CA2231394A1 - High level expression of proteins - Google Patents

Abstract

The invention features a synthetic gene encoding a protein normally expressed in a mammalian cell or eukaryotic cell wherein at least one non-preferred or less preferred codon in the natural gene encoding the mammalian protein has been replaced by a preferred codon encoding the same amino acid.

Description

HIGH LEVEL EXPRESSION OF PROTEINS
Field of the Invention 5The invention concerns genes and methods for expressing eukaryotic and viral proteins at high levels in eukaryotic cells.
Backqround of the Invention Expression of eukaryotic gene products in 10 prokaryotes is sometimes limited by the presence of codons that are infrequently used in E. coli. Expression of such genes can be enhanced by systematic substitution of the endogenous codons with codons over represented in highly expressed prokaryotic genes (Robinson et al.
15 1984). It is commonly supposed that rare codons cause pausing of the ribosome, which leads to a failure to complete the nascent polypeptide chain and a uncoupling of transcription and translation. The mRNA 3' end of the stalled ribosome is exposed to cellular ribonucleases,~0 which decreases the stability of the transcript.
s~ ~ry of the Invention The invention features a synthetic gene encoding a protein normally expressed in a mammalian cell or other eukaryotic cell wherein at least one non-preferred or 25 less preferred codon in the natural gene encoding the protein has been replaced by a preferred codon encoding the same amino acid.
Preferred codons are: Ala (gcc); Arg (cgc); Asn (aac); Asp (gac) Cys (tgc); Gln (cag); Gly (ggc); His (cac); Ile (atc); Leu (ctg); Lys (aag); Pro (ccc); Phe (ttc); Ser (agc); Thr (acc); Tyr (tac); and Val (gtg).
Less preferred codons are: Gly (ggg); Ile (att); Leu (ctc); Ser (tcc); Val (gtc). All codons which do not fit the description of preferred codons or less preferred 35 codons are non-preferred codons. In general, the degree of preference of particular codon is indicated by the prevalence of the codon in highly expressed human genes as indicated in Table 1 under the heading ~High.~ For example, ~atc~ represents 77~ of the Ile codons in highly 5 expressed ~ ~lian genes and is the preferred Ile codon;
~att~ represents 18% of the Ile codons in highly expressed mammalian genes and is the less preferred Ile codon. The sequence ~ata~ represents only 5% of the Ile codons in highly expressed human genes as is a non-10 preferred codon. Replacing a codon with another codonthat is more prevalent in highly expressed human genes will generally increase expression of the gene in mammalian cells. Accordingly, the invention includes replacing a less preferred codon with a preferred codon 15 as well as replacing a non-preferred codon with a preferred or less preferred codon.
By ~protein normally expressed in a mammalian cell~ is meant a protein which is expressed in ~ ~lian under natural conditions. The term includes genes in the 20 mammalian genome such as Factor VIII, Factor IX, interleukins, and other proteins. The term also includes genes which are expressed in a ~ ~lian cell under disease conditions such as oncogenes as well as genes which are encoded by a virus (including a retrovirus) 25 which are expressed in ~ ~lian cells post-infection.
By ~protein normally expressed in a eukaryotic cell~ is meant a protein which is expressed in a eukaryote under natural conditions. The term also includes genes which are expressed in a - ~lian cell under ~;~A~e 30 conditions such as In preferred embodiments, the synthetic gene is capable of expressing the mammalian or eukaryotic protein at a level which is at lea3t 110~, 150%, 200~, 500%, 1,000%, 5,000% or 10,000% of that expressed by said 35 natural gene in an in vitro mammalian cell culture system W O 97/11086 PCTrUS96/15088 under identical conditions (i.e., same cell type, same culture conditions, same expression vector).
Suitable cell culture systems for measuring expression of the synthetic gene and corresponding 5 natural gene are described below. Other suitable expression systems employing ~ ~lian cells are well known to those skilled in the art and are described in, for example, the stA~rd molecular biology reference works noted below. Vectors suitable for expressing the 10 synthetic and natural genes are described below and in the st~n~d reference works described below. By "expression" is meant protein expression. Expression can be measured using an antibody specific for the protein of interest. Such antibodies and measurement t~c-hn;ques are 15 well known to those skilled in the art. By "natural gene" is meant the gene sequence (including naturally occurring allelic variants) which naturally encodes the protein.
In other preferred embodiments at least 10%, 20%, 20 30%, 40%, 50%, 60%, 70%, 80%, or 90% of the codons in the natural gene are non-preferred codons.
In a preferred embodiment the protein is a retroviral protein. In a more preferred embodiment the protein is a lentiviral protein. In an even more 25 preferred embodiment the protein is an HIV protein. In other preferred embodiments the protein is gag, pol, env, gpl20, or gpl60. In other preferred embodiments the protein is a human protein.
The invention also features a method for preparing 30 a synthetic gene encoding a protein normally expressed by a ~ ~lian cell or other eukaryotic cell. The method includes identifying non-preferred and less-preferred codons in the natural gene encoding the protein and replacing one or more of the non-preferred and less-O 97/11086 PCT~US96/15088 preferred codons with a preferred codon encoding the same amino acid as the replaced codon.
Under some circumstances (e.g., to permit introduction of a restriction site) it may be desirable to replace a non-preferred codon with a less preferred codon rather than a preferred codon.
It is not necessary to replace all less preferred or non-preferred codons with preferred codons. Increased expression can be accomplished even with partial replacement. Under some circumstances it may be desirable to only partially replace non-preferred codons with preferred or less preferred codons in order to obtain an intermediate level of expression.
In other preferred embodiments the invention features vectors (including expression vectors) comprising one or more the synthetic genes.
