. Construction of the plasmid-encoded library of TS containing random nucleotide substitutions is shown corresponding to the residues 197-199 and 204. Synthesis of the random nucleotide-containing oligonucleotides (inserts) is illustrated in steps 1-4. (X)12 in oligomer `Random-100' denotes the 12 non-contiguous nucleotides containing a 25% mixture of each four nucleotides. Ligation of the inserts into the dummy vector, to replace the wild-type TS sequence at residues 197-199 and 204, is shown in step 5. Preparation of the plasmid library in TS- E.coli is outlined in steps 6-8.
Replacement of wild-type TS sequence with the random library
The purified, partially random oligodeoxyribonucleotides were used as inserts for construction of the human TS plasmid library. The stuffer vector was removed by digestion with BglII and MroI and the resulting 3.6-kb fragment was isolated from a 0.8% agarose gel and was ligated to the 95-bp restricted random insert using T4 DNA ligase (Gibco BRL) (Fig. 1, step 5). The ligation mixture was directly transformed(Bio-Rad Genepulser, 2 kV, 25 µFD, 400 Ohm) into fresh electrocompetent NM522 cells (Stratagene) (Fig. 1, step 6), in 10 separate transformations using 2 µl of the ligation mixture and 100 µl NM522 E.coli. After combining the transformation reactions, the size of the library containing the TS plasmid was determined by plating an aliquot of the transformation mixture on media containing carbenicillin (50 µg/ml; Island Scientific, Bainbridge Island, WA). The remainder of the library was amplified by growing the transformed NM522 cells overnight in 1× YT media in the presence of carbenicillin and the plasmid was harvested (Fig. 1, step 7). Electrocompetent [chi]2913 (TS-) cells were transformed with the plasmid library and pooled, plated to confirm adequate transformation efficiency, grown overnight in appropriate antibiotics and 50 µg/ml thymidine, and stored in aliquots at -80°C in 10% glycerol (Fig. 1, step 8). The extent of randomization was verified by sequencing plasmid DNA isolated from 22clones of NM522 E.coli cells containing the random library. All sequencing reactions were conducted using the ABI Prism Dye Terminator Cycle Sequencing Kit with AmpliTaq DNA polymerase, and the sequencing primer `3[prime]HSG', d(TAACGCCAGGGTTTTCCCAG). The primer `3[prime]HSG' is a 20mer which corresponds to the antisense sequence of the TS plasmid 33 bp downstream of the SacI site.
Genetic selection in E.coli
Clones encoding active TS protein were selected by growth of [chi]2913 E.coli cells containing the random library on M9 minimal medium containing 50 µg/ml carbenicillin and 10 µg/ml tetracycline. This medium, lacking thymidine, allows colony formation of [chi]2913 (TS-) E.coli cells only if complemented by a functional TS (15). Thawed [chi]2913 cells containing the random library were inoculated (1:100) into 1× YT medium containing 50 µg/ml thymidine and the appropriate antibiotics and grown at 37°C until the absorbance at 600 nm attained a value of 0.8-1.0. Aliquots of 1 ml of the exponentially growing cells were pelleted and resuspended in M9 salts, plated on minimal medium containing carbenicillin and tetracycline, and incubated at 37°C for 36 h. Plasmids were isolated from 32 surviving colonies and the corresponding random region was sequenced to determine amino acid changes tolerable in maintaining catalytic activity of TS.
In order to select for library members that are resistant to killing with 5-FdUR, transfectants were plated on minimal medium containingincreasing amounts of 5-FdUR (0-200 nM 5-FdUR) and incubated at 37°C for 36-48 h. Colonies formed in 5-FdUR were isolated and theplasmid from each clone was retransformed into fresh [chi]2913 E.coli in order to confirm the drug-resistant phenotype. Each retransformed bacterium was then subjected to the same selection procedure. Plasmid DNA from cells which survived 175 or 200 nM FdUR (n = 53), which is lethal to E.coli harboring the wild-type TS, was sequenced.
