Biocatalyst Design for Stability and Specificity - American Chemical

1Current address: Department of Molecular Genetics, University of Texas Southwestern ..... (15) Hum, D. W.; Bell, A. W.; Rozen, R.; MacKenzie, R. E. J...
0 downloads 0 Views 1004KB Size
Chapter 17

C -Tetrahydrofolate Synthase 1

Dissection of Active Site and Domain by Protein Engineering 1

Structure

2

3

Anice E. Thigpen , Charles Κ. Barlowe , and Dean R. Appling

Downloaded by IOWA STATE UNIV on March 9, 2017 | http://pubs.acs.org Publication Date: December 11, 1993 | doi: 10.1021/bk-1993-0516.ch017

Department of Chemistry and Biochemistry, University of Texas, Austin, TX 78712 C -tetrahydrofolate (H folate) synthase is a trifunctional protein possess­ ing the activities 10-formyl-H folate synthetase, 5,10-methenyl-H folate cyclohydrolase, and 5,10-methylene-H folate dehydrogenase. The cur­ rent model divides the eukaryotic protein into two functionally independ­ ent domains with dehydrogenase/cyclohydrolase activities sharing an overlapping site on the N-terminal domain and synthetase activity asso­ ciated with the C-terminal domain. In prokaryotes, these activities are generally catalyzed by separate monofunctional enzymes. The genes or cDNAs encoding members of this family of enzymes have been isolated from mammalian, yeast, and bacterial sources and reveal considerable conservation of amino acid sequence. In an effort to elucidate structure/ function relationships in these proteins, we have utilized several recom­ binant D N A methodologies to dissect active site and domain structure in this family of enzymes. Site-directed and random mutagenesis has been used to identify important amino acid residues in the overlapping dehy­ drogenase/cyclohydrolase active site. Deletion analysis and construction and expression of chimeric forms of the trifunctional enzyme in which the two domains derive from different species have provided insights into domain structure and function. 1

4

4

4

4

Folate-mediated one-carbon metabolism plays an essential role in several major cellular processes including nucleic acid biosynthesis, mitochondrial and chloroplast protein biosynthesis, amino acid biosynthesis and interconversions, vitamin metabolism, and 1

Current address: Department of Molecular Genetics, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75235 Current address: Department of Biochemistry, University of California, Berkeley, CA 94720 Corresponding author

2

3

0097-6156/93/0516-0210$06.00/0 © 1993 American Chemical Society

Himmel and Georgiou; Biocatalyst Design for Stability and Specificity ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

17. THIGPEN ET AL.

211

C Tetrahydrofolate Synthase r

methyl group biogenesis. These pathways are summarized in Figure 1. The variety of the 3-carbon of serine, derived from glycolytic intermediates (i). The one-carbon unit is transferred to H folate in a reaction catalyzed by serine hydroxymethyltransferase (reaction 4, Figure 1), generating 5,10-methylene-H folate and glycine. This form of the coenzyme can be used directly in de novo dTMP synthesis by thymidylate synthase (reaction 6, Figure 1). 5,10-Methylene-H folate may also be reduced to 5-methylH folate or oxidized to 10-formyl-H folate depending on the needs of the cell. In rapidly growing cells, the synthesis of purines is a critical folate-dependent pathway, requiring two moles of 10-formyl-H folate per mole of purine ring. 5,10-Methylene-H folate is oxidized to 10-formyl-H folate via the sequential enzymes 5,10-methylene-H folate dehydrogenase and 5,10-methenyl-H folate cyclohydrolase (reactions 3 and 2, respec­ tively). Formate represents another potential one-carbon donor (2) and is activated in an ATP-dependent reaction catalyzed by 10-formyl-H folate synthetase (reaction 1, Figure 1). In prokaryotes, reactions 1-3 (Figure 1) are catalyzed by three separate monofunctional enzymes, with the known exceptions of Escherichia coli and Clostridium thermoaceticum in which the cyclohydrolase and dehydrogenase activities are catalyzed by bifunctional proteins. In eukaryotes, these three activities are present on one polypeptide in the form of a trifunctional enzyme. This enzyme, termed Cj-tetrahydrofolate synthase (Cj-THF synthase), is thus responsible for interconversion of the one-carbon unit between the formate and formaldehyde oxidation levels. A l l of the known eukaryotic C,-THF synthases exist as homodimers of approxi­ mately 100,000 Da. The current model divides the enzyme into two functionally independent domains with dehydrogenase/cyclohydrolase activities sharing an overlap­ ping active site on the N-terminal domain and synthetase activity associated with the Cterminal domain. This model is supported by several lines of evidence. For example, limited proteolysis results in physical separation of activities, with the synthetase activity associated with a large proteolytic fragment (subunit M = 60,000-80,000) and the 4

