WO1991001374A1

WO1991001374A1 - Prevention of internal initiation

Info

Publication number: WO1991001374A1
Application number: PCT/US1990/004113
Authority: WO
Inventors: Frank Genbauffe
Original assignee: Seragen, Inc.
Priority date: 1989-07-24
Filing date: 1990-07-20
Publication date: 1991-02-07
Also published as: EP0484411A1; AU6067590A; NZ234616A; EP0484411A4; CA2063799A1; JPH05501054A

Abstract

Described is a method of preventing undesired initiation of translation at an internal initiator codon in a DNA sequence encoding a polypeptide, wherein a Shine-Dalgarno sequence is located upstream from the internal initiator codon. The method includes altering either the internal initiator codon or the Shine-Dalgarno sequence, or both, such that inappropriate initiation of translation and/or ribosome binding is prevented.

Description

PREVENTION OF INTERNAL INITIATION Background of the Invention This invention relates to the use of recombinant DNA techniques to maximize the expression and purity of useful polypeptides.

A critical stage in the expression of a protein is the initiation of translation. Initiation is the point at which a ribosome, a mRNA molecule, a charged tRNA, and other factors are assembled in the proper relationship to allow the insertion of the initial amino acid residue of the polypeptide to be synthesized.

In prokaryotes the initiation of translation is controlled in large part by signals encoded in the sequence of mRNA molecules. Initiation of prokaryotic translation begins with the binding of a 3OS ribosomal subunit to an mRNA molecule. A ribosomal binding site usually precedes the initiator codon of each independently synthesized polypeptide product. Thus a polycistronic prokaryotic message encoding three polypeptides will generally possess a ribosome binding site just 5' to each of its independently synthesized polypeptide products.

Only the core of the ribosomal binding site, the Shine- Dalgarno sequence, is conserved between various bacterial genes, and even the Shine-Dalgarno sequence is not invariant. A consensus Shine-Dalgarno sequence is AGGAGG. The Shine-Dalgarno sequence is complimentary to a pyrimidine-rich portion of the 16S ribosomal RNA component of the 3OS ribosomal subunit. To function optimally in initiation, the Shine-Dalgarno sequence is generally centered approximately 4 to 20 nucleotides upstream from an initiator codon. The initiator codon in bacteria is generally AUG, but may be GUG or even TJUG or ATJU. The amino terminal amino acid, the first added in the synthesis of the polypeptide, is almost invariably N-formylmethionine (fMet) . Formylation occurs after the methionine has been attached to its tRNA. The formyl group blocks the amino group of the methionine. This does not prevent the use of N-formylmethionine at the first position in the polypeptide chain but precludes its insertion at any other point,because of the inability of the blocked amino group to participate in a peptide bond. Shortly after synthesis of the polypeptide chain begins, the formyl group is usually removed from the amino terminal methionine by a deformylase. In the case of many proteins, the amino terminal methionine itself is subsequently removed by an aminopeptidase.

Binding of the 3OS ribosomal subunit at the ribosomal binding site is followed by assembly of a complete ribosome on the mRNA molecule. Once the ribosome has assembled at the Shine-Dalgarno sequence it scans the message until it encounters an initiator codon. Upon encountering the initiator codon the ribosome undergoes a conformational change and N-formylmethionine is inserted at the first position of the amino acid chain.

After initiation the ribosome proceeds codon by codon along the mRNA inserting amino acids into the growing polypeptide chain. Termination is signaled by sequences encoded in the mRNA in the form of one of three stop codons, UAA, UAG or UGA. Upon reaching an in-frame stop codon, the ribosome falls off the mRNA and addition of amino acid residues to the polypeptide chain ceases. The codons that serve as prokaryotic initiator codons are found not only at the initial position but also at other positions in the coding region. Although each of these codons directs the insertion of N-formylmethionine at the initial position, none of them normally result in the insertion of N-formylmethionine when present at other positions. When present internally, i.e., at any position other than the first or initiator position, the codons AUG, GUG, UUG, and AUU normally direct the insertion, respectively, of methionine, valine, leucine, and isoleucine.