By "vector" is meant a DNA molecule, derived, e.g., from a plasmid, bacteriophage, or ~ An or insect virus, into which fragments of DNA may be inserted or cloned. A vector will contain one or more unique restriction sites and may be capable of autonomous replication in a defined host or vehicle organism such that the cloned sequence is reproducible. Thus, by "expression vector" is meant any autonomous element capable of directing the synthesis of a protein. Such DNA expression vectors include mammalian plasmids and viruses.
The invention also features synthetic gene fragments which encode a desired portion of the protein.
Such synthetic gene fragments are similar to the synthetic genes of the invention except that they encode only a portion of the protein. Such gene fragments preferably encode at least 50, 100, 150, or 500 contiguous amino acids of the protein.
-In constructing the synthetic genes of the invention it may be desirable to avoid CpG sequences as these sequences may cause gene silencing.
The codon bias present in the HIV gpl20 envelope 5 gene is also present in the gag and pol proteins. Thus, replacement of a portion of the non-preferred and less preferred codons found in these genes with preferred codons should produce a gene capable of higher level expression. A large fraction of the codons in the human 10 genes encoding Factor VIII and Factor IX are non-preferred codons or less preferred codons. Replacement of a portion of these codons with preferred codons should yield genes capable of higher level expression in mammalian cell culture.
The synthetic genes of the invention can be introduced into the cells of a living organism. For example, vectors (viral or non-viral) can be used to introduce a synthetic gene into cells of a living org~n;s~ for gene therapy.
Conversely, it may be desirable to replace preferred codons in a naturally occurring gene with less-preferred codons as a ~~n~ of lowering expression.
St~n~rd reference works describing the general principles of recombinant DNA technology include Watson, 25 J.D. et al., Molecular Biolo~Y of the Gene, Volumes I and II, the Benjamin/C ;nqs Publishing Company, Inc., publisher, Menlo Park, CA (1987); Darnell, J.E. et al., Molecular Cell Biolo~v, Scientific American Books, Inc., Publisher, New York, N.Y. (1986); Old, R.W., et al., 30 Princi~les of Gene Manipulation: An Introduction to Çenetic Enqineering, 2d edition, University of California Press, publisher, Berkeley, CA (1981); Maniatis, T., et al., Molecular Cloning: A Laboratory Manual, 2nd Ed.
Cold Spring Harbor Laboratory, publisher, Cold Spring 35 Harbor, NY (1989); and Current Protocols in Molecular O 97/11086 PCTnUS96/15088 ~;oloqY, Ausubel et al., Wiley Press, New York, NY
(1992).
Detailed Description Description of the Drawinqs Figure 1 depicts the sequence of the synthetic gpl20 and a synthetic gpl60 gene in which codons have been replaced by those found in highly expressed human genes.
Figure 2 is a schematic drawing of the synthetic gpl20 (HIV l MN) gene. The shaded portions marked vl to v5 indicate hypervariable regions. The filled box indicates the CD4 binding site. A limited number of the unique restriction sites ares shown: H (Hind3), Nh (Nhel), P (Pstl), Na (Nael), M (Mlul), R (EcoRl), A
(Agel) and No (Notl). The chemically synthesized DNA
fragments which served as PCR templates are shown below the gpl20 sequence, along with the locations of the primers used for their amplification.
Figure 3 is a photograph of the results of transient transfection assays used to measure gpl20 expression. Gel electrophoresis of ; noprecipitated supernatants of 293T cells transfected with plasmids expressing gpl20 encoded by the IIIB isolate of HIV-l (gpl20IIIb), by the MN isolate (gpl20mn), by the MN
isolate modified by substitution of the endogenous leader peptide with that of the CD5 antigen (gpl20mnCD5L), or by the chemically synthesized gene en~oA;ng the MN variant with the human CD5Leader (syngpl2Omn). Supernatants were harvested following a 12 hour labeling period 60 hours post-transfection and immunoprecipitated with CD4:IgGl ~usion protein and protein A ~ephArose.
Figure 4 is a graph depicting the results of ELISA
assays used to measure protein levels in supernatants of transiently transfected 293T cells. Supernatants of 293T
cells transfected with plasmids expressing gpl20 encoded W O 97/11086 PCT~US96/15088 by the IIIB isolate of HIV-1 (gpl20 IIIb), by the MN
isolate (gpl20mn), by the MN isolate modified by substitution of the endogenous leader peptide with that of CD5 antigen (gpl20mn CD5L), or by the chemically 5 synthesized gene encoding the MN variant with human CDS
leader (syngpl20mn) were harvested after 4 days and tested in a gpl20/CD4 ELISA. The level of gpl20 is expressed in ng/ml.
Figure 5, panel A is a photograph of a gel 10 illustrating the results of a immunoprecipitation assay used to measure expression of the native and synthetic gpl20 in the presence of rev in trans and the RRE in cis.
In this experiment 293T cells were transiently transfected by calcium phosphate coprecipitation of 10 ~g 15 of plasmid expressing: (A) the synthetic gpl20MN sequence and RRE in cis, (B) the gpl20 portion of HIV-l IIIB, (C) the gpl20 portion of HIV-l IIIB and RRE in cis, all in the presence or absence of rev expression. The RRE
constructs gpl20IIIbRRE and syngpl20 - RRF~ were generated 20 using an Eagl/Hpal RRE fragment cloned by PCR from a HIV-l HXB2 proviral clone. Each gpl20 expression plasmid was cotransfected with 10 ~g of either pCMVrev or CDM7 plasmid DNA. Supernatants were harvested 60 hours post transfection, immunoprecipitated with CD4:IgG fusion 25 protein and protein A agarose, and run on a 7% reducing SDS-PAGE. The gel exposure time was extended to allow the induction of gpl2OIIIbrre by rev to be demonstrated.
Figure 5, panel B is a shorter exposure of a similar experiment in which syngpl2Omnrre was 30 cotransfected with or without pCMVrev. Figure 5, panel C
is a schematic diagram of the constructs used in panel A.
Figure 6 is a comparison of the sequence of the wild-type rat THY-l gene (wt) and a synthetic rat THY-1 gene (env) constructed by chemical synthesis and having 35 the most prevalent codons found in the HIV-1 env gene.