4

4

4

4

4

4

4

4

Downloaded by IOWA STATE UNIV on March 9, 2017 | http://pubs.acs.org Publication Date: December 11, 1993 | doi: 10.1021/bk-1993-0516.ch017

4

4

r

methionine + THF 4 I 5-methyl-THF

glycolysis . i serine

·* s \



5

1

* ' °-



m e t h

y

l e n e

"

T H F



Ή IF glycine

1

purines + THF A v \ »

*

+

NADP

5,10-methenyl-THF «

> NADPH

2

» 10-formyl-THF

formate





ADP ATP + THF

thymidylate + DHF

Figure 1. Tetrahydrofolate coenzymes in one-carbon metabolism. Pathways which utilize or generate one or more folate coenzymes are indicated schematically by arrows. Reactions 1, 2 and 3; 10-formyl-THF synthetase (EC 6.3.4.3), 5,10methenyl-THF cyclohydrolase (EC 3.5.4.9) and 5,10-methylene-THF dehydroge­ nase (EC 1.5.1.5), are catalyzed by C^THF synthase. Other reactions shown are 4, serine hydroxymethyltransferase (EC 2.1.2.1); 5, 5,10-methylene-THF reductase (EC 1.5.1.20); 6, thymidylate synthase (EC 2.1.1.45). THF, tetrahydrofolate; DHF, dihydrofolate.

Himmel and Georgiou; Biocatalyst Design for Stability and Specificity ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

212

BIOCATALYST DESIGN FOR STABILITY AND SPECIFICITY

dehydrogenase and cyclohydrolase activities with a small fragment (subunit M = 30,000) (5-5). Immunochemical experiments with the yeast enzyme (6) support the concept of two functionally independent domains. Villar et al. (5) used differential scanning calorimetry to demonstrate the existence of two domains in the rabbit enzyme. Schirch (7) demonstrated coordinate protection by NADP+ of the rabbit liver dehydrogenase/ cyclohydrolase activities against heat inactivation. 5,10-Methenyl-H folate, a product of the dehydrogenase reaction, does not accumulate in the coupled dehydrogenase/cyclohydrolase reaction (7-9). Chemical modification studies with the trifunctional enzyme (6,10-12) also support the concept of an overlapping active site for the dehydrogenase/ cyclohydrolase activities. These observations raise two major questions. First, what are the spatial and functional relationships between the two structural domains of C,-THF synthase? Second, what is the structure of the overlapping dehydrogenase/cyclohydrolase active site? A more complete understanding of the enzymology of these reactions requires new experimental approaches. The amino acid sequences of several species of C,-THF synthase have been derived from their cloned genes or complementary DNAs (13-16). In addition, sequences are available for several monofunctional and bifunctional versions of these enzymes (17-20). In the absence of a three dimensional structure, these clones provide the opportunity for a detailed analysis of catalytic mechanisms and active site architecture of this family of enzymes. This chapter describes several approaches utilizing recombinant D N A methods to address these questions. r

Downloaded by IOWA STATE UNIV on March 9, 2017 | http://pubs.acs.org Publication Date: December 11, 1993 | doi: 10.1021/bk-1993-0516.ch017