Crude extracts of wild-type and mutant TS
From a fresh overnight culture, [chi]2913 cells containing wild-type or mutant forms of TS were grown in 30 ml 2× YT medium containing carbenicillin. After attaining an absorbance at 600 nm of 0.4, TS expression was induced by addition of 1 mM IPTG. After 2.5 h, cells were harvested by centrifugation, resuspended in 50 mM Tris-HCl, pH 7.5, 1 mM EDTA, 200 mM NaCl, 10% glycerol and 200 µg/ml lysozyme, aliquoted and stored at -80°C. In all subsequent experiments, frozen cells were thawed and lysed on ice for ~3 h. The lysed cells were centrifuged (27 000 g) and the pellet discarded.
Determination of TS protein concentration
The concentration of active TS dimer in the crude extracts was determined by [3H]FdUMP binding. Crude extract (25 and 50 µl) was incubated with 150 µM CH2H4-folate and 300 nM [3H]-FdUMP in a standard reaction (500 µl) containing 50 nM TES, pH 7.4, 25 mM MgCl2, 6.5 mM formaldehyde, 1 mM EDTA, and 150 µM 2-mercaptoethanol. The concentration of all reagents used was determined by weight with the exception of [3H]FdUMP. The specific activity of this compound was 18.6 Ci/mmol. Following incubation for 1 h at room temperature, 125 µl 50% TCA was added and the mixture was centrifuged at 13 000 g for 5 min. The pellet was washed four times with 10% trichloroacetic acid and resuspended in a mixture of 2 M NaOH/50% ethanol, and added to 5 ml Scinti-Verse scintillation fluid (Fischer) and the radioactivity quantified. For calculations, we assumed 1.7 mol of FdUMP bind 1 mol of TS (13). The assay was repeated using one-half of the amount of crude extract to ensure [3H]FdUMP saturation. Reactions were conducted in duplicate. No counts above background were detected in a control reaction incubated in the absence of cell extract. Analysis of purified wild-type TS was conducted as a control, and results were consistent with those obtained by Bradford analysis.
TS enzyme assays
TS activity was monitored spectrophotometrically by the increase in absorbance at 340 nm that occurs concomitant with the production of H2-folate ([Delta][epsiv] = 6400 M-1 cm-1) (16,17). To the standard reaction buffer was added 300 µM of the racemic mixture (150 µM of the active isomer) of CH2H4-folate and 200 µM dUMP. The crude extract (60, 80 and 100 µl) was added to a total reaction volume of 1 ml to initiate the reaction. Reactions were conducted at 25°C in a Peltier temperature controlled microcuvette. Initial velocitiy measurements were obtained on a Perkin Elmer Lamba Bio 20 UV/Vis spectrophotometer and slopes determined using the software Cricket Graph III (Computer Associates, NY). Catalytic activity was determined by dividing the Vmax by the concentration of TS determined above.
Computational analysis
For purposes of analyzing the structural consequences of mutagenesis of thymidylate synthase, two coordinate sets were used. Inasmuch as the PDB file 1tls.pdb was no longer available through the Protein Data Bank, coordinates were released to us directly from the authors (14). This structure does not have substrate or inhibitor bound, and some portions of the structure are disordered. Moreover, the loop from residue 181 to 196 has a substantially different conformation than its counterpart in the E.coli TS. The E.coli structure does include 5-FdUMP and the folate analog 10-propargyl-5,8-dideazafolate (18). We constructed a `chimera' to represent the human TS with 5-FdUMP bound by using the E.coli structure in the region of residues 181-196 (human TS numbering), substituting in the human sequence and performing gentle energy minimization. This also resulted in an artificially short loop connecting the missing residues 114-133 (human numbering). The E.coli and human TS structures were superimposed and the bound 5-FdUMP and folate coordinates were included in the figures but not in the calculations.
