Rapid generation of incremental truncation libraries for protein engineering using α-phosphothioate nucleotides

Incremental truncation for the creation of hybrid enzymes (ITCHY) is a novel tool for the generation of combinatorial libraries of hybrid proteins independent of DNA sequence homology. We herein report a fundamentally different methodology for creating incremental truncation libraries using nucleotide triphosphate analogs. Central to the method is the polymerase catalyzed, low frequency, random incorporation of α-phosphothioate dNTPs into the region of DNA targeted for truncation. The resulting phosphothioate internucleotide linkages are resistant to 3′→5′ exonuclease hydrolysis, rendering the target DNA resistant to degradation in a subsequent exonuclease III treatment. From an experimental perspective the protocol reported here to create incremental truncation libraries is simpler and less time consuming than previous approaches by combining the two gene fragments in a single vector and eliminating additional purification steps. As proof of principle, an incremental truncation library of fusions between the N-terminal fragment of Escherichia coli glycinamide ribonucleotide formyltransferase (PurN) and the C-terminal fragment of human glycinamide ribonucleotide formyltransferase (hGART) was prepared and successfully tested for functional hybrids in an auxotrophic E.coli host strain. Multiple active hybrid enzymes were identified, including ones fused in regions of low sequence homology.     

A combinatorial approach to hybrid enzymes independent of DNA homology

We present a methodology, termed incremental truncation for the creation of hybrid enzymes (ITCHY), that creates combinatorial fusion libraries between genes in a manner that is independent of DNA homology. We compared the ability of ITCHY and DNA shuffling to create interspecies fusion libraries between fragments of the Escherichia coli and human glycinamide ribonucleotide transformylase genes, which have only 50% identity on the DNA level. Sequencing of several randomly selected positives from each library illustrated that ITCHY identified a more diverse set of active fusion points including those in regions of nonhomology and those with crossover points that diverged from the sequence alignment. Furthermore, some of the hybrids found by ITCHY that were fused at nonhomologous locations had activities that were greater than or equal to the activity of the hybrids found by DNA shuffling.     

Incremental truncation as a strategy in the engineering of novel biocatalysts

The application and success of combinatorial approaches to protein engineering problems have increased dramatically. However, current directed evolution strategies lack a combinatorial methodology for creating libraries of hybrid enzymes which lack high homology or for creating libraries of highly homologous genes with fusions at regions of non-identity. To create such hybrid enzyme libraries, we have developed a series of combinatorial approaches that utilize the incremental truncation of genes, gene fragments or gene libraries. For incremental truncation, Exonuclease III is used to create a library of all possible single base-pair deletions of a given piece of DNA. Incremental truncation libraries (ITLs) have applications in protein engineering as well as protein folding, enzyme evolution, and the chemical synthesis of proteins. In addition, we are developing a methodology of DNA shuffling which is independent of DNA sequence homology.     

Construction of hybrid gene libraries involving the circular permutation of DNA

The method of incremental truncation for the creation of hybrid enzymes (ITCHY) allows the creation of comprehensive fusion libraries between 5 and 3 fragments of two genes in a manner that is independent of DNA sequence homology. A methodology is presented for the creation of ITCHY libraries called circularly permuted ITCHY (CP-ITCHY) that allows the creation of ITCHY libraries in a manner that does not require extensive time point sampling. In addition, CP-ITCHY requires only a single vector and productively biases the library towards those fusions that are approximately the same size as the original genes. In the model system of creating fusions between fragments of the Escherichia coli and human glycinamide ribonucleotide transformylase genes, the CP-ITCHY libraries are shown to contain a diverse set of active fusions including those in regions of low-homology. In addition, a high percentage of active fusions were temperature-sensitive as they complemented an auxotrophic strain of Escherichia coli at 22 °C but not at 37 °C.     