4

Site-directed mutagenesis The structure of the dehydrogenase/cyclohydrolase active site, mechanism of the coupling of these two reactions, and the identity of catalytic residues are unknown. Chemical modification experiments with the yeast cytoplasmic C,-THF synthase indi­ cate the existence of at least two critical histidyl residues and at least two critical cysteinyl residues at the dehydrogenase/cyclohydrolase active site (6). The putative dehydroge­ nase/cyclohydrolase domain, encompassing the first 300 or so amino acids, contains 11 histidines, but only 3 cysteines (73). Oligonucleotide-mediated site-directed mutagene­ sis was used to individually change each cysteine contained within the dehydrogenase/ cyclohydrolase domain (Cys-ll,Cys-144 and Cys-257) to serine (27). The resulting proteins were over-expressed in yeast and purified for kinetic analysis. Site-specific mutations in the dehydrogenase/cyclohydrolase domain do not affect synthetase activity, consistent with the proposed domain structure. The C144S and C257S mutations result in 7-fold and 2-fold increases, respectively, in the dehydrogenase K for NADP*. C144S lowers the dehydrogenase maximal velocity roughly 50% while C257S has a maximal velocity similar to the wild type. Cyclohydrolase catalytic activity is reduced 20-fold by the C144S mutation but is increased 2-fold by the C257S mutation. Conversion of Cys11 to serine has a negligible effect on dehydrogenase/cyclohydrolase activity. A double mutant, C144S/C257S, results in catalytic properties roughly multiplicative of the individual mutations. The dehydrogenase K for the folate substrate is not significantly changed in any of the mutants generated. These results implicate Cys-144 and Cys-257 as important active site residues for both the dehydrogenase and cyclohydrolase reacm

m

Himmel and Georgiou; Biocatalyst Design for Stability and Specificity ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

17. THIGPENETAL.

C^Tetrahydrofolate Synthase

213

tions, consistent with the overlapping active site model. Cys-144 is conserved in all five NADP-dependent 5,10-methylene-H folate dehydrogenases for which sequence infor­ mation is known (Figure 2). In contrast, the two known NAD*-dependent 5,10methylene-H folate dehydrogenases, which are approximately 50% identical to their NADP-dependent counterparts, have an alanine in the equivalent position. Since the major effect of the C144S mutation is on the K for N A D P , it is tempting to speculate that Cys-144 is involved in the binding specificity of N A D P vs. NAD*. 4

4

Downloaded by IOWA STATE UNIV on March 9, 2017 | http://pubs.acs.org Publication Date: December 11, 1993 | doi: 10.1021/bk-1993-0516.ch017

m

HBF

MLPATIWK

MBF

MLP|A|TÎWK

Y M

ÎFIPCTFJYJG

Y C

F L P C T P K G

H C

F I P C T P K G

R C

F I P C T P K G

E C

1..PCTPRG

(166-173) (172-179) (173-180) (141-148) (144-151) (144-151) (140-145)

Figure 2. Amino acid sequence conservation surrounding active site cysteine (bold) in N A D P - vs. NAD-dependent 5,10-methylene-H folate dehydrogenases. H B F , human bifunctional NAD-dependent dehydrogenase/cyclohydrolase (20); M B F , mouse bifunctional NAD-dependent dehydrogenase/cyclohydrolase (79); Y M , yeast mitochondrial C T H F synthase (74); Y C , yeast cytoplasmic Cj-THF synthase (13); HC, human cytoplasmic C^THF synthase (75); rat cytoplasmic Cj-THF synthase (16); EC, E. coli bifunctional NADP-dependent dehydrogenase/cyclohydrolase (33). A l l of the Cj-THF synthases contain the NADP-dependent dehydrogenase. Boxed residues indicate identical residues or conservative substitutions. 4

r

Random mutagenesis/functional selection Without a crystal structure of the enzyme or additional chemical modification data that might suggest important residues, picking targets for site-directed mutagenesis becomes a guessing game. In families of related enzymes, stretches of high sequence similarity often correspond to active site regions. However, the C,-THF synthases are highly conserved across their entire sequence, making it impossible to predict active site regions. We therefore developed a method in which random mutagenesis is coupled to a functional screening to identify those mutations that affect one or more of the activities of the trifunctional enzyme (22). The method relies on plasmid-borne expression of genes in strains of Saccharomyces cerevisiae that are missing one or more of the activities of C THF synthase. These specially constructed strains allow a very simple screening procedure for the detection of mutants in any of the three reactions. Specific segments of the gene or cDNA are subjected to random mutagenesis in vitro before expression and screening. The chemical mutagen used, nitrous acid, deaminates deoxycytosine, deoxyadenosine, and deoxyguanosine, resulting in C to Τ and A to G transitions (25). Plasmids encoding mutant enzymes are easily recovered for sequence analysis and subsequent overexpression of the mutant protein for enzymatic or structural analysis. The screening procedure is based on the earlier observation (24) that the nutritional requirement of serl mutants can be met by either serine, or glycine plus formate, although t

