Reconstruction of an enzyme by domain substitution effectively switches substrate specificity

Biosynthesis of Arginine and Polyamines

Early investigations on arginine biosynthesis brought to light basic features of metabolic regulation. The most significant advances of the last 10 to 15 years concern the arginine repressor, its structure and mode of action in both E. coli and Salmonella typhimurium, the sequence analysis of all arg structural genes in E. coli and Salmonella typhimurium, the resulting evolutionary inferences, and the dual regulation of the carAB operon. This review provides an overall picture of the pathways, their interconnections, the regulatory circuits involved, and the resulting interferences between arginine and polyamine biosynthesis. Carbamoylphosphate is a precursor common to arginine and the pyrimidines. In both Escherichia coli and Salmonella enterica serovar Typhimurium, it is produced by a single synthetase, carbamoylphosphate synthetase (CPSase), with glutamine as the physiological amino group donor. This situation contrasts with the existence of separate enzymes specific for arginine and pyrimidine biosynthesis in Bacillus subtilis and fungi. Polyamine biosynthesis has been particularly well studied in E. coli, and the cognate genes have been identified in the Salmonella genome as well, including those involved in transport functions. The review summarizes what is known about the enzymes involved in the arginine pathway of E. coli and S. enterica serovar Typhimurium; homologous genes were identified in both organisms, except argF (encoding a supplementary OTCase), which is lacking in Salmonella. Several examples of putative enzyme recruitment (homologous enzymes performing analogous functions) are also presented.

cse, a Chimeric and variable gene, encodes an extracellular protein involved in cellular segregation in Streptococcus thermophilus

The isolation of a Streptococcus thermophilus CNRZ368 mutant displaying a long-chain phenotype allowed us to identify the cse gene (for cellular segregation). The N terminus of Cse exhibits high similarity to Streptococcus agalactiae surface immunogenic protein (SIP), while its C terminus exhibits high similarity to S. thermophilus PcsB. In CNRZ368, deletion of the entire cse open reading frame leads to drastic lengthening of cell chains and altered colony morphology. Complementation of the Deltacse mutation with a wild-type allele restored both wild-type phenotypes. The central part of Cse is a repeat-rich region with low sequence complexity. Comparison of cse from CNRZ368 and LMG18311 strains reveals high variability of this repeat-rich region. To assess the impact of this central region variability, the central region of LMG18311 cse was exchanged with that of CNRZ368 cse. This replacement did not affect chain length, showing that divergence of the central part does not modify cell segregation activity of Cse. The structure of the cse locus suggests that the chimeric organization of cse results from insertion of a duplicated sequence deriving from the pcsB 3' end into an ancestral sip gene. Thus, the cse locus illustrates the module-shuffling mechanism of bacterial gene evolution.

Extreme functional sensitivity to conservative amino acid changes on enzyme exteriors

Mutagenesis studies and alignments of homologous sequences have demonstrated that protein function typically is compatible with a variety of amino-acid residues at most exterior non-active-site positions. These observations have led to the current view that functional constraints on sequence are minimal at these positions. Here, it is shown that this inference assumes that the set of acceptable residues at each position is independent of the overall sequence context. Two approaches are used to test this assumption. First, highly conservative replacements of exterior residues, none of which would cause significant functional disruption alone, are combined until roughly one in five have been changed. This is found to cause complete loss of function in vivo for two unrelated monomeric enzymes: barnase (a bacterial RNase) and TEM-1 beta-lactamase. Second, a set of hybrid sequences is constructed from the 50 %-identical TEM-1 and Proteus mirabilis beta-lactamases. These hybrids match the TEM-1 sequence except for a region at the C-terminal end, where they are random composites of the two parents. All of these hybrids are biologically inactive. In both experiments, complete loss of activity demonstrates the importance of sequence context in determining whether substitutions are functionally acceptable. Contrary to the prevalent view, then, enzyme function places severe constraints on residue identities at positions showing evolutionary variability, and at exterior non-active-site positions, in particular. Homologues sharing less than about two-thirds sequence identity should probably be viewed as distinct designs with their own sets of optimising features.