Two classes of tRNA^Met exist: tRNA_p ^Met, which carries N-formylmethionine and tRNA^^ which carries methionine. When bound to the initiation site the ribosome takes part in a base pairing interaction between the 16S RNA component of the ribosome and the Shine-Dalgarno sequence. This interaction is thought to alter the conformation of the ribosome in such a way that the only transfer RNA that can bind to the 3OS subunit of the ribosome is tRNA_F ^Met. When that ribosome encounters an AUG, or one of the other codons that may serve as an initiator codon, N-formylmethionine is then inserted as the initial amino acid of a polypeptide. When the Shine- Dalgarno induced interaction is not present, i.e., at an internal site, tRNA_f ^Met does not bind to the ribosome. In these instances the tRNA appropriate to the encountered codon (tRNA_m ^Mβt in the case of AUG, tRNA^Val in the case of GUG, tRNA^Lβu in the case of UUG, and tRNA^Ilβ in the case of AUU) binds to the ribosome with the concomitant insertion of its charged amino acid.

Summary of the Invention The invention depends on the observation that whether one of the four codons AUG, GUG, UUG, AND AUU functions as an internal codon depends on the presence or absence of an appropriately placed Shine-Dalgarno sequence. The essential sequence requirements for the initiation of translation consists of an initiator codon downstream from a Shine-Dalgarno sequence. Thus an internal AUG,

GUG, UUG or AUU, if coincidentally preceded by sequences that can function as a Shine-Dalgarno sequence, can result in second site or internal initiation. Second site or internal initiation results in the production of a protein corresponding to the sequence between the second site start codon and the first in-frame stop codon encountered.

In general, the invention features a method of preventing such undesirable initiation of translation at an internal initiator codon which is preceded by an internal Shine-Dalgarno sequence, by effecting a change in the DNA of either the internal initiator codon or the internal Shine-Dalgarno sequence. The change preferably does not alter the amino acid sequence of the full length translation product.

The invention provides a number of methods for eliminating internal initiation. The method employed in a particular application is dictated by the nature of the nucleotide sequence of the DNA encoding the desired full length desired translation product and the amino acid sequence of the desired full length translation product. In the most preferred embodiment, internal initiation is eliminated by effecting a change in the sequence of the internal initiator codon and in the Shine-Dalgarno sequence such that the internal initiator codon is converted to a codon that does not support initiation and such that any possibility of inappropriate ribosome binding is eliminated, all without changing the amino acid sequence of the full length translation product. In a less preferred embodiment, where the sequence is such that no change in the sequence of the internal initiator codon will eliminate the presence of the internal initiator codon without changing the amino acid sequence of the full length transcription product, a change in the DNA sequence of the Shine-Dalgarno sequence is effected. This change is such that the functional Shine-Dalgarno sequence is destroyed without resulting in a change in the amino acid sequence of the full length translation product. In other less preferred embodiments, any change that will eliminate the existence of an internal initiator codon or destroy the superfluous functional Shine- Dalgarno sequence will result in a change in the sequence of the full length translation product. In these embodiments a change in the DNA sequence is effected that will eliminate internal initiation and produce a polypeptide with the most conservative departure from the original full length translation product. The invention improves yield of recombinant proteins and facilitates purification by preventing the production of non-functional truncated protein fragments beginning at internal initiation sites.

Other features and advantages of the invention will be apparent from the following description of the preferred embodiments thereof, and from the claims.

Description of the Preferred Embodiments The drawings will first be described Drawings FIG. 1 is the sequence of the DNA expressing IL-2- toxin. The internal GTG codon, starting at nucleotide 526, is indicated.

FIG. 2 is a codon dictionary. FIG. 3 is a table of conservative amino acid substitutions.

FIG. 4 is a Shine-Dalgarno sequence in different registers with the in-frame codons of a mRNA molecule. Prevention of Internal Initiation in IL-2 Toxin

IL-2-toxin is a 68,170 dalton fusion protein expressed from a hybrid gene. The recombinant gene contains both a portion of the diphtheria toxin gene and the interleukin- 2 (IL-2) gene. DNA encoding the diphtheria toxin's generalized eukaryotic receptor binding domain is replaced with interleukin-2 (IL-2) encoding DNA, using recombinant DNA methods, as described in Murphy, U.S. Patent No. 4,675,382, hereby incorporated by reference. As described in Strom et al. PTC/US89/02166, hereby incorporated by reference, IL-2-toxin can act as an IL- 2-receptor-positive-cell-destroying substance. The purification of IL-2-toxin yields a 59,000 dalton polypeptide as a contaminant. Amino acid sequence analyses of the copurifying polypeptide indicates that the N-terminal amino acid residue of the 59,000 dalton polypeptide is threonine, and thereafter the 59,000 dalton polypeptide is identical to the remaining 534 amino acid residues comprising the carboxyl terminal portion of IL-2-toxin. The 59,000 dalton polypeptide is cross-reactive with anti-diphtheria toxin antibodies and anti-interleukin-2 antibodies. The antigenic determinants of these antibodies are both present in the region of IL-2-toxin corresponding to the sequence of the 59,000 dalton internal start polypeptide.