Himmel and Georgiou; Biocatalyst Design for Stability and Specificity ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

Downloaded by IOWA STATE UNIV on March 9, 2017 | http://pubs.acs.org Publication Date: December 11, 1993 | doi: 10.1021/bk-1993-0516.ch017

214

BIOCATALYST DESIGN FOR STABILITY AND SPECIFICITY

growth on glycine plus formate is considerably slower. The phosphorylated pathway to serine at the phosphoserine aminotransferase reaction is blocked in serl mutants. Under the glycine plus formate condition, serine is synthesized via the combined activities of Cj-THF synthase and serine hydroxymethyltransferase (reactions 1-4, Figure 1). Therefore, serl mutants which are also deficient in one or more of the activities in this pathway are unable to grow on glycine plus formate. Using this method, we have generated several mutant forms of the yeast cytoplasmic C,-THF synthase that result in deficient synthetase and/or dehydrogenase/cyclohydrolase activities (22,2526). When the synthetase domain was targeted, inactivating mutations were isolated at codons 428, 487, 492, and 565. None of these mutations affected the dehydrogenase/cyclohydrolase activities. Conversely, when the dehydrogenase/cyclohydrolase domain was targeted, inactivating mutations were isolated at codons 138,145,177,188,267,271, and 284. The effect on cyclohydrolase activity has not been determined in these mutants. These mutations did not affect the synthetase activity. Some of the mutant enzymes harbor multiple amino acid replacements, so it will be necessary to determine which mutated residues affect activity and which are silent. Beyond contributing information on active site residues, these redesigned enzymes have been extremely valuable in studying folate-mediated one-carbon metabolism in yeast. We have replaced the normal chromosomal gene of yeast with the mutant forms, resulting in new strains of yeast expressing full-length enzyme, but lacking one, two or all three of the C,-THF synthase activities. These 'metabolic engineering' experiments have revealed a structural role for C,-THF synthase in a putative multienzyme complex in de novo purine biosynthesis (25) and a new folate-dependent enzyme in yeast (27). Deletion analysis Deletion mutagenesis has been widely used to probe domain structure in multifunctional proteins. For example, many transcriptional activators have been dissected with this technique to delineate DNA binding domains, dimerization domains, and activation domains. We have used deletion mutagenesis to further define the domain structure of C,-THF synthases from yeast and rat. Several deletions were constructed in the yeast ADE3 gene, encoding the cytoplasmic Cj-THF synthase, shown schematically in Figure 3 (26). A small internal deletion in the dehydrogenase/cyclohydrolase domain (yAS/X; deletion of residues 114-144) resulted in a very unstable protein when expressed in yeast, with none of the three activities detectable over background. Deletion of 91 amino acids near the C-terminus of the synthetase domain (yACla) completely eliminated synthetase activity. Again, the protein was quite unstable and the dehydrogenase/cyclohydrolase activities were less than 10% of wild-type. Deletion of the entire synthetase domain (residues 335-946; yAH3) results in a protein with no synthetase activity, with dehydrogenase/cyclohydrolase activity at about 10% of wild-type. Deletion of the entire dehydrogenase/cyclohydrolase domain (residues 2-328; yADC) resulted in a stable protein of approximately 70 kDa when expressed in yeast. This deletion completely abolished dehydrogenase/cyclohydrolase activity, as expected, but also reduced synthetase activity to less than 10% of wild-type. It is quite likely that this deletion extended through the connecting peptide into the N -

Himmel and Georgiou; Biocatalyst Design for Stability and Specificity ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

17. THIGPENETAL

r

Dehydrogenase/ Cyclohydrolase 1 300 320

Downloaded by IOWA STATE UNIV on March 9, 2017 | http://pubs.acs.org Publication Date: December 11, 1993 | doi: 10.1021/bk-1993-0516.ch017