In order to evaluate the effect of the mutations, specific residues were mutated in the bound or unbound form of the human TS structure, gently energy minimized using X-PLOR (v.3.1), raising the simulated annealing temperature to only 500°C and cooling to 300°C. Energy minimized mutated and unmutated structures were compared on a computer graphics system using `O' (19). Coordinates for the E.coli TS (1tls.pdb) were obtained from the Protein Data Bank (20). Figure 5 was constructed using MOLSCRIPT (21) and Raster3D v.2.0 (22).
RESULTS
Construction of the random sequence library
A library of 2 × 105 human TS mutants was created by random sequence mutagenesis. This library contains 25% of each base at codons corresponding to residues 197-199 and 204. Unlike the wild-type TS construct, the recombinant plasmids also containa silent SnaBI site that can be used to rapidly confirm that activeclones are the result of selection from the library and not wild-type DNA plasmid contamination. These positions were targeted based on previous studies that demonstrated that certain substitutions conferred drug resistance.
Sequencing of non-selected clones to determine diversity of library
Prior to selection, plasmid DNA was isolated from 22 transformedclones and sequenced. The number of substitutions per clone is presented in Figure 2A, and the types of substitutions are tabulated in Figure 3A. An average of 10 nucleotide changes and 3.8 amino acid changes per clone were detected. Amino acid substitutions in the non-selected cloneswere evenly distributed among the 22 residues encoded by the randomized nucleotides (Fig. 3A). Nine of the 22 non-selected clones analyzed (40%) contained frameshift mutations (three insertions, six deletions) and two contained termination codons (Table 1). Thus approximately half of the TS random library clones will not produce full-length protein. Based on the number and frequency of random substitutions, we calculated thatthe probability of obtaining a wild-type nucleotide sequence in the non-selected library is ~6.0 × 10-8. The incidence of a wild-type protein, although higher due to the degeneracy of the nucleotide code, would still be a meagre 0.001% of the library. Therefore it is expected that no wild-type molecules would be detected amongst the 22 non-selected clones sequenced nor present in the library.
Figure 2. Number of amino acid substitutions. (A) Non-selected mutants; (B) active mutant library; (C) 5-FdUR-selected mutant library. Twenty-two clones were sequenced from the non-selected library. Forty-two clones were sequenced from the pool of mutants that demonstrated the ability to complement the TS- phenotype of [chi]2913 E.coli (comprising the active mutant library). Fifty-three clones were sequenced from the 5-FdUR-selected library.
A
 |
B
 |
C
|
Figure 3. Amino acid substitutions. (A) Non-selected; (B) active mutant library; (C) 5-FdUR-selected library. The wild-type TS sequence is shown below the solid line, and the substitutions observed at each position are indicated above each site. Boxed residues are those subjected to mutagenesis. The catalytic cysteine, Cys-195, is absolutely conserved (underlined). Green indicates charged residues, blue indicates polar resides and red indicates hydrophobic changes.
Table 1. Library statistics
| Library | Number sequenced | Avg. # AA [Delta][prime]s | Avg. # NT [Delta]'s | # Frameshift mutations |
| Non-selected | 22 | 3.8 | 10 | 9 |
| Functional | 32 | 3.5 | 9 | 0 |
| 5-FdUR resistant | 52 | 2.8 | 8.5 | 0 |
Selection and sequencing of E.coli expressing active mutant TS
We isolate active TS enzymes from the large plasmid libraries using a positive genetic complementation (13). The wild-type TS construct is able to rescue the TS- E.coli phenotype and form colonies on minimal medium (containing no thymidine), yetE.coli harboring the stuffer vector that inactivates TS does not. By plating the random library on media with or without thymidine, we determined that ~22% of the non-selected random library encodes for functional enzymes and thus our active, or functional, library contains ~2.2 × 104 clones. DNA from 32 mutants was sequenced, without detecting any nonsense or frameshift mutations. The average number of nucleotide and amino acid substitutions was 9.0 and 3.5, respectively. No clone was identical at the nucleotide level (Table 1).