Substrate-induced conformational change in a trimeric ornithine transcarbamoylase

The crystal structure of Escherichia coli ornithine transcarbamoylase (OTCase, EC 2.1.3.3) complexed with the bisubstrate analog N-(phosphonacetyl)-L-ornithine (PALO) has been determined at 2.8-A resolution. This research on the structure of a transcarbamoylase catalytic trimer with a substrate analog bound provides new insights into the linkages between substrate binding, protein-protein interactions, and conformational change. The structure was solved by molecular replacement with the Pseudomonas aeruginosa catabolic OTCase catalytic trimer (Villeret, V., Tricot, C., Stalon, V. & Dideberg, O. (1995) Proc. Natl. Acad. Sci. USA 92, 10762-10766; Protein Data Bank reference pdb 1otc) as the model and refined to a crystallographic R value of 21.3%. Each polypeptide chain folds into two domains, a carbamoyl phosphate binding domain and an L-ornithine binding domain. The bound inhibitor interacts with the side chains and/or backbone atoms of Lys-53, Ser-55, Thr-56, Arg-57, Thr-58, Arg-106, His-133, Asn-167, Asp-231, Met-236, Leu-274, Arg-319 as well as Gln-82 and Lys-86 from an adjacent chain. Comparison with the unligated P. aeruginosa catabolic OTCase structure indicates that binding of the substrate analog results in closure of the two domains of each chain. As in E. coli aspartate transcarbamoylase, the 240s loop undergoes the largest conformational change upon substrate binding. The clinical implications for human OTCase deficiency are discussed.

Expression, purification and kinetic characterization of wild-type human ornithine transcarbamylase and a recurrent mutant that produces 'late onset' hyperammonaemia

Ornithine Transcarbamylase Deficiency, an X-linked disorder, is the most common cause of inherited urea cycle disorders. Approx. 90 mutations that produce reduced levels of ornithine transcarbamylase (OTCase) activity have been identified in patients [Tuchman (1993) Hum. Mutat. 2, 174-178; Tuchman and Plante (1995) Hum. Mutat. 5, 293-295]. A model of the three-dimensional structure of OTCase, developed on the basis of its homology to the catalytic subunit of Escherichia coli aspartate transcarbamylase (ATCase) [Tuchman, Morizono, Reish, Yuan and Allewell (1995) J. Med. Genet. 32, 680-688], and in good agreement with the crystal structure of Pseudomonas aeruginosa OTCase [Villeret, Tricot, Stalon and Dideberg (1995) Proc. Natl. Acad. Sci. U.S.A. 92, 10762-10766], indicates that many mutations that produce severe clinical symptoms are at the active site or buried in the interior of the protein. However, one of the few recurrent mutations, R277W, an alteration that produces a milder phenotype of ornithine transcarbamylase deficiency, is located in the model in a loop remote from the active site that is analogous to a similar loop (the 240's loop, a flexible loop of the catalytic chain of Escherichia coli aspartate transcarbamylase, comprised of residues 230-250) of ATCase. Human wild-type OTCase and the R277W mutant have been cloned and overexpressed in E. coli and a rapid and efficient purification method utilizing the bisubstrate analogue, Ndelta-(phosphonacetyl)-L-ornithine, has been developed and used to purify both proteins. Gel chromatography indicates both are trimeric. The pH dependence of the kinetic parameters of the wild-type enzyme is similar to that of E. coli OTCase [Kuo, Herzberg and Lipscomb (1985) Biochemistry 24, 4754-4761], suggesting that its catalytic mechanism is similar, although its maximal activity is approx. 10-fold less. Compared with the wild-type, the R277W mutant has nearly 70-fold lower affinity for L-ornithine, shows no substrate inhibition, and its thermal stability is reduced by 5 degrees C. Its reduced affinity for L-ornithine, which in turn results in lower activity at physiological concentrations of ornithine, as well as its reduced stability, may contribute to the clinical effects that it produces.