215

C Tetrahydrofolate Synthase

Phenotype Synthetase 946

Purine synthesis

Gly/form ser

yAS/X

no

no

yACla

yes

no

γΔΗ3

yes

yADC

no

Υ//////////////Λ

ΓΔΗ2

V^7Z^Z^7y

rASYN

no

yes

V/////////////A

rADC

rDCySYN

no

Y//////AU

WÎ//////////////À

yDCrSYN

yes

yes

yes

I I yeast

0

rat

Β

deleted sequences

Figure 3. Growth phenotypes of ade3' serl yeast expressing deletion mutants and domain chimeras of rat and yeast cytoplasmic C^-THF synthases from high copy number plasmids. The two columns on the right indicate whether the construct supports the synthesis of purines or the synthesis of serine from glycine + formate. The wild-type forms of both enzymes are positive for both phenotypes. Lower case r and y refer to rat and yeast proteins, respectively. Descriptions of each construct can be found in the text. terminus of the synthetase domain, perhaps deleting critical residues for that activity. Recendy, Hum and MacKenzie (28) used deletion analysis to define the interdomain region of the human Cj-THF synthase to residues 292-310 (corresponding to residues 297-315 of the yeast enzyme). Furthermore, homology at the N-termini of the two bacterial monofunctional synthetases begins around position 319 of the yeast enzyme. Examination of the amino acid sequence of the yeast enzyme reveals a likely connecting region between amino acids 306-321 that contains 6 proline residues and several potential protease digestion sites. Three deletions in the rat cytoplasmic Cj-THF synthase have recendy been con-

Himmel and Georgiou; Biocatalyst Design for Stability and Specificity ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

Downloaded by IOWA STATE UNIV on March 9, 2017 | http://pubs.acs.org Publication Date: December 11, 1993 | doi: 10.1021/bk-1993-0516.ch017

216

BIOCATALYST DESIGN FOR STABILITY AND SPECIFICITY

structed. The first (ΓΔΗ2) deleted amino acids 133-272 from the dehydrogenase/ cyclohydrolase domain, the second (rASyn) deleted amino acids 563-921 from the synthetase domain, and the third (rADC) deleted the entire dehydrogenase/cyclohydro­ lase domain (amino acids 2-277) (Figure 3). None of these constructs produced stable protein that could be detected by activity or immunoblot. Genetic complementation assays provide a sensitive measure of in vivo function(s) of a protein. We previously used genetic complementation of ade3 mutants for functional selection in our random mutagenesis experiments (see above). Complementation of a yeast strain harboring an ade3 chromosomal deletion with these deletion constructs on high-copy-number plasmids reveals some surprising results. Several of the constructs that did not exhibit enzyme activity in crude extracts nonetheless support growth of the ade3 deletion strain in media lacking purines. In fact, all the constructs involving synthetase deletions support growth in this strain, albeit at reduced rates. On the other hand, none of the dehydrogenase/cyclohydrolase domain deletions supported growth in the absence of exogenous adenine. We have proposed that cytoplasmic Cj-THF synthase is required as a structural component of a multienzyme complex involved in de novo purine biosynthesis, at least in yeast (25). This putative structural function can be differentiated from the enzymatic function of C,-THF synthase by testing for growth on glycine plus formate in place of serine. None of the deletion constructs support serine synthesis from glycine plus formate, since all three activities of Cj-THF synthase are required in this pathway (Figure 1). Thus, the structural role in purine synthesis is independent of any enzymatic function of Cj-THF synthase. The deletion results suggest that the dehydrogenase/cyclohydrolase domain is the structural portion of the protein involved in critical protein-protein interactions with other participants of the putative purine multienzyme complex. The nature and composition of this complex remains to be determined. Domain Chimeras One drawback to deletion analysis is that the deletions often have global effects. For example, the folding, stability, or activity of one domain may be dependent on the integrity of the deleted structure. This is apparently the case for some of the constructs described above, since many of the proteins proved to be quite unstable. A more subtle strategy is to substitute whole regions or domains of a protein with sequences from other proteins. We have used this approach to study the modular nature of the Cj-THF synthase family. These enzymes share a great deal of sequence similarity. The rat and human enzymes are greater than 9 0 % identical; the monofunctional synthetases from Clostrid­ ium share as high as 4 6 % identity with the synthetase domain of the rat enzyme (16). It is quite likely then, that these proteins are structurally related as well, and it might be possible to generate stable chimeric enzymes in which the dehydrogenase/cyclohydro­ lase and synthetase domains derive from different proteins. We have recently con­ structed two of these domain chimeras. rDCyS Y N is composed of the rat dehydrogenase/ cyclohydrolase domain (amino acids 1-277) fused to the yeast synthetase domain (amino acids 283-946). yDCrSYN is composed of the yeast dehydrogenase/cyclohydrolase domain (amino acids 1-334) fused to the rat synthetase domain (amino acids 331-935).