WO 97/11086 PCT~US96/15088 Figure 7 is a schematic diagram of the synthetic ratTHY-l gene. The solid black box denotes the signal peptide. The shaded box denotes the sequences in the precursor which direct the attachment of a phophatidyl-5 inositol glycan anchor. Unique restriction sites usedfor assembly of the THY-1 constructs are marked H
(Hind3), M (Mlul), S (Sacl) and No (Notl). The position of the synthetic oligonucleotides employed in the construction are shown at the bottom of the figure.
Figure 8 is a graph depicting the results of flow cytometry analysis. In this experiment 293T cells transiently transfected with either wild-type rat THY-1 (dark line), ratTHY-1 with envelope codons (light line) or vector only (dotted line). 293T cells were 15 transfected with the different expression plasmids by calcium phosphate coprecipitation and stained with anti-ratTHY-l monoclonal antibody OX7 followed by a polyclonal FITC- conjugated anti-mouse IgG antibody 3 days after transfection.
Figure 9, panel A is a photograph of a gel illustrating the results of immunoprecipitation analysis of supernatants of human 293T cells transfected with either syngpl2Omn (A) or a construct syngpl2Omn.rTHY-lenv which has the rTHY-lenv gene in the 3' untranslated 25 region of the syngpl20mn gene (B). The syngpl20mn.rTHY-lenv construct was generated by inserting a Notl adapter into the blunted Hind3 site of the rTHY-lenv plasmid. Subsequently, a 0.5 kb Notl fragment cont~; n; ng the rTHY-lenv gene was cloned into the Notl 30 site of the syngpl20mn plasmid and tested for correct orientation. Supernatants of 35S labeled cells were harvested 72 hours post transfection, precipitated with CD4:IgG fusion protein and protein A agarose, and run on a 7% reducing SDS-PAGE.