A universal,vector-based system for nucleic acid reading-frame selection

The identification of a nucleic acid sequence's correct reading frame has important implications for homology-independent protein engineering techniques such as incremental truncation and SCRATCHY. We report the development and experimental implementation of a general in-frame selection system, pSALect, a plasmid vector that utilizes two marker sequences flanking the DNA of interest. This dual selection approach overcomes inconsistencies observed with traditional C-terminally fused reporter proteins. In the pSALect vector, sequences of interest are positioned between an N-terminal Tat-signal sequence and a C-terminal beta-lactamase reporter. In-frame selection of the resulting three-domain protein is performed by growing colonies on ampicillin-containing plates, requiring full-length translation in order to link covalently the signal sequence to the lactamase for export into the periplasm. This dual selection scheme has been validated successfully using defined sequences of glycinamide ribonucleotide formyltransferases (GARTs) from Escherichia coli and human and, in contrast to C-terminal fusion systems, proved effective when applied towards the selection of in-frame constructs in an incremental truncation library.     

Isolation of cDNA clones encoding an enzyme from bovine cells that repairs oxidative DNA damage in vitro: homology with bacterial repair enzymes.

Ionizing radiation and radiomimetic compounds, such as hydrogen peroxide and bleomycin, generate DNA strand breaks with fragmented deoxyribose 3' termini via the formation of oxygen-derived free radicals. These fragmented sugars require removal by enzymes with 3' phosphodiesterase activity before DNA synthesis can proceed. An enzyme that reactivates bleomycin-damaged DNA to a substrate for Klenow polymerase has been purified from calf thymus. The enzyme, which has a Mr of 38,000 on SDS-PAGE, also reactivates hydrogen peroxide-damaged DNA and has an associated apurinic/apyrimidinic (AP) endonuclease activity. The N-terminal amino acid sequence of the purified protein matches that reported previously for a calf thymus enzyme purified on the basis of AP endonuclease activity. Degenerate oligonucleotide primers based on this sequence were used in the polymerase chain reaction to generate from a bovine cDNA library a fragment specific for the 5' end of the coding sequence. Using this cDNA fragment as a probe, several clones containing 1.35 kb cDNA inserts were isolated and the complete nucleotide sequence of one of these determined. This revealed an 0.95 kb open reading frame which would encode a polypeptide of Mr 35,500 and with a N-terminal sequence matching that determined experimentally. The predicted amino acid sequence shows strong homology with the sequences of two bacterial enzymes that repair oxidative DNA damage, ExoA protein of S. pneumoniae and exonuclease III of E. coli.     

Isolation of a short, cytosine-rich repeating unit from the DNA of Escherichia coli

Hybridization of heterologous nucleic acids has provided the means for isolating a repeating sequence which is located next to template regions of DNA. Separated single strands of 32P-labelled DNA from Escherichia coli were to a limited extent able to anneal with DNA of Micrococcus lysodeikticus immobilized on nitrocellulose membrane filters. The resulting hybrid was resistant to enzymes specific for unpaired strands, nuclease S1 (Aspergillus oryzae) and exonuclease I (E. coli). The E. coli DNA so hybridized was isolated and characterized. It contained all four bases with cytosine predominating; strand length was about 50-60 nucleotides. Since these units occupied about 1-2% of the length of the E. coli chromosome, they would have to be repeated about 2000 times in a single cell. Formation of the unusual hybrid was not diminished by prior saturation of the E. coli DNA with homologous 3H-labelled RNA. In fact both RNA and additional increments of DNA were detected on the filters approximately in a 1:1 ratio, showing that some of the repeating sequences were physically continuous with transcribed regions of DNA.     