The sequence of IL-2-toxin is shown in FIG. l. The 84th codon in the mRNA specifying IL-2-toxin is GUG, which encodes valine. (The 84th codon corresponds to nucleotides 526-528 in FIG. 1) The 85th codon in the mRNA specifying IL-2-toxin is ACG, which encodes threonine. The 59,000 dalton polypeptide contaminant is, we believe, derived from an internal initiation of translation at codon 84 in the IL-2-toxin mRNA. Internal initiation of translation occurs at the GUG codon at position 84 probably because of the presence of a highly conforming representative of the Shine-Dalgarno sequence, UGGAGG, centered 13 bases 5¹ of the GUG codon. GUG may serve as an initiator codon in bacteria. Although GUG normally specifies valine, the Shine-Dalgarno sequence 5¹ to the GUG codon at position 84 results in initiation- ribosome binding, ribosomal assembly, and concomitant insertion of N-formylmethionine. This N-formylmethionine residue, which is the first amino acid of the internal start polypeptide, is cleaved by an aminopeptidase, and thus the initial amino acid residue of the ultimately recovered internal start contaminant is threonine. This threonine corresponds to the threonine at position 85 of IL-2-toxin.

Second site initiation is undesirable for a number of reasons. Second site starts .may disrupt translation from properly initiated riboso es, as well as compete with initiation at the true initiation site for ribosomes, initiation factors, charged tRNAs, and any other factors needed for production of the final protein product. Thus second site initiators may reduce the overall yield of a desired product of protein synthesis. The product of the second site starts may also result in the need for additional purification steps. Products of internal initiation, when in the same reading frame, can be very similar e.g., in size, physical properties and immunological specificity, to the desired product. These similarities can present particularly difficult purification problems.

Second site or internal initiation can be eliminated according to the invention by effecting an appropriate alteration in the sequence of either the codon being used as the second site initiator codon or in the nearest adjacent Shine-Dalgarno sequence or in both. The most preferred alteration is one that, without altering the sequence of the desired translation product, is known unambiguously to be capable of eliminating internal initiation and ribosome binding. If only a single change can be made, changes at the internal initiator codon are thus preferable to changes at the internal Shine- Dalgarno, although changes at both sites is most preferable. Changing the third "G" in the GUG codon at position 84 to "C" yields GUC, which encodes valine, but which cannot act as an initiator codon in the presence of a Shine- Dalgarno sequence.

This change is consistent with the fact that the final nucleotide of the codons specifying most amino acids are far less specific than are the first and second nucleotides of the codon (see FIG. 2) . For example, four codons specify valine: ,GUG, GUA, GUC, and GUU. Any change at the first or second nucleotide of the GUG codon would result in a change in the amino acid inserted into the polypeptide. The final nucleotide of an in-frame GUG may however be changed to any one of A, C, or U and the resultant codon will still direct the insertion of valine into the nascent polypeptide chain. The codons GUA, GUC, and GUU never, even when proceeded by a Shine-Dalgarno sequence, function as initiator codons. Thus any substitution at the third nucleotide of the GUG codon at position 84 of IL-2-toxin will eliminate internal initiation at position 84 without effecting a change in the amino acid sequence of IL-2-toxin. Standard methods for the manipulation of cloned DNA sequences, known to those skilled in the art, can be employed to effect the desired change in DNA sequence.

In other preferred embodiments, where the codon serving as the initiator codon of an internal start is UUG or AUU that is in-frame, an analogous approach is used. UUG, which normally encodes leucine, can be changed to UUA, which also encodes leucine but which cannot function as an initiator codon, even in the presence of a Shine- Dalgarno sequence. Likewise, AUU which normally encodes isoleucine can be changed to AUC, which also encodes isoleucine but never serves as an initiator codon, even in the presence of a Shine-Dalgarno sequence.