W O 97/11086 PCTrUS96/15088 g Figure 9, panel B is a schematic diagram of the constructs used in the experiment depicted in panel A of FIG. 9.
Figure 10, panel A is a photograph of COS cells 5 transfected with vector only showing no GFP fluorescence.
Figure 10, panel B i a photograph of COS cells transfected with a CDM7 expression plasmid encoding native GFP engineered to include a consensus translational initiation sequence. Figure 10, panel C is 10 a photograph of COS cells transfected with an expression plasmid having the same flanking sequences and initiation consensus as in FIG. 10, panel B, but bearing a codon optimized gene sequence. Figure 10, panel D is a photograph of COS ce]ls transfected with an expression 15 plasmid as in FIG. 10, panel C, but bearing a Thr at residue 65 in place of Ser.
Descri~tion of the Preferred Embodiments Construction of a Synthetic q~120 Gene Having Codons Found in HiqhlY Expressed Human Genes A codon frequency table for the envelope precursor of the LAV subtype of HIV-l was generated using software developed by the University of Wisconsin Genetics Computer Group. The results of that tabulation are contrasted in Table 1 with the pattern of codon usage by 25 a collection of highly expressed human genes. For any amino acid encoded by degenerate codons, the most favored codon of the highly expressed genes is different from the most favored codon of the HIV envelope precursor.
Moreover a simple rule describes the pattern of favored 30 envelope codons wherever it applies: preferred codons maximize the number of adenine residues in the viral RNA. In all cases but one this means that the codon in which the third position is A is the most frequently used. In the special case of 35 serine, three codons equally contribute one A residue to W O 97/11086 PCTrUS96/15088 the mRNA; together these three comprise 85% of the serine codons actually used in envelope transcripts. A
particularly striking example of the A bias is found in the codon choice for arginine, in which the AGA triplet 5 comprises 88% of the arginine codons. In addition to the preponderance of A residues, a marked preference is seen for uridine among degenerate codons whose third residue must be a pyrimidine. Finally, the inconsistencies among the less frequently used variants can be accounted for by 10 the observation that the dinucleotide CpG is under represented; thus the third position is less likely to be G whenever the second position is C, as in the codons for alanine, proline, serine and threonine; and the CGX
triplets for arginine are hardly used at all.

CA 0223l394 l998-03-09 W O 97/11086 PCT~US96/15088 TABLE 1: Codon FreauencY in the HIV-l IIIb env qene and in hi~hlY exPressed human qenes.
~igh ~nv High Env Ala ~5 G 17 5 Gln Arq G 88 45 T 7 4 Glu G 18 8 Glv Asn T 12 13 A ~is Ile Leu Ser Lvs Thr Pro ~vr Phe Val Codon frequency was calculated using the GCG program established the University of Wisconsin Genetics Computer Group. Numbers represent the percentage of cases in which the particular codon i~ used. Codon usage frequencies of envelope genes of other HIV-l virus isolaten are comparable and show a similar bias.