Identification of genes involved in the sequestration of iron in mycobacteria: the ferric exochelin biosynthetic and uptake pathways

Mycobacteria produce two siderophores, mycobactin and exochelin. Mycobacterium smegmatis mutants defective in the production of exochelin were isolated using agar medium containing chrome azural S for the sensitive detection of siderophores. Cosmids of genomic libraries from M. smegmatis and Mycobacterium bovis BCG were screened for complementation of the mutation. Subcloning of the complementing M. smegmatis cosmid identified a 4.3 kb fragment required for restoring exochelin biosynthesis. Sequencing of the DNA revealed four open reading frames whose genes were named fxuA. fxuB, fxuC, and fxbA. FxuA, FxuB, and FxuC share amino acid sequence homology with the iron permeases FepG, FepC. and FepD from Escherichia coli. respectively. Deletion analysis identified fxbA as the gene required to restore exochelin biosynthesis in our mutant. Although fxbA does not share amino acid sequence homology with any of the published sequences for siderophore biosynthetic genes, it does show limited homology with the phosphoribosyl-glycineamide formyltransferases (GAR enzymes) and methionyl-tRNA formyltransferase over a limited region of the sequence, suggesting that fxbA may code for an enzyme which adds a formyl group in the synthesis of exochelin. A fusion of fxbA with the E. coli lacZ gene demonstrated regulation of gene expression by iron. The ability of M. smegmatis mutants to produce mycobactin in the absence of exochelin supports the hypothesis that exochelin is not a precursor of mycobactin and suggests that the siderophores have independent biosynthetic pathways. In addition, complementation of the M.     

The Unexpected Catalytic Properties of a Heterodimer of GAR Transformylase

We have developed an efficient expression and purification protocol for a heterodimer of glycinamide ribonucleotide transformylase that was identified in incremental truncation libraries, a general combinatorial method for protein fragment complementation (M. Ostermeier, A. E. Nixon, J. H. Shim, and S. J. Benkovic, [1999], Proc. Natl. Acad. Sci. USA 96, 3562-3567). This heterodimer (B13) containing both a bisection point and a deletion in conserved residues close to the active site was expressed and purified in high yield using Intein methodology. The N-terminus fragment (1-111) and C-terminus fragment (M114-212) were also expressed separately as stable proteins. When these two fragments were mixed together, they associate at a highly specific 1:1 ratio to give only the active heterodimer, B13. The activity of B13 is comparable to that of the wild type and the pH-dependent kinetics of B13 turned out to be nearly identical to those of the wild type, indicating that B13 operates in the same mechanism as the wild type. This result demonstrated that cutting within conserved regions is a viable domain separation and confirmed the generality of using incremental truncation for protein fragment complementation.     

Polymerization activity of an alpha-like DNA polymerase requires a conserved 3''-5'' exonuclease active site.

Most DNA polymerases are multifunctional proteins that possess both polymerizing and exonucleolytic activities. For Escherichia coli DNA polymerase I and its relatives, polymerase and exonuclease activities reside on distinct, separable domains of the same polypeptide. The catalytic subunits of the alpha-like DNA polymerase family share regions of sequence homology with the 3'-5' exonuclease active site of DNA polymerase I; in certain alpha-like DNA polymerases, these regions of homology have been shown to be important for exonuclease activity. This finding has led to the hypothesis that alpha-like DNA polymerases also contain a distinct 3'-5' exonuclease domain. We have introduced conservative substitutions into a 3'-5' exonuclease active site homology in the gene encoding herpes simplex virus DNA polymerase, an alpha-like polymerase. Two mutants were severely impaired for viral DNA replication and polymerase activity. The mutants were not detectably affected in the ability of the polymerase to interact with its accessory protein, UL42, or to colocalize in infected cell nuclei with the major viral DNA-binding protein, ICP8, suggesting that the mutation did not exert global effects on protein folding. The results raise the possibility that there is a fundamental difference between alpha-like DNA polymerases and E. coli DNA polymerase I, with less distinction between 3'-5' exonuclease and polymerase functions in alpha-like DNA polymerases.     