4-Oxalocrotonate tautomerase, a 41-kDa homohexamer: backbone and side-chain resonance assignments, solution secondary structure, and location of active site residues by heteronuclear NMR spectroscopy

4-Oxalocrotonate tautomerase (4-OT), a homohexamer consisting of 62 residues per subunit, catalyzes the isomerization of unsaturated alpha-keto acids using Pro-1 as a general base (Stivers et al., 1996a, 1996b). We report the backbone and side-chain 1H, 15N, and 13C NMR assignments and the solution secondary structure for 4-OT using 2D and 3D homonuclear and heteronuclear NMR methods. The subunit secondary structure consists of an alpha-helix (residues 13-30), two beta-strands (beta 1, residues 2-8; beta 2, residues 39-45), a beta-hairpin (residues 50-57), two loops (I, residues 9-12; II, 34-38), and two turns (I, residues 30-33; II, 47-50). The remaining residues form coils. The beta 1 strand is parallel to the beta 2 strand of the same subunit on the basis of cross stand NH(i)-NH(j) NOEs in a 2D 15N-edited 1H-NOESY spectrum of hexameric 4-OT containing two 15N-labeled subunits/hexamer. The beta 1 strand is also antiparallel to another beta 1 strand from an adjacent subunit forming a subunit interface. Because only three such pairwise interactions are possible, the hexamer is a trimer of dimers. The diffusion constant, determined by dynamic light scattering, and the rotational correlation time (14.5 ns) estimated from 15N T1/T2 measurements, are consistent with the hexameric molecular weight of 41 kDa. Residue Phe-50 is in the active site on the basis of transferred NOEs to the bound partial substrate 2-oxo-1,6-hexanedioate. Modification of the general base, Pro-1, with the active site-directed irreversible inhibitor, 3-bromopyruvate, significantly alters the amide 15N and NH chemical shifts of residues in the beta-hairpin and in loop II, providing evidence that these regions change conformation when the active site is occupied.

Structural similarity between ornithine and aspartate transcarbamoylases of Escherichia coli: implications for domain switching

Each catalytic (c) polypeptide chain of Escherichia coli aspartate transcarbamoylase (ATCase) is composed of two globular domains connected by two interdomain helices. Helix 12, near the C-terminus, extends from the second domain back through the first domain, bringing the two termini close together. This helix is of critical importance for the assembly of a stable enzyme. The trimeric E. coli enzyme ornithine transcarbamoylase (OTCase) is proposed to be similar in tertiary and quaternary structure to the ATCase trimer and has a predicted alpha-helical segment near its C-terminus. In our companion paper, we have shown that this putative helix is essential for OTCase folding and assembly (Murata L, Schachman HK, 1996, Protein Sci 5:709-718). Here, the similarity between OTCase and the ATCase trimer, which are 32% identical in sequence, was tested further by the construction of several chimeras in which various structural elements were switched between the enzymes by genetic techniques. These elements included the two globular domains and regions containing the C-terminal helices. In contrast to results reported previously (Houghton J, O'Donovan G, Wild J, 1989, Nature 338:172-174), none of the chimeric proteins exhibited in vivo activity and all were insoluble when overexpressed. Attempts to make hybrid trimers composed of c chains from ATCase and OTCase were also unsuccessful. These results underscore the complexities of specific intrachain and interchain side-chain interactions required to maintain tertiary and quaternary structures in these enzymes.