CA 02231394 l99X-03-09 In order to produce a gpl20 gene capable of high level expression in ~m~1 ian cells, a synthetic gene encoding the gpl20 segment of HIV-1 was constructed (syngpl20mn), based on the sequence of the most common 5 North American subtype, HIV-l MN (Shaw et al., Science 226:1165, 1984; Gallo et al., Nature 321:119, 1986). In this synthetic gpl20 gene nearly all of the native codons have been systematically replaced with codons most frequently used in highly expressed human genes (FIG. 1).
10 This synthetic gene was assembled from chemically synthesized oligonucleotides of 150 to 200 bases in length. If oligonucleotides exceeding 120 to 150 bases are chemically synthesized, the percentage of full-length product can be low, and the vast excess of material 15 consists of shorter oligonucleotides. Since these shorter fragments inhibit cloning and PCR procedures, it can be very difficult to use oligonucleotides exc~ing a certain length. In order to use crude synthesis material without prior purification, single-stranded 20 oligonucleotide pools were PCR amplified before cloning.
PCR products were purified in agarose gels and used as templates in the next PCR step. Two adjacent fragments could be co-amplified because of overlapping sequences at the end of either fragment. These fragments, which were 25 between 350 and 400 bp in size, were subcloned into a pCDM7-derived plasmid containing the leader sequence of the CD5 surface molecule followed by a Nhel/Pstl/Mlul/EcoR1/BamH1 polylinker. Each of the restriction enzymes in this polylinker represents a site 30 that i8 present at either the 5' or 3' end of the PCR--generated fragments. Thus, by sequential subcloning of each of the 4 long fragments, the whole gpl20 gene was assembled. For each fragment 3 to 6 different clones were subcloned and sequenced prior to assembly. A schematic 35 drawing of the method used to construct the synthetic W O 97/11086 PCTrUS96/15088 gpl20 is shown in FIG. 2. The sequence of the synthetic gpl20 gene (and a synthetic gpl60 gene created using the same approach) is presented in FIG. 1.
The mutation rate was considerable. The most 5 commonly found mutations were short (1 nucleotide) and long (up to 30 nucleotides) deletions. In some cases it was necessary to exchange parts with either synthetic adapters or pieces from other subclones without mutation in that particular region. Some deviations from strict 10 adherence to optimized codon usage were made to accommodate the introduction of restriction sites into the resulting gene to facilitate the replacement of various segments (FIG. 2). These unique restriction sites were introduced into the gene at approximately 100 bp 15 intervals. The native HIV leader sequence was exchanged with the highly efficient leader peptide of the human CD5 antigen to facilitate secretion (Aruffo et al., Cell.61:1303, 1990) The plasmid used for construction is a derivative of the ~ ~lian expression vector pCDM7 20 transcribing the inserted gene under the control of a strong human CMV ; o~;ate early promoter.
To c- -~e the wild-type and synthetic gpl20 coding sequences, the synthetic gpl20 coding sequence was inserted into a ~ ~lian expression vector and tested in 25 transient transfection assays. Several different native gpl20 genes were used as controls to exclude variations in expression levels between different virus isolates and artifacts induced by distinct leader sequences. The gpl20 HIV IIIb construct used as control was generated by 30 PCR using a Sall/Xhol HIV-l HXB2 envelope fragment as template. To exclude PCR induced mutations, a Kpnl/Earl fragment containing approximately 1.2 kb of the gene was exchanged with the respective sequence from the proviral clone. The wild-type gpl20mn constructs used as controls 35 were cloned by PCR from HIV-1 MN infected C8166 cells W O 97/11086 PCT~US96/15088 (AIDS Repository, Rockville, MD) and expressed gpl20 either with a native envelope or a CD5 leader sequence.
Since proviral clones were not available in this case, two clones of each construct were tested to avoid PCR
5 artifacts. To determine the amount of secreted gpl20 semi-quantitatively supernatants of 293T cells transiently transfected by calcium phosphate co-precipitation were immunoprecipitated with soluble CD4:immunoglobulin fusion protein and protein A
10 sepharose.
The results of this analysis (FIG. 3) show that the synthetic gene product is expressed at a very high level compared to that of the native gpl20 controls. The molecular weight of the synthetic gpl20 gene was 15 comparable to control proteins (FIG. 3) and appeared to be in the range of 100 to 110 kd. The slightly faster migration can be explained by the fact that in some tumor cell lines like 293T glycosylation is either not complete or altered to some extent.
To compare expression more accurately gpl20 protein levels were quantitated using a gpl20 ELISA with CD4 in the demobilized phase. This analysis shows (FIG.
4) that ELISA data were comparable to the immunoprecipitation data, with a gpl20 concentration of 25 approximately 125 ng/ml for the synthetic gpl20 gene, and less than the background cutoff (5 ng/ml) for all the native gpl20 genes. Thus, expression of the synthetic gpl20 gene appears to be at least one order of magnitude higher than wild-type gpl20 genes. In the experiment 30 shown the increase was at least 25 fold.
The Role of rev in gpl20 Expression Since rev appears to exert its effect at several steps in the expression of a viral transcript, the possible role of non-translational effects in the 35 improved expression of the synthetic gpl20 gene was W O 97/11086 PCT~US96/15088 tested. First, to rule out the possibility that negative signals elements conferring either increased mRNA
degradation or nucleic retention were eliminated by changing the nucleotide sequence, cytoplasmic mRNA levels 5 were tested. Cytoplasmic RNA was prepared by NP40 lysis of transiently transfected 293T cells and subsequent elimination of the nuclei by centrifugation. Cytoplasmic RNA was subsequently prepared from lysates by multiple phenol extractions and precipitation, spotted on 10 nitrocellulose using a slot blot apparatus, and finally hybridized with an envelope-specific probe.
Briefly, cytoplasmic mRNA 293 cells transfected with CDM&, gpl20 IIIB, or syngpl20 was isolated 36 hours post transfection. Cytoplasmic RNA of Hela cells 15 infected with wild-type vaccinia virus or recombinant virus expressing gpl20 IIIb or the synthetic gpl20 gene was under the control of the 7.5 promoter was isolated 16 hours post infection. Equal amounts were spotted on nitrocellulose using a slot blot device and hybridized 20 with randomly labeled 1.5 kb gpl20IIIb and syngpl20 fragments or human beta-actin. RNA expression levels were quantitated by scAnn;ng the hybridized membranes with a phospoimager. The procedures used are described in greater detail below.
This experiment demonstrated that there was no significant difference in the mRNA levels of cells transfected with either the native or synthetic gpl20 gene. In fact, in some experiments cytoplasmic mRNA
level of the synthetic gpl20 gene was even lower than 30 that of the native gpl20 gene.
These data were confirmed by measuring expression from recombinant vaccinia viruses. Human 293 cells or Hela cells were infected with vaccinia virus expressing wild-type gpl20 IIIb or syngpl20mn at a multiplicity of 35 infection of at least 10. Supernatants were harvested 24 hours post infection and immunoprecipitated with CD4:immunoglobin fusion protein and protein A sepharose.
The procedures used in this experiment are described in greater detail below.
This experiment showed that the increased expression of the synthetic gene was still observed when the endogenous gene product and the synthetic gene product were expressed from vaccinia virus recombinants under the control of the strong mixed early and late 7.5k 10 promoter. Because vaccinia virus mRNAs are transcribed and translated in the cytoplasm, increased expression of the synthetic envelope gene in this experiment cannot be attributed to improved export from the nucleus. This experiment was repeated in two additional human cell 15 types, the kidney cancer cell line 293 and HeLa cells.
As with transfected 293T cells, mRNA levels were s; il~
in 293 cells infected with either recombinant vaccinia virus.
Codon Usaqe in Lentivirus Because it appears that codon usage has a significant impact on expression in mammalian cells, the codon frequency in the envelope genes of other retroviruses was ~Y~ ;ned This study found no clear pattern of codon preference between retroviruses in 25 general. However, if viruses from the lentivirus genus, to which HIV-l belongs to, were analyzed separately, codon usage bias almost identical to that of HIV-l was found. A codon frequency table from the envelope glycoproteins of a variety of (predo~;n~ntly type C) 30 retroviruses excluding the lentiviruses was prepared, and compared a codon frequency table created from the envelope sequences of four lentiviruses not closely related to HIV-1 (caprine arthritis encephalitis virus, equine infectious anemia virus, feline ; odeficiency 35 virus, and visna virus) (Table 2). The codon usage W O 97/11086 PCTrUS96/15088 pattern for lentiviruses is strikingly similar to that of HIV-1, in all cases but one, the preferred codon for HIV-1 is the same as the preferred codon for the other lentiviruses. The exception is proline, which is encoded ~ 5 by CCT in 41~ of non-HIV lentiviral envelope residues, and by CCA in 40% of residues, a situation which clearly also reflects a significant preference for the triplet ending in A. The pattern of codon usage by the non-lentiviral envelope proteins does not show a similar 10 pred~_ ;nAnce of A residues, and is also not as skewed toward third position C and G residues as is the codon usage for the highly expressed human genes. In general non-lentiviral retroviruses appear to exploit the different codons more equally, a pattern they share with 15 less highly expressed human genes.