Selection and sequencing of E.coli expressing TS resistant to 5-fluorodeoxyuridine
To isolate members of the random enzyme library demonstrating selective resistance to 5-FdUMP, an additional positive genetic selection was employed by supplementation of the minimal medium agar plates with a gradient of 5-FdUR concentrations. Previously, we have reported thatthe survival of the E.coli harboring wild-type TS is only modestly reduced to 90% at 75 nM 5-FdUR, however it precipitously declines to 0.1% at 100 nM of the analog (13). Consistent with these results, no surviving wild-type clones were detected at dosages above 100 nM. Unlike the wild-type, ~0.1% of the E.coli cells transformed withthe random library formed colonies at 5-FdUR concentrations as high as 200 nM. It has been shown that survival in E.coli at dosages above where wild-type is viable correlates with kinetic resistance in vitro (13). As an additional control, the two well-described mutants Y33H and G52S have been constructed via site directed mutagenesis and tested in this 5-FdUR gradient plating assay and survive at dosages clearly lethal to the wild-type enzyme (D.Landis, unpublished).
Individual colonies were isolated from plates containing 175 or 200 nM 5-FdUR and plasmid DNA was re-transformed into fresh [chi]2913 to confirm the drug-resistant phenotype. DNA from those mutants that formed colonies upon a second exposure to 5-FdUR at a dose which was clearly lethal to the wild-type enzyme (175 or 200 nM, n = 53) was sequenced in the random region and the corresponding amino acid changes deduced (Fig. 3C).
Among the 5-FdUR survivors, many clones were found to encode the identical DNA sequence, indicating that the resistant pool was dominated by a few sequences. In fact, the 53 clones which were each retransformed, reselected and sequenced consisted of only 14 unique mutants (Fig. 3C). Thus, the diversity of our library at the drug selected level was limited, consistent with the observation that the survival of the library at 175 nM 5-FdUR was approximately one-thousandth of that of the untreated library. The average number of nucleotide and amino acid substitutions in the 5-FdUR resistant pool was 8.5 and 2.8, respectively (Table 1).
The most common change seen in the drug-resistant library was C199L, observed in 9 of the 14 unique (65%) clones (Fig. 3C).This was enriched from a 12% presence in the functional TS library (Fig. 3B). Three of the six possible codons encoding leucine were observed in the selected clones. Other selected alterations at this site included the hydrophobic residues isoleucine, methionine and phenylalanine. Although polar residues such as serine and threonine made up nearly half of the functional library at residue 199 (Fig. 3B), no polar residues were found in the drug selected library at this codon (Fig. 3C). Leucine 198 was the least mutable, being reselected as leucine in 10 of the 14 clones, a trend similar to that seen in the active library. As with C199L, three of the six possible codons encoding leucine were observed in the selected clones. L198 is buried in the dimer interface in a tightly constrained pocket and has direct contacts with L198 from the other monomer. The few substitutions that were tolerated at this site (isoleucine, threonine, phenylalanine) were detected in both the functional and resistant libraries, indicating that these alterations are not specifically associated with drug resistance. Although alanine 197 could be substituted by either hydrophobic or polar residues in the functional TS library (Fig. 3B), only non-aromatic hydrophobicresidues were selected amongst the drug-resistant enzymes (Fig. 3C). Finally, virtually any change in V204 allowed for an active mutant in the functional library (16 of the possible 20 amino acids were observed in Fig. 3B), spanning a wide range of size, charge and hydrophobicity. However, in the drug-resistant population, all mutations at V204 were restricted to polar or charged residues (Fig. 3C).