In other preferred embodiments second site initiation occurs at an in-frame AUG codon. AUG is the only codon that directs the insertion of methionine. A change at any of the three nucleotides of the AUG codon will result in the substitution of some other amino acid for methionine in the full length polypeptide.

Second site starts at in-frame AUG codons may be eliminated in one of three ways. In a preferred embodiment, the functional integrity of the Shine- Dalgarno sequence and the internal initiator codon can be altered to eliminate internal starts. Alternatively, the internal initiator alone can be altered. The AUG serving as the second site initiator codon can be altered to a codon that does not support initiation. As mentioned above, any change in an AUG codon will be expressed in the full length polypeptide. The choice of alterations should be limited to those resulting in an amino acid whose properties most resemble those of methionine. FIG. 3 provides a table of the most conservative amino acid substitutions.

In other embodiments it may be preferable to eliminate internal starts by alteration of the functional internal Shine-Dalgarno sequence. Depending on the exact sequence of the Shine-Dalgarno sequence and the way in which it is superimposed on the in-frame codons specifying the translation product, different options will arise. It will be possible in some instances to eliminate the functional Shine-Dalgarno sequence without altering the amino acid sequence of the desired full length polypeptide. FIG. 4 depicts a Shine-Dalgarno sequence in each possible register with the in-frame codons of the corresponding polypeptides. Inspection of the dictionary of codons in FIG. 2 indicates that the Shine-Dalgarno sequence in 4a can be disrupted without a change in amino acid sequence, while the changes that can be made without changing the sequence of the amino acids in the polypeptide do not unambiguously destroy the Shine- Dalgarno sequence of figure 4c. In cases where any change, either in the internal initiator codon or the Shine-Dalgarno sequence, will result in a change in the amino acid sequence of the polypeptide reference can be made to FIG. 3 to discover the most conservative of the resulting substitutions.

In other embodiments, the internal initiator codon will be out-of-frame with the desired full length polypeptide. For example, an in-frame sequence specifying the insertion of isoleucine and cysteine could be comprised of AUA»UGU. The final nucleotide of the isoleucine codon together with the first two nucleotides of the cysteine codon form an out-of-frame AUG. In many cases it will be possible to alter the in-frame sequences in such a way that the out-of-frame initiator is destroyed without effecting a change in the amino acids inserted into the desired full length polypeptide. In the example just given, alteration to yield the in-frame sequence AUC»UGU will replace the out-of-frame AUG with CUG, which never serves as an initiator codon. The new sequence will result in the insertion of the same amino acids, isoleucine and cysteine, into the full length polypeptide. In other preferred embodiments it will not be possible to eliminate second site starts without effecting a change in the amino acid sequence of the full length polypeptide, e.g., where the in-frame sequence AAA»UGU specifying lysine and cysteine creates an out-of-frame AUG formed from the last nucleotide of the lysine codon and the first two nucleotides of the cysteine codon. The only change that can be made within the 6 nucleotides of the in-frame codons that is not expressed in the in-frame polypeptide is the change of the third nucleotide of the lysine codon (here A may be changed to G without a change in the polypeptide) . This change destroys the out-of- frame AUG initiator codon but creates an out-of-frame GUG initiator codon. In these and analogous situations, second site starts may be eliminated by changes in the in-frame sequence that eliminate the out-of-frame initiator codon but result in only the most conservative changes in the amino acids inserted into the desired full length polypeptide. The results of alterations at the internal initiator codon are compared to the results of alterations at the Shine-Dalgarno sequence. Generally, the alteration that results in no or the most conservative change in amino acid sequence is chosen. Other embodiments are within the following claims.

Claims

1. A method of preventing undesired initiation of translation at an internal initiator codon in a DNA sequence encoding a polypeptide, wherein there is located upstream of said internal initiator codon a Shine- Dalgarno sequence, said method comprising altering either said internal initiator codon or said Shine-Dalgarno sequence, or both, such that, inappropriate initiation of translation and/or ribosome binding is prevented.

2. The method of claim 1, wherein said internal initiator codon is altered by a substitution of one or more base pairs.

3. The method of claim 1, wherein said Shine-Dalgarno sequence is altered by a substitution of one or more base pairs.