CA 0223l394 l998-03-09 WO 97/11086 PCT~US96/15088 TABLE 2: Codon freA~uencv in the enveloA~e ~ene of lentiviruse~
(lenti) and non-lentiviral retroviru~e~ (other).
Other Lenti Other Lonti Ala Cv~

G 9 3 Gln Ar~ G48 31 T 6 3 91~

G 15 26 Gly A~n T13 9 A~ ~is Ile Lou ~cr Lv~ ~hr Pro G10 8 T 30 41 ~yr Val Codon frequency was calculated using the GCG ~ oy~ established by the University of Wiscon~in Genetics Computer Group. N~"~eL~
represent the percentage in which a particular codon is used. Codon usage of non-lentiviral retroviru~es was compiled from the envelope precursor se~uences of bovine le~7kA~;~A~ virus feline le~ virus, human T-cell Ie~A~i A virus type I, human T-cell lymphotropic viru~
type II, the mink cell focus-forming i~olate of murine le~kA~i~ virus (MuLV), the Rauscher spleen focus-forming isolate, the 10A1 isolate, the 4070A amphotropic isolate and the myeloproliferative le~

CA 0223l394 l998-03-09 W O 97/1l086 PCT~US96/l5088 ~ 19 -virus i~olate, and from rat leukemia virus, Qimian ~arcoma virus, simian T-cell le~lk~mi~ viru~, lel~k -,genic retrovirus T1223/B and gibbon ape le~ virus. The codon frequency tables for the non-HIV, non-SIV lentiviruQes were compiled from the envelope precursor seguence~ for caprine arthriti~ encephalitis viru~, equine infectious anemia virus, feline ~ noA~ficiency viruR, and visna virus.