In order to assess the level of activity in these mutant forms of TS compared to the wild-type, crude extracts were prepared from seven of the mutants selected in the 5-FdUR library. Activity was determined in a spectrophotometric assay that measures the formation of H2 folate, which is produced stoichiometrically with dTMP. Values were the average of three determinations. The catalytic activites of most of the mutant forms were similar to the wild-type enzyme (Table 2.). Only two of the mutants demonstrated catalytic activites less than half of that of the wild-type (M32: A197V; L198F; C199L; V204S and M43: A197M; L198I; C199L; V204T). Interestingly, these mutants are unique in that they do not contain the L197-L198-L199 motif; mutants containing the three contiguous leucines demonstrated activity equivalent to the wild-type construct. The double mutant A197M; C199L, over-represented in the drug-resistant library, demonstrated approximately the same activity as well.
Table 2. Catalytic activities of key mutants
| Clone no. | Sequence (residue 197-199, 204) | Catalytic activity (s-1) |
| M23 | M L L - V | 1.4 0.3 |
| M21 | L L L - Y | 0.8 0.2 |
| M28 | L L L - N | 1.2 0.1 |
| M29 | L L L - Q | 1.6 0.1 |
| M33 | L L L - R | 1.3 0.2 |
| M32 | V F L - S | 0.32 0.03 |
| M43 | M I L - T | 0.62 0.01 |
| Wild-type | A L C - V | 1.3 0.1 |
In general, the rank order of survival of TS resistant mutants in this E.coli plating assay correlates qualitatively to survival in mammalian cells or kinetic inhibition values. The survival of the previously described mutant M64 against FdUMP (A197V; L198I; C199F) (Kd ~20 × WT) was determined to be superior to the mutant Y33H (Kd ~4 × WT) but inferior to G52S (Ki ~20 × WT and IC50 ~97 × WT) (D.M.Landis, unpublished).
DISCUSSION
In a previous study, we used random sequence mutagenesis to alter the active site of the human TS (13). A large library of TS substitutions was generated spanningresidues 197-199 and 204-212. Based on the crystal structure of the E.coli and human TS, these residues form two strands of the [beta]-sheet that forms the dimer interface and at one end contacts the substrate or 5-FdUMP binding site. Therefore, amino acid substitutions can cause changes in the binding of pyrimidine based analogs. In order to target 13 residues in the previous study, it was necessary to bias the random library toward the wild-type nucleotides to ensure that clones would contain few enough substitutions to retain activity (Fig. 4A). This library was designed with 14% randomness to yield an average of ~4.2 amino acid changes per clone, and 1% of the library should encode wild-type protein. Although this method is a powerful technique to identify important residues involved in 5-FdUR resistance, due to inherent statistical limitations, it cannot allow for the assessment of complete combinatorial sets of multiple amino acid changes in a library. Also, because certain sets of nucleotide substitutions will occur at a lower probability than others, it is difficult to know if all 20 amino acids were tested at each site. To circumvent these limitations, we have re-randomized a subset of the previous targeted region and removed the bias toward wild-type nucleotides. Subjecting the entire 13 amino acid segment to unbiased (100%) randomization is impractical because the large number of nucleotide changes (Fig. 4B) would result in a majority of non-functional and truncated proteins. Additionally, due to technical limitations in creation of a random library, only a fraction of the possible permutations of mutants could be analyzed. For this reason, it is necessary to limit the size of the region of interest to a maximum of four residues (Fig. 4A and C). In this 100% randomization, the only bias present will be that due to a variable number of codons for each residue in the genetic code.
Figure 4. Theoretical distribution of nucleotide changes among randomized residues. (A) The original library, encoding 13 residues, was randomized at 14% in order to yield an average of 5.4 nt changes, corresponding to 3.8 amino acid changes. Wild-type clones will be created at ~1% amino acid level. (B) Although using an unbiased library (100% random) is preferable to create multiple substitutions and equally test residues, it is not feasible because the average clone will contain 29 nucleotide changes, or 12 amino acid alterations, likely leading to a non-functional enzyme. (C) In order to regain an appropriate level of nucleotide changes in an unbiased random oligomer library, the number of residues targeted must be reduced to four. This will yield an average of 5.4 nucleotide mutations corresponding to 3.8 amino acid changes. The probability of obtaining a wild-type protein in this library is essentially zero (0.001% at the protein level).