The molecular basis of ornithine transcarbamylase deficiency: modelling the human enzyme and the effects of mutations

Human ornithine transcarbamylase is a trimer with 46% amino acid sequence homology to the catalytic subunit of E coli aspartate transcarbamylase. Secondary structure predictions, distributions of hydrophilic and hydrophobic regions, and the pattern of conserved residues suggest that the three dimensional structures of the two proteins are likely to be similar. A three dimensional model of ornithine transcarbamylase was generated from the crystal structure of the catalytic subunit of E coli aspartate transcarbamylase in the holoenzyme, by aligning the sequences, building in gaps, and minimising the energy. The binding sites for carbamyl phosphate in both enzymes are similar and the ornithine binding site in ornithine transcarbamylase appears to be in the same location as the L-aspartate binding site in aspartate transcarbamylase, with negatively charged side chains replaced by positively charged residues. Mutations in the ornithine transcarbamylase gene found in patients with hyperammonaemia of the "neonatal type" are clustered in important structural or functional domains, either in the interior of the protein, at the active site, or at the interchain interface, while mutations found in patients with milder "late onset" disease are located primarily on the surface of the protein. The predicted effects of all known missense mutations and in frame deletions in the ornithine transcarbamylase gene on the structure and function of the mature enzyme are described.

Domain exchange: characterization of a chimeric lipase of hepatic lipase and lipoprotein lipase

Hepatic lipase and lipoprotein lipase hydrolyze fatty acids from triacylglycerols and are critical in the metabolism of circulating lipoproteins. The two lipases are similar in size and amino acid sequence but are distinguished by functional differences in substrate preference and cofactor requirement. Presumably, these distinctions result from structural differences in functional domains. To begin localization of these domains, a chimeric lipase was constructed composed of the N-terminal 329 residues of rat hepatic lipase linked to the C-terminal 136 residues of human lipoprotein lipase. The chimera hydrolyzed both monodisperse short-chain (esterase) and emulsified long-chain (lipase) triacylglycerol substrates with catalytic and kinetic properties closely resembling those of native hepatic lipase. However, monoclonal antibodies to lipoprotein lipase inhibited the lipase activity, but not the esterase function, of the chimera. Therefore, the chimeric molecule is a functional lipase and contains elements and characteristics from both parental enzymes. It is proposed that the N-terminal domain, containing the active center from hepatic lipase, governs the catalytic character of the chimera, and the C-terminal domain is essential for hydrolysis of long-chain substrates.

Mutational remodeling of enzyme specificity

With the advent of genetic engineering techniques has come the ability to modify proteins as desired. Given this stunning capability, the question remains what residues should be altered, and how should they be changed to achieve a particular specificity pattern. The goals of such modifications are likely to fall into either of two categories: probing the function of a protein or attempting to alter its properties. In either case, our understanding of the consequences of a mutation, as ascertained by our ability to predict the results, is currently quite limited. The problem is extraordinarily complex; our understanding of how to calculate the energetics involved is still incomplete, and we are just beginning to accumulate experimental data which may help guide us. On the positive side, theoretical methods are now being developed and refined that should prove useful in the drive to engineer enzyme specificity. What may be most important at this juncture is to expand the experimental database interrelating sequence, function, and structure. That is, there should be a concerted effort to combine functional analysis of mutant proteins with structural analysis. Only from this combined examination of the effects of mutations can sufficient data be accumulated to test and improve both qualitative and quantitative approaches or methods for remodeling enzyme specificity.

Independent folding of individual components in hybrid proteins. Evidence that the carboxy-terminal 135 residues of the LexA repressor constitute a single autonomous domain