In addition to the prevalence of A containing codons, lentiviral codons adhere to the HIV pattern of strong CpG under representation, so that the third position for alanine, proline, serine and threonine 5 triplets is rarely G. The retroviral envelope triplets show a similar, but less pronounced, under representation of CpG. The most obvious difference between lentiviruses and other retroviruses with respect to CpG prevalence lies in the usage of the CGX variant of arginine 10 triplets, which is reasonably frequently represented among the retroviral envelope coding sequences, but is almost never present among the comparable lentivirus sequences.
Differences in rev Dependence Between Native and 15 Synthetic gpl20 To ~ ;ne whether regulation by rev is connected to HIV-1 codon usage, the influence of rev on the expression of both native and synthetic gene was investigated. Since regulation by rev requires the rev-20 binding site RRE in cis, constructs were made in whichthis binding site was cloned into the 3' untranslated region of both the native and the synthetic gene. These plasmids were co-transfected with rev or a control plasmid in trans into 293T cells, and gpl20 expression 25 levels in supernatants were measured semiquantitatively by immunoprecipitation. The procedures used in this experiment are described in greater detail below.
As shown in FIG. 5, panel A and FIG. 5, panel B, rev up regulates the native gpl20 gene, but has no effect 30 on the expression of the synthetic gpl20 gene. Thus, the WO 97/11086 PCT~US96/15088 action of rev is not apparent on a substrate which lacks the coding sequence of endogenous viral envelope sequences.
Expression of a svnthetic rat THY-1 qene with HIV
envelope codons The above-described experiment suggest that in fact "envelope sequences" have to be present for rev regulation. In order to test this hypothesis, a synthetic version of the gene encoding the small, typically highly expressed cell surface protein, rat THY-1 antigen, was prepared. The synthetic version of the rat THY-1 gene was designed to have a codon usage like that of HIV gpl20. In designing this synthetic gene AUUUA sequences, which are associated with mRNA
instability, were avoided. In addition, two restriction sites were introduced to simplify manipulation of the resulting gene (FIG 6). This synthetic gene with the HIV
envelope codon usage (rTHY-lenv) was generated using three 150 to 170 mer oligonucleotides (FIG. 7). In contrast to the syngpl2Omn gene, PCR products were directly cloned and assembled in pUC12, and subsequently cloned into pCDM7.
Expression levels of native rTHY-1 and rTHY-1 with the HIV envelope codons were quantitated by immunofluorescence of transiently transfected 293T cells.
FIG 8 shows that the expression of the native THY-1 gene is almost two orders of magnitude above the background level of the control transfected cells (pCDM7). In contrast, expression of the synthetic rat THY-1 is substantially lower than that of the native gene (shown by the shift to of the peak towards a lower channel number).
To prove that no negative sequence elements promoting mRNA degradation were inadvertently intro~llc~, a construct was generated in which the rTHY-lenv gene was W O 97/11086 PCT~US96/15088 cloned at the 3' end of the synthetic gpl20 gene (FIG. 9, panel B). In this experiment 293T cells were transfected with either the syngpl20mn gene or the syngpl20/rat THY-1 env fusion gene (syngpl2Omn.rTHY-lenv). Expression was 5 measured by immunoprecipitation with CD4:IgG fusion protein and protein A agarose. The procedures used in this experiment are described in greater detail below.
Since the synthetic gpl20 gene has an UAG stop codon, rTHY-lenv is not translated from this transcript.
10 If negative elements conferring enhanced degradation were present in the sequence, gpl20 protein levels expressed from this construct should be decreased in comparison to the syngpl2Omn construct without rTHY-lenv. FIG. 9, panel A, shows that the expression of both constructs is 15 similar, indicating that the low expression must be linked to translation.
Rev-de~endent ex~ression of svnthetic rat THY-l qene with envelo~e codons To explore whether rev is able to regulate 20 expression of a rat THY-l gene having env codons, a construct was made with a rev-binding site in the 3' end of the rTHYlenv open reading frame. To measure rev-responsiveness of the a rat THY-lenv construct having a 3' RRE, human 293T cells were cotransfected 25 ratTHY-lenvrre and either CDM7 or pCMVrev. At 60 hours post transfection cells were detached with 1 mM EDTA in PBS and stained with the OX-7 anti rTHY-l mouse monoclonal antibody and a secondary FITC-conjugated antibody. Fluorescence intensity was measured using a 30 EPICS XL cytofluorometer. These procedures are described in greater detail below.
In repeated experiments, a slight increase of rTHY-lenv expression was detected if rev was cotransfected with the rTHY-lenv gene. To further 35 increase the sensitivity of the assay system a construct expressing a secreted version of rTHY-lenv was generated.
This construct should produce more reliable data because the accumulated amount of secreted protein in the supernatant reflects the result of protein production 5 over an extended period, in contrast to surface expressed protein, which appears to more closely reflect the current production rate. A gene capable of expressing a secreted form was prepared by PCR using forward and reverse primers annealing 3' of the endogenous leader 10 sequence and 5' of the sequence motif required for phosphatidylinositol glycan anchorage respectively. The PCR product was cloned into a plasmid which already contained a CD5 leader sequence, thus generating a construct in which the membrane anchor has been deleted 15 and the leader sequence exchanged by a heterologous (and probably more efficient) leader peptide.
The rev-responsiveness of the secreted form ratTHY-lenv was measured by immunoprecipitation of supernatants of human 293T cells cotransfected with a plasmid expressing a secreted form of ratTHY--lenv and the RRE sequence in cis (rTHY-lenvPI-rre) and either CDM7 or pCMVrev. The rTHY-lenvPI-RRE construct was made by PCR
using the oligonucleotide:
cgcggggctagcgcaaagagtaataagtttaac (SEQ ID NO:38) as a 25 forward primer, the oligonucleotide:
cgcggatcccttgtattttgtactaata (S~Q ID NO:39) as reverse primer, and the synthetic rTHY-lenv construct as template. After digestion with Nhel and Notl the PCR
fragment was cloned into a plasmid contA;n;ng CD5 leader 30 and RRE sequences. Supernatants of 35S labeled cells were harvested 72 hours post transfection, precipitated with a mouse monoclonal antibody OX7 against rTHY-1 and anti mouse IgG sepharose, and run on a 12% reducing SDS-PAGE.