Figure 5. Structural analysis of human TS. (A) Location of the sites for 5-FdUR resistance is plotted on the backbone of the `chimera' of the human thymidylate synthase (see text for `chimera' construction). Gray ribbon is monomer A, colored ribbon is monomer B. Magenta ball and stick is 5-FdUMP and the folate analog from the E.coli TS structure (1tls.pdb). (B) Analysis of the human TS mutant M29 (A197L; C199L; V204Q), in the `bound' form. Unmutated sites are colored by atomic coloring, mutated sites are cyan. The locations of residues that change systematically after energy minimization are shown in red. The most important consistent effect is a cumulative shift of residues towards the 5-FdUMP binding pocket so that Cys 195 S[gamma] may no longer be favorably disposed to bind to fluorodeoxyuridylate.
Substitutions that retain activity differed in many respects compared to the non-selected random library
The region selected for mutagenesis is highly plastic and accepts many amino acid alterations. It was previously observed that most of the four targeted residues could tolerate many substitutions. For this reason, it is not surprising that despite an average number of amino acid substitutions in the unselected library of 3.8 out of the four residues, 22% of these clones were active as demonstrated by survival on media without thymidine. In addition, although nucleotide sequence alignments indicate thatTS is phylogenetically one of the most highly conserved enzymes known, many authors have noted that even highly conserved residues are extremely tolerant to a variety of substitutions in vitro when they are not directly involved in the reaction mechanism (5).
The mutants which maintained the ability to complement the TS- E.coli strain [chi]2913 (the functional library) consisted of a far more limited number of amino acid substitutions at each position (Fig. 3A and B). Charged residues were eliminated at residues 198 and 199, and all aromatic structures were largely excluded in position 197. Only V204 retained a similar spectrum of mutations. It is able to accommodate almost any mutation. All frameshift and nonsense mutations were eliminated from the functional library.
The 5-FdUR selected library is further restricted
Unlike the functional TS library, in which each of the residues was indiscriminate in selection between polar and hydrophobic alterations (and charged residues in the case of V204), the spectrum of allowable mutations in the 5-FdUR selected library is more limited. Residues A197, L198 and C199 were completely reselected only to hydrophobic substitutions, with the single exception L198T. Interestingly, whereas V204 was substituted for 16 of the 20 amino acids in the functional library, mutations were only to polar residues and arginine in the 5-FdUR library. The rationale for including V204 in this unbiased library was based on the observation that a much higher percentage of mutations in V204 was present in the previous drug selected 14% library than would be statistically expected. Although these mutations included many types, because the nucleotide composition was biased such that 86% of the wild-type nucleotide was present at each randomized position, mostly single nucleotide alterations dominated the observed substitutions. These substitutions were predominately Met, Leu, Gly and Ala-all hydrophobic residues. By using an unbiased library, we can allow an unbiased test of each of the 20 amino acids at this position, and can justly conclude that changes associated with 5-FdUR resistance at V204 include only polar and charged groups, although one specific substitution does not appear to lead to significant advantages over another. This residue is at the end of a [beta]-strand and faces one side of the Arg50 loop of the other monomer, with much space to accommodate various side chains. It is of interest that ethyl methanesulfonate mutagenesis has identified that alterations in the Arg50 loop lead to 5-FdUR resistant mutants (11). Many species contain an insertion of several residues (ranging from 1 to 17) following the residue analogous to 204 before returning to the conserved sequence.