Himmel and Georgiou; Biocatalyst Design for Stability and Specificity ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

17. THIGPEN ET AL.

C -Tetrahydrofolate Synthase 1

217

Downloaded by IOWA STATE UNIV on March 9, 2017 | http://pubs.acs.org Publication Date: December 11, 1993 | doi: 10.1021/bk-1993-0516.ch017

Both of these proteins have been expressed from high-copy-number plasmids in an ade3 deletion strain. Both constructs are able to complement the purine auxotrophy of the host strain, and both support the synthesis of serine from glycine plus formate (Figure 3). The enzymatic activity of these chimeric proteins, as measured in crude extracts, is quite low, suggesting that the constructs are unstable and/or poorly expressed. Some of these extracts have been examined by immunoblotting with antisera raised against yeast or rat Cj-THF synthase and only very low levels of the full-length chimeras can be detected. Nevertheless, the in vivo complementation results confirm the modular nature of C,-THF synthase and indicate that the domains derived from different sources function together efficiently enough to support growth of the yeast Future Directions The work presented here represents only our initial efforts at studying the structure and functions of the multifunctional enzyme C,-THF synthase though protein engineering. We are currently working on replacement of the NADP*-dependent dehydrogenase/ cyclohydrolase domain of Q-THF synthase with a human NAD*-dependent dehydroge­ nase/cyclohydrolase protein and replacement of the synthetase domain of the eukaryotic enzyme with a bacterial synthetase structure. These constructs are of interest from both the structural and metabolic points of view, and the yeast expression system provides a valuable window on in vivo function. However, several technical problems remain to be solved, the foremost being that of protein instability. Although most of the constructs described clearly function in vivo, their relative instability makes it difficult to obtain them in sufficient amounts for structural analysis. Towards this end, we are constructing a series of yeast expression vectors in which the various C,-THF synthase sequences are fused to the C-terminus of the 76-amino acid human ubiquitin. This strategy has proven successful for the stable expression in yeast of a number of heterologous proteins (29-52). An improved expression system should allow the purification of enough of these proteins to carry out detailed structural studies so that we may better understand how these important multifunctional proteins function in the cell. Acknowledgments This work was supported by grants to D.R.A. from the National Institutes of Health (DK36913) and from the Foundation for Research.

Literature Cited (1) Schirch, L. In Folates and Pterins; R. L. Blakley and S. J. Benkovic, Ed.; Wiley: New York, 1984; Vol. 1; pp 399-431. (2) Barlowe, C. K.; Appling, D. R. Biofactors 1988, 1, 171-176. (3) Paukert, J. L.; Williams, G. R.; Rabinowitz, J. C. Biochem. Biophys. Res. Comm. 1977, 77, 147-154. (4) Tan, L. U. L.; Drury, E. J.; MacKenzie, R. E.J.Biol. Chem. 1977, 252, 1117-1122. (5) Villar, E.; Schuster, B.; Peterson, D.; Schirch, V. J. Biol. Chem. 1985, 260, 22452252. (6) Appling, D. R.; Rabinowitz, J. C. Biochemistry 1985, 24, 3540-3547. (7) Schirch, L. Arch. Biochem. Biophys. 1978, 189, 283-290.

Himmel and Georgiou; Biocatalyst Design for Stability and Specificity ACS Symposium Series; American Chemical Society: Washington, DC, 1993.