Inactivation of the Escherichia coli repressor protein, LexA, takes place through a cleavage reaction which hydrolyzes the Ala84-Gly85 peptide bond near the center of the molecule. The mechanism of cleavage has previously been shown to be an intramolecular reaction stimulated in vitro by elevated pH or by the addition of activated RecA protein. The entire self-cleavage activity of LexA has been found to lie within a 135-residue tryptic fragment extending from Leu68 to the end of the protein at Leu202. Since the activity of self-cleavage is dependent on the proper three-dimensional structure of the protein, we have used it as a probe to investigate the extend of folding autonomy and functional independence of this 135-residue carboxy-terminal domain of LexA by applying a protein fusion approach. A series of twelve different hybrid proteins, containing LexA sequences in a variety of predefined primary structural arrangements, were constructed and evaluated for whether or not self-cleavage activity has been retained. The results revealed that retention or loss of activity is independent of the nature or size of the foreign protein used. Loss of self-cleavage was found to be a function of amino- or carboxy-terminal deletions in the self-cleaving LexA component of the fusion proteins. The present findings, together with the observations of other artificial fusions proteins and the naturally occurring bifunctional and multifunctional proteins, along with the data on helix packing, provide further support for the notion of modular architecture of proteins and suggest that when these autonomous units are fused, they retain their tendency to fold independently of the remainder of the polypeptide to generate physically linked active domains, rather than to fold dependently and yield scrambled structures.

Chimeric phage-bacterial enzymes: a clue to the modular evolution of genes

Pneumococcal peptidoglycan amidase (N-acetylmuramoyl-L-alanine amidase, EC 3.5.1.28) and phage CPL1 lysozyme degrade a common substrate (choline-containing pneumococcal cell walls); the former hydrolyzes the bond between muramic acid and alanine, whereas the latter breaks down the linkage between muramic acid and glucosamine. The amino acid sequences of their C-terminal domains are homologous. Chimeric genes were constructed by site-directed mutagenesis: a unique SnaBI restriction site in the cpl1 gene, coding for the phage lysozyme, was introduced at a location equivalent to the SnaBI site present in the lytA gene, which codes for the pneumococcal amidase. The resulting genes expressed lytic activities at levels similar to those of the parental genes. The gene products, which have been purified to electrophoretical homogeneity, exhibited unusual combined biochemical properties--e.g., by exchange of protein domains, we have switched the regulatory properties of these enzymes without altering their catalytic activities. Chimeric gene construction in Streptococcus pneumoniae and its bacteriophages is an excellent model to study the modular organization of genes and proteins and to help to establish evolutionary relationships between phage and bacteria. These constructions provide an experimental approach to the molecular processes involved in cassette recruitment during evolution and contribute support to the concept of bacteria as adaptable chimeras.

Engineering aspartate transcarbamylase

Aspartate transcarbamylase from Escherichia coli is one of the most extensively studied regulatory enzymes as a model of cooperativity and allostery. Numerous methods are used to engineer variants of this molecule: random and site-directed mutagenesis, dissociation and reassociation of the catalytic and regulatory subunits and chains, construction of hybrids made from normal and modified subunits or chains, interspecific hybrids and construction of chimeric enzymes. These methods provide detailed information on the regions, domains, interfaces and aminoacid residues which are involved in the mechanism of co-operativity between the catalytic sites, and of regulation by the antagonistic effectors CTP and ATP. These effectors induce the transmission of intramolecular signals whose pathways begin to be delineated.

Interdomain hybrid Tet proteins confer tetracycline resistance only when they are derived from closely related members of the tet gene family

Inner membrane Tet proteins encoded by tet genes in gram-negative bacteria mediate resistance to tetracycline (Tcr) by directing its export. Total sequences for class A, B, and C tet genes demonstrate that their products have a common ancestor, with Tet(A) and Tet(C) being more closely related (78% identical) than either is to Tet(B) (45% identical). The N- and C-terminal halves of Tet(B) and Tet(C) appear to comprise separate domains, and trans-complementation observed between tetracycline sensitive mutants in either domain of Tet(B) suggests separate but interactive functions for these domains. In this present study, interdomain hybrid genes were constructed to express hybrid tet products whose N- and C-terminal halves were derived from different family members [Tet(A/C), Tet(B/C), and Tet(C/B)]. Tet(A/C) specified a level of Tcr comparable to wild-type Tet(C) and 60% that of Tet(A), indicating that domains from these closely related tet products can function in cis. Although neither Tet(B/C) nor Tet(C/B) hybrids conferred significant Tcr, cells producing both of these types of hybrid proteins expressed substantial Tcr, indicating that productive interactions can occur in trans between Tet(B/C) and Tet(C/B). Taken together, these results suggest that highly specific interactions between the N- and C-terminal domains are necessary for Tcr and do not occur in individual hybrids derived from the more distant relatives, Tet(B) and Tet(C). This requirement for specific interactions suggests that N- and C-terminal domains have coevolved in each member of the Tet family.