Deletion analysis of sucrose metabolic genes from a Salmonella plasmid cloned in Escherichia coli K12

The sucrose operon from pUR400, a 78-kbp conjugative Salmonella plasmid, was cloned in Escherichia coli K12. The operon was located in a 5.7-kbp SalI restriction fragment and was subcloned, in each of two possible orientations, using the expression vector pUC18. The insert DNA was restriction mapped and duplicate restriction sites in the insert and in the polylinker of the vector were used to create various deletions promoter distal in the operon sequence. Additional deletions were made with the restriction exonuclease Bal31. Cells containing hybrid plasmids with specified deletions lacked the ability to transport sucrose or were constitutive for hydrolase and/or uptake activities. The scrA (enzyme IIScr) and scrR (regulatory) genes resided within 2900-bp SmaI-SalI DNA fragment and were assigned the order scrB, scrA, scrR. An amplified sucrose-inducible gene product, Mr 68,000, was detected only in the membrane fraction from recombinant cells that contained plasmid with the intact operon sequence. This protein represented 11% of the total membrane protein and was resistant to extraction with 0.5 M sodium chloride, 2% Triton X-100, and 0.5% sodium deoxycholate. The protein did not appear to be the product of either the scrA, scrB, or scrR gene and may therefore represent a previously unidentified membrane-bound sucrose protein. A new gene, scrC, is proposed. In addition, the cloned 5.7-kbp SalI and 2.5-kbp SmaI-SalI DNA fragments failed to hybridize to chromosomal DNA from Bacillus subtilis, Streptococcus lactis, Streptococcus mutans, and Lactobacillus acidophilus as well as to DNA from a sucrose plasmid from Salmonella tennessee. However, the probes showed weak homology with a 20-kbp EcoRI restriction fragment from Klebsiella pneumoniae.     

Enhanced crossover SCRATCHY: construction and high-throughput screening of a combinatorial library containing multiple non-homologous crossovers

SCRATCHY is a methodology for the construction of libraries of chimeras between genes that display low sequence homology. We have developed a strategy for library creation termed enhanced crossover SCRATCHY, that significantly increases the number of clones containing multiple crossovers. Complementary chimeric gene libraries generated by incremental truncation (ITCHY) of two distinct parental sequences are created, and are then divided into arbitrarily defined sections. The respective sections are amplified by skewed sets of primers (i.e. a combination of gene A specific forward primer and gene B specific reverse primer, etc.) allowing DNA fragments containing non-homologous crossover points to be amplified. The amplified chimeric sections are then subjected to a DNA shuffling process generating an enhanced crossover SCRATCHY library. We have constructed such a library using the rat theta 2 glutathione transferase (rGSTT2) and the human theta 1 glutathione transferase (hGSTT1) genes (63% DNA sequence identity). DNA sequencing analysis of unselected library members revealed a greater diversity than that obtained by canonical family shuffling or with conventional SCRATCHY. Expression and high-throughput flow cytometric screening of the chimeric GST library identified several chimeric progeny that retained rat-like parental substrate specificity.     

A conserved 3'----5' exonuclease active site in prokaryotic and eukaryotic DNA polymerases

The 3'----5' exonuclease active site of E. coli DNA polymerase I is predicted to be conserved for both prokaryotic and eukaryotic DNA polymerases based on amino acid sequence homology. Three amino acid regions containing the critical residues in the E. coli DNA polymerase I involved in metal binding, single-stranded DNA binding, and catalysis of the exonuclease reaction are located in the amino-terminal half and in the same linear arrangement in several prokaryotic and eukaryotic DNA polymerases. Site-directed mutagenesis at the predicted exonuclease active site of the phi 29 DNA polymerase, a model enzyme for prokaryotic and eukaryotic alpha-like DNA polymerases, specifically inactivated the 3'----5' exonuclease activity of the enzyme. These results reflect a high evolutionary conservation of this catalytic domain. Based on structural and functional data, a modular organization of enzymatic activities in prokaryotic and eukaryotic DNA polymerases is also proposed.     