In our prior study, the substitution C199L was found in 46% of the drug-resistant clones, whereas it was not detected in the sampling of the functional clones (13). Because of the bias of the 14% library, we remained uncertain if another amino acid alteration at this position may have been superior to leucine, but this would have required multiple nucleotide changes and was not tested. The present unbiased study demonstrates that leucine is indeed the most favorable change at residue 199, and is found in 65% of the resistant clones. In neither the drug-resistant nor active pool was C199 reselected to remain wild-type. C199 is present in human and mouse TS, as well as organisms including Candida albicans, Plasmodium falciparum and Toxoplasma gondii (5) The corresponding residue is phenylalanine in E.coli, tyrosine in Lactobacillus casei, and serine in Saccharomyces cerevisiae. In most species of TS, these larger side chains are usually accompanied by a smaller residue at methionine 179 that packs against 199 (valine in E.coli and L.casei) (14,18). Thus it is apparent that compensatory changes are observed between the homologs of C199 and M179 such that smaller residues can accommodate the larger changes observed in C199. However, in our drug-resistant mutants, C199 packs against the relatively large methionine, forming a rigid hydrophobic pocket. Interestingly, leucine at position 199 has so far not been observed in any of the 29 sequenced species of TS. This is of interest because other combinatorial methods to create novel TS molecules, such as recombination of segments between species (family shuffling) (23,24), would have missed this advantageous mutant.
Catalytic activity of mutant thymidylate synthases
Analysis of seven mutants that dominated the drug selection has demonstrated that all were highly active enzymes (Table 2). This is not unexpected because they have been selected via complementation and would likely retain enough catalytic activity to complement a human cell. All mutants with the L197; L198; L199 motif demonstrated wild-type activity. The two mutants demonstrating less activity than wild-type harbored substitutions at L198. This, coupled with the fact that substitutions at L198 were disfavored in the drug-resistant library compared to reselection for leucine, indicates that mutation of this residue is deleterious not only to activity but also does not aid in drug resistance. Among the substitutions observed at the highly mutable V204 in the active library, none resulted in a dramatic decrease in activity. In this study, the wild-type catalytic activity was determined to be 1.3 s-1 in crude extracts. In order to confirm this value, the wild-type enyzme was purified, as previously described (13), and catalytic activity was determined to be 1.2 s-1. These values are considerably lower than those obtained in our previous experiment (13) and accordingly we revise our previous estimate of 3.9 s-1 to 1.2 s-1. This value is in accordance with the findings of others (2,5,8,9,11).
Conclusions
Using random mutagenesis we have modified residues near the active site of TS to create 5-FdUR resistant enzymes. The 5-FdUR-selected library essentially consists of only a small subset of the active library mutant population. Survival of the library exposed to 175 nM 5-FdUR was approximately one-thousandth that of the non-drug treated active mutant library. Because the 53 clones sequenced only revealed 14 unique mutants, it is apparent that the library diversity has been narrowed down in the 5-FdUR library such that only a small number of clones are viable. Because each of the 53 clones was retransformed and rescreened against a range of dosages of 5-FdUR before sequencing, we can be quite confident that each of the mutants selected does indeed confer resistance to 5-FdUR in E.coli. The rank order of survival of TS resistant mutants in our E.coli assay correlates with survival in mammalian cells or kinetic inhibition values (D.M.Landis, unpublished). When created by site-directed mutagenesis and tested beside the novel mutants described here, the novel mutants demonstrated survival superior to that of Y33H [Ki ~4 × WT and IC50 ~4 × WT (9)] but not to G52S [Ki ~20 × WT and IC50 ~100 × WT (11)]. The survival of these mutants was equivalent to that of previously reported mutant A197V; L198I; C199F [Ki ~20 × WT (13)]. Thus, while 100% randomization of this region has defined the substitutability of residues in this region, it has not yet produced mutants with resistance to 5-FdUR greater than those previously described.