Altering enzymatic activity: recruitment of carboxypeptidase activity into an RTEM beta-lactamase/penicillin-binding protein 5 chimera

The D-Ala-D-Ala carboxypeptidases/transpeptidases (penicillin-binding proteins, PBPs) share considerable structural homology with class A beta-lactamases (EC 3.5.2.6), although these beta-lactamases have no observable D-Ala-D-Ala carboxypeptidase activity. With the objective of recruiting such activity into a beta-lactamase background, we have prepared a chimeric protein by inserting a 28-amino acid segment of PBP-5 of Escherichia coli in place of the corresponding region of the RTEM-1 beta-lactamase. The segment thus inserted encompasses two residues conserved in both families: Ser-70, which forms the acyl-enzyme intermediate during beta-lactam hydrolysis, and Lys-73, whose presence has been shown to be necessary for catalysis. This chimera involves changes of 18 residues and gives a protein that differs at 7% of the residues from the parent. Whereas RTEM beta-lactamase has no D-Ala-D-Ala carboxypeptidase activity, that of the chimera is significant and is, in fact, about 1% the activity of PBP-5 on diacetyl-L-Lys-D-Ala-D-Ala; in terms of free energy of activation, the chimera stabilizes the transition state for the reaction to within about 2.7 kcal/mol of the stabilization achieved by PBP-5. Furthermore, the chimera catalyzes hydrolysis exclusively at the carboxyl-terminal amide bond which is the site of cleavage by D-Ala-D-Ala carboxypeptidase. Though containing all those residues that are conserved throughout class A beta-lactamases and are thought to be essential for beta-lactamase activity, the chimera has considerably reduced activity (approximately 10(-5) on penams such as penicillins and ampicillins as substrates. As a catalyst, the chimera shows an induction period of approximately 30 min, reflecting a slow conformational rearrangement from an inactive precursor to the active enzyme.

Use of bromovirus RNA2 hybrids to map cis- and trans-acting functions in a conserved RNA replication gene

Brome mosaic virus (BMV) and cowpea chlorotic mottle virus (CCMV) are related positive-strand RNA viruses with tripartite genomes. RNA replication by either virus requires genomic RNAs 1 and 2, which encode protein 1a and the polymeraselike, 94-kilodalton 2a protein, respectively. Proteins 1a and 2a share extensive sequence similarity with proteins encoded by a wide range of other positive-strand RNA viruses of animals and plants. Heterologous combinations of BMV and CCMV RNAs 1 and 2 do not support viral RNA replication, and although BMV RNA2 is amplified in CCMV-infected cells, CCMV RNA2 is not amplified by BMV. Construction of hybrids by precise exchange of segments between BMV and CCMV RNA2 has now allowed preliminary mapping of such virus-specific replication functions in RNA2 and the 2a protein. The ability to support replication in trans with BMV RNA1 segregated with a 5' BMV RNA2 fragment encoding the first 358 2a gene amino acids, while a 5' fragment extending over 281 BMV 2a codons transferred only cis-acting competence for RNA2 amplification in cells coinfected with wild-type BMV. Successful trans-acting function with CCMV RNA1 segregated with a CCMV RNA2 3' fragment that included the last 206 2a gene codons. Thus, the less conserved N- and C-terminal 2a segments appear to be involved in required interaction(s) of this polymeraselike protein with the 1a protein or RNA1 or both. Moreover, when individual hybrid RNA2 molecules that function with either BMV or CCMV RNA1 were tested, BMV- and CCMV-specific differences in recognition and amplification of RNA3 templates appeared to segregate with RNA1.