On the structural and functional modularity of glycinamide ribonucleotide formyltransferases

Glycinamide ribonucleotide formyltransferases (GARTs) are part of the de novo purine biosynthetic pathway, catalyzing the direct transfer of a formyl group from the tetrahydrofolate cofactor to the glycinamide ribonucleotide substrate. Despite the low amino acid-sequence identity between the GARTs from Escherichia coli and human, their tertiary structures are superimposable. As part of our functional studies of these enzymes, we have investigated the interchangeability of individual protein fragments or modules between the two enzymes and the functional properties of the resulting hybrids. The modular nature of GART facilitated the creation of combinatorial libraries of chimeras between the Escherichia coli and human enzymes, which were functionally selected through complementation of an auxotrophic Escherichia coli strain. From a pool of several dozen sequence distinct hybrids, six in vivo-functional fusion genes were selected, overexpressed, and purified to homogeneity. The kinetic analysis of these constructs and the comparison of their k(cat) and K(M) values to the parental enzymes suggest that the characteristic kinetic properties from the two parents are "modular encoded" and can be exchanged by domain swapping. The chimeras in general, however, are subject to temperature instability and misfolding; thus, they serve primarily as useful candidates for further rounds of optimization.     

Molecular cloning and sequencing of a pectinesterase gene from Pseudomonas solanacearum

Two pectinesterase-positive Escherichia coli clones, differing in expression levels, were isolated from a genomic library of Pseudomonas solanacearum. Both clones contained a common DNA fragment which included the pectinesterase-encoding region. The different expression levels found with the two clones could be ascribed to different positioning of the pectinesterase gene with respect to a vector promoter. Restriction analysis, subcloning, and further exonuclease deletion mapping revealed that the genetic information for pectinesterase was located within a 1.3 kb fragment. A protein of 41 to 42 kDa was expressed from this fragment. Nucleotide sequence analysis of the respective region disclosed an open reading frame of 1188 bp. The deduced polypeptide had a calculated molecular mass of 41,004 Da, which is consistent with the determined size of the pectinesterase protein. The predicted amino acid sequence showed significant homology to pectinesterases from Erwinia chrysanthemi and tomato. In cultures of E. coli clones up to 30% of total pectinesterase activity was transported into the medium. However, no significant pectinesterase activity could be detected in the periplasm.     

Identification of a new motif required for the 3'-5' exonuclease activity of Escherichia coli DNA polymerase I (Klenow fragment): the RRRY motif is necessary for the binding of single-stranded DNA substrate and the template strand of the mismatched duplex

The Klenow fragment of Escherichia coli DNA polymerase I houses catalytic centers for both polymerase and 3'-5' exonuclease activities that are separated by about 35 A. Upon the incorporation of a mismatched nucleotide, the primer terminus is transferred from the polymerase site to an exonuclease site designed for excision of the mismatched nucleotides. The structural comparison of the binary complexes of DNA polymerases in the polymerase and the exonuclease modes, together with a molecular modeling of the template strand overhang in Klenow fragment, indicated its binding in the region spanning residues 821-824. Since these residues are conserved in the "A" family DNA polymerases, we have designated this region as the RRRY motif. The alanine substitution of individual amino acid residues of this motif did not change the polymerase activity; however, the 3'-5' exonuclease activity was reduced 2-29-fold, depending upon the site of mutation. The R821A and R822A/Y824A mutant enzymes showed maximum cleavage defect with single-stranded DNA, mainly due to a large decrease in the ssDNA binding affinity of these enzymes. Mismatch removal by these enzymes was only moderately affected. However, data from the exonuclease-polymerase balance assays with mismatched template-primer suggest that the mutant enzymes are defective in switching mismatched primer from the polymerase to the exonuclease site. Thus, the RRRY motif provides a binding track for substrate ssDNA and for nonsubstrate single-stranded template overhang, in a polarity-dependent manner. This binding then facilitates cleavage of the substrate at the exonuclease site.     