To gain insight into the mechanism for FdUMP resistance, in the absence of crystal structures for the mutants, three mutants were analyzed computationally (M23, M29 and M43) (Table 2). As the human bound structure of TS has yet to be published, the model is a chimera of the human unbound and E.coli FdUMP bound structures and therefore is an approximation. Interestingly, in this model, none of the mutants showed any significant structural differences in the `unbound' form whereas all three showed a consistent shift of residues toward the FdUMP binding site such that Cys 195 may no longer be in a suitable location for binding to 5-fluorouracil (Fig. 5B). No conclusions could be made about the lower activity of M43. The preference for polar, hydrogen bonding residues at position 204 in 34 of the 53 mutants developed might be understood as an indirect solvent mediated interaction on the R50 loop inasmuch as the C[alpha] of 204 is ~16 Å from the phosphate of FdUMP, but ~8 Å from residue 47 on monomer A. Residue 47 is on the Arg50 loop shown by Tong et al. to produce several mutants that are FdUMP resistant (11). It is interesting that mutation of residue 204 to Ser or Thr leading to FdUMP resistance occurs only in conjunction with mutations at residue 198 away from wild-type leucine.
The results of the genetic selection and kinetic assays of the TS mutants indicate that the design of a human 5-FdUR resistant TS mutant should include four basic features: replace A197 with a larger, non-aromatic hydrophobic residue such as leucine, methionine or valine; retain L198 as wild-type; create the highly selected alteration C199L; and finally, if V204 is to be mutated, alter it from a hydrophobic to a hydrophilic or charged residue.
Lastly, it is becoming increasingly apparent that many alterations in the TS enzyme can lead to 5-FdUR resistance. Although the precise mechanism for each mutant will differ, many common features will come to light. Structurally, TS undergoes what has been termed segmental accommodation upon substrate binding; the entire molecule has been observed to condense via many small additive movements (25). However, the loop corresponding to residues 181-195 has been shown to undergo a rather large twist upon binding the inhibitor-conjugate (14). It is apparent that alterations in many different regions of the protein can lead to drug resistance. These include mutations at the Arg50 loop residues 47-52 (11), mutants I108A and F225W (12), the naturally occurring mutants Y33H (6) and P303L (T.Spencer and F.Berger, personal communication), and the currently described region including residues 197-199 and 204 (13). None of these mutations are located directly in the active site since such mutations would likely lead to loss of wild-type activity. Our own mutations appear to make the local structure more rigid suggesting that binding of 5-FdUMP requires a `sloppier' active site.
ACKNOWLEDGEMENTS
We wish to thank Dr Robert Stroud for providing the coordinates of the human TS and for stimulating discussion and the reviewers for helpful suggestions. These studies were supported by a grant from the National Institute of Health CA78885 (L.A.L. and D.M.L.), the UW center grant P30 ES07033 from the National Institute of Environmental Health Sciences (E.T.A.), a Medical Scientist Training Program grant (NIH NIGMS 5 T32 07266 D.M.L.), and the Cora May Poncin Scholarship Fund (D.M.L.)
REFERENCES
*To whom correspondence should be addressed. Tel: +1 206 543 6015; Fax: +1 206 543 3967; Email: laloeb{at}u.washington.edu
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: jnl.info{at}oup.co.uk
Last modification:
Copyright© Oxford University Press, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  P. H. Petra, E. T. Adman, W. R. Orr, K. T. Woodcock, C. Groff, and L.-M. Sui Arginine-140 and isoleucine-141 determine the 17{beta}-estradiol-binding specificity of the sex-steroid-binding protein (SBP, or SHBG) of human plasma Protein Sci., September 1, 2001; 10(9): 1811 - 1821. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Skandalis and L. A. Loeb Enzymatic properties of rat DNA polymerase {beta} mutants obtained by randomized mutagenesis Nucleic Acids Res., June 1, 2001; 29(11): 2418 - 2426. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. M. Landis, C. C. Heindel, and L. A. Loeb Creation and Characterization of 5-Fluorodeoxyuridine-resistant Arg50 Loop Mutants of Human Thymidylate Synthase Cancer Res., January 1, 2001; 61(2): 666 - 672. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||