Understanding structural relationships in proteins of unsolved three-dimensional structure

The locations of functionally important sequences and general structural motifs have been assigned to Ile-tRNA synthetase. However, a function has not been established for some segments of the protein (e.g., CP1). The method of structural modeling described here cannot establish the details of a 3 A crystal structure, and, in contrast to a crystal structure, the precision of the model varies according to the extent of a sequence similarity or the functional importance of a region. In Ile-tRNA synthetase, the signature sequence and the flanking regions are likely to be similar in structure to the proteins on which the model is based. For other regions, it may be possible to build a three-dimensional model by connecting well defined regions and refining the positions of the connecting elements by energy minimization. Structural modelling of this kind must be done cautiously, because the order and orientation of the elements of a structural motif can change in subtle ways. In the case of Tyr-tRNA synthetase, the beta-strand nearest the N-terminus is the outermost strand of the nucleotide binding fold; in Met-tRNA synthetase, the same strand is innermost. Furthermore, the orientation of this strand may be antiparallel (Tyr-tRNA synthetase) or parallel (Met-tRNA synthetase). Because multiple structures that differ in their orientations of structural elements are possible, the structural analogies between proteins should not be naively extrapolated without independent experimental support. As described above, some regions of proteins tolerate internal deletions and insertions. This provides further experimental support for the practice of allowing for gaps in computer-generated sequence alignments. Nevertheless, because some regions are more tolerant of insertions and deletions than others, the structural and functional significance of a region of broken alignment must be assessed carefully. All gaps in sequence alignments cannot be treated equally, and each must be evaluated within its own context. In the synthetases of known structure, structural analogy can be used to identify important functional elements. For example, the amino acid binding site of Met-tRNA synthetase might be formed, at least in part, by a peptide that encompasses Ala50; this amino acid aligns with Gly94 of the Ile-tRNA synthetase. This is an example in which results on a protein of unknown structure (Ile-tRNA synthetases) can lead to identification of a potential substrate binding site in a protein of known structure (Met-tRNA synthetase).

An experimental approach to testing modular evolution: directed replacement of alpha-helices in a bacterial protein

We have used oligonucleotide site-directed mutagenesis to ask whether certain structural motifs in proteins are determined mainly by local interactions among amino acids. Multiple consecutive amino acids in three alpha-helices in the alkaline phosphatase (EC 3.1.3.1) of Escherichia coli have been replaced with helical sequences from four other sources. Altogether, 12 distinct helical replacements were created, 9 of which retain enzymatic activity. Most short stretches of helical sequence can be replaced with unrelated helical sequences without eliminating enzyme activity. Replacements of the carboxyl half of an alpha-helix are less harmful than those of the amino half, and the two together are synergistic rather than additive. These results are consistent with the hypothesis that proteins originally evolved by the assembly of small functional folding units.

Construction of intra-domain chimeras of aminoacyl-tRNA synthetases

To explore new approaches to enzyme engineering, intra-domain chimeras of two aminoacyl-tRNA synthetases were constructed. Connections were made within the nucleotide folds of these enzymes at sites earlier shown either to be dispensable for activity or able to accommodate oligopeptide insertions. (R.M. Starzyk, T.A. Webster and P. Schimmel, Science 237, 1614 (1987); R.M. Starzyk, J.J. Burbaum and P. Schimmel, Biochemistry, in press). Based on the known structure of one synthetase and structural modeling of the other, the locations of the connection sites allow the possibility of functional "compound" ATP and tRNA binding sites. Of five chimeric genes which were constructed, three direct synthesis of polypeptides that accumulate in vivo. These stable hybrids provide prototypes to which mutagenesis procedures may be applied to produce enzymatically active chimeric synthetases.