Inactivation of DNA polymerase I (Klenow fragment) by adenosine 2',3'-epoxide 5'-triphosphate: evidence for the formation of a tight-binding inhibitor

The suicidal inactivation of Escherichia coli DNA polymerase I by epoxy-ATP has been previously reported (Abboud et al., 1978). We have examined in detail the mechanism of this inactivation utilizing a synthetic DNA template-primer of defined sequence. Epoxy-ATP inactivates the large fragment of DNA polymerase I (the Klenow fragment) in a time- and concentration-dependent manner (KI = 21 microM; kinact = 0.021 s-1). Concomitant with inactivation is the incorporation of epoxy-AMP into the primer strand. The elongated DNA duplex directly inhibits the polymerase activity of the enzyme (no time dependence) and is resistant to degradation by the 3'----5' exonuclease and pyrophosphorylase activities of the enzyme. Inactivation of the enzyme results from slow (4 X 10(-4) s-1) dissociation of the intact epoxy-terminated template-primer from the enzyme and is thus characterized as a tight-binding inhibition. Surprisingly, while the polymerase activity of the enzyme is completely suppressed by epoxy-ATP, the 3'----5' exonuclease activity remains intact. The data presented demonstrate that even though the polymerase site is occupied with duplex DNA, the enzyme can bind a second DNA duplex and carry out exonucleolytic cleavage.     

Theoretical distribution of truncation lengths in incremental truncation libraries

Incremental truncation is a method for constructing libraries of every one base pair truncation of a segment of DNA. Incremental truncation libraries can be created using a time-dependent nuclease method or through the incorporation of alpha-phosphothioate dNTPs by PCR or by primer extension (THIO(pcr) truncation and THIO(extension) truncation, respectively). Libraries created by the fusion of two truncation libraries, known as ITCHY libraries, can be created using the above methods or by the incremental truncation-like method SHIPREC. Knowing and being able to tailor the distribution of truncations in incremental truncation, ITCHY and SHIPREC libraries would be beneficial for their use in protein engineering and other applications. However, the experimental determination of the distributions would require extensive, cost-prohibitive, DNA sequencing to obtain statistically relevant data. Instead, a theoretical prediction of the distributions was developed. Time-dependent incremental truncation libraries had the most uniform distribution of truncation lengths, but were biased against longer truncations. Essentially uniform distribution over the desired truncation range (from zero to N(max) base pairs) required that truncations be prepared up to at least 1.2-1.5 N(max). THIO(pcr) and THIO(extension) truncation libraries had a very nonuniform distribution of truncation lengths with a bias against longer truncations. Such nonuniformity could be significantly diminished by decreasing the incorporation rate of alphaS-dNTPs but at the expense of having a large fraction of the DNA truncated beyond the desired range or completely degraded. ITCHY libraries created using time-dependent truncation had the most uniform distribution of possible fusions and had the highest fraction of the library being parental-length fusions. However, the distribution of parental-length fusions was biased against fusions near the beginning/ends of genes unless the truncation libraries are prepared with a uniform distribution up to N(max). In contrast, SHIPREC libraries and THIO(pcr) ITCHY libraries, by the very nature of the nonuniform distributions of the truncated DNA, are ensured of having a uniform distribution of fusion points in parental-length fusions. This comes at the expense of having a smaller fraction of the library being parental-length fusions; however, this limitation can be overcome by performing size selection on the library.     

Chromosomal transformation of Escherichia coli recD strains with linearized plasmids.

Wild-type Escherichia coli are resistant to genetic transformation by purified linear DNA, probably in part because of exonuclease activity. We demonstrate that E. coli containing a recD mutation could be easily transformed by linearized plasmids containing a selectable marker. The marker was transferred to the chromosome by homologous recombination, whereas plasmid markers not in the region of homology were lost.