Downloaded by IOWA STATE UNIV on March 9, 2017 | http://pubs.acs.org Publication Date: December 11, 1993 | doi: 10.1021/bk-1993-0516.ch017

218

BIOCATALYST DESIGN FOR STABILITY AND SPECIFICITY

(8) Cohen, L.; MacKenzie, R. E. Biochim. Biophys. Acta 1978, 522, 311-317. (9) Wasserman, G. F.; Benkovic, P. Α.; Young, M.; Benkovic, S. J. Biochemistry 1983, 22, 1005-1013. (10) Schirch, L.; Mooz, E. D.; Peterson, D. In Chemistry and Biology ofPteridines; R. L. Kisluik and G. M. Brown, Ed.; Elsevier/North-Holland: Amsterdam, 1979; pp 495-500. (11) Smith, D. D. S.; MacKenzie, R. E. Can. J. Biochem. Cell Biol. 1983, 61, 1166-1171. (12) Smith, D. D. S.; MacKenzie, R. E. Biochem. Biophys. Res. Comm. 1985, 128, 148154. (13) Staben, C.; Rabinowitz, J. C. J. Biol. Chem. 1986, 261, 4629-4637. (14) Shannon, K. W.; Rabinowitz, J. C. J. Biol. Chem. 1988, 263, 7717-7725. (15) Hum, D. W.; Bell, A. W.; Rozen, R.; MacKenzie, R. E. J. Biol. Chem. 1988, 263, 15946-15950. (16) Thigpen, A. E.; West, M. G.; Appling, D. R. J. Biol. Chem. 1990, 265, 7907-7913. (17) Whitehead, T. R.; Rabinowitz, J. C. J. Bacteriol. 1988, 170, 3255-3261. (18) Lovell, C. R.; Przybyla, Α.; Ljungdahl, L. G. Biochemistry 1990, 29, 5687-5694. (19) Belanger, C.; MacKenzie, R. E. J. Biol. Chem. 1989, 264, 4837-4843. (20) Peri, K. G.; Belanger, C.; MacKenzie, R. E. Nucleic Acids Res. 1989, 17, 8853. (21) Barlowe, C. K.; Williams, M. E.; Rabinowitz, J. C.; Appling, D. R. Biochemistry 1989, 28, 2099-2106. (22) Barlowe, C. K.; Appling, D. R. Biofactors 1989, 2, 57-63. (23) Myers, R. M.; Lerman, L. S.; Maniatis, T. Science 1985, 229, 242-247. (24) McKenzie, K. Q.; Jones, E. W. Genetics 1977, 86, 85-102. (25) Barlowe, C. K.; Appling, D. R. Mol. Cell.Biol.1990, 10, 5679-5687. (26) Barlowe, C. K. Ph.D. Thesis, The University of Texas at Austin, 1990. (27) Barlowe, C. K.; Appling, D. R. Biochemistry 1990, 29, 7089-7094. (28) Hum, D. W.; MacKenzie, R. E. Prot. Eng. 1991, 4, 493-500. (29) Ecker, D. J.; Stadel, J. M.; Butt, T. R.; Marsh, J. Α.; Monia, B. P.; Powers, D. Α.; Gorman, J. Α.; Clark, P. E.; Warren, F.; Shatzman, Α.; Crooke, S. T. J. Biol. Chem. 1989, 264, 7715-7719. (30) Butt, T. R.; Khan, M. I.; Marsh, J.; Ecker, D. J.; Crooke, S. T.J.Biol. Chem. 1988, 263, 16364-16371. (31) Sone, T.; McDonnell, D. P.; O'Malley, B. W.; Pike, J. W. J. Biol. Chem. 1990, 265, 21997-22003. (32) Baker, R. T.; Varshavsky, A. Proc. Natl. Acad. Sci. USA 1991, 88, 1090-1094. (33) D'Ari, L.; Rabinowitz, J. C. J. Biol. Chem. 1991, 266, 23953-23958. RECEIVED March 31, 1992

Himmel and Georgiou; Biocatalyst Design for Stability and Specificity ACS Symposium Series; American Chemical Society: Washington, DC, 1993.