W O 97/11086 PCT~US96/15088 In this experiment the induction of rTHY-lenv by rev was much more prominent and clear-cut than in the above-described experiment and strongly suggests that rev is able to translationally regulate transcripts that are 5 suppressed by low-usage codons.
Rev-independent ex~ression of a rTHY-lenv:immunoglobulin fusion protein To test whether low-usage codons must be present throughout the whole coding sequence or whether a short 10 region is sufficient to confer rev-responsiveness, a rTHY-lenv:immunoglobulin fusion protein was generated.
In this construct the rTHY-lenv gene (without the sequence motif responsible for phosphatidylinositol glycan anchorage) is linked to the human IgG1 hinge, CH2 15 and CH3 domains. This construct was generated by anchor PCR using primers with Nhel and BamHI restriction sites and rTHY-lenv as template. The PCR fragment was cloned into a plasmid containing the leader sequence of the CD5 surface molecule and the hinge, CH2 and CH3 parts of 20 human IgGl immunoglobulin. A Hind3/Eagl fragment contA;n;ng the rTHY-lenvegl insert was subsequently cloned into a pCDM7-derived plasmid with the RRE
sequence.
To measure the response of the rTHY-lenv/
25 immunoglobin fusion gene (rTHY-lenveglrre) to rev human 293T cells cotransfected with rTHY-lenveglrre and either pCDM7 or pCMVrev. The rTHY-lenveglrre construct was made by anchor PCR using forward and reverse primers with Nhel and BamHl restriction sites respectively. The PCR
30 fragment was cloned into a plasmid containing a CD5 leader and human IgGl hinge, CH2 and CH3 domains.
Supernatants of 35S labeled cells were harvested 72 hours post transfection, precipitated with a mouse monoclonal antibody OX7 against rTHY-1 and anti mouse IgG sepharose, 35 and run on a 12% reducing SDS-PAGE. The procedures used are described in greater detail below.

W O 97/11086 PCT~US96/15088 As with the product o~ the rTHY-lenvPI- gene, this rTHY-lenv/immunoglobulin fusion protein is secreted into the supernatant. Thus, this gene should be responsive to rev-induction. However, in contrast to rTHY-lenvPI-, 5 cotransfection of rev in trans induced no or only a negligible increase of rTHY-lenvegl expression.
The expression of rTHY-l:immunoglobulin fusion protein with native rTHY-1 or HIV envelope codons was measured by immunoprecipitation. Briefly, human 293T
10 cells transfected with either rTHY-lenvegl (env codons) or rTHY-lwtegl (native codons). The rTHY-lwtegl construct was generated in manner similar to that used for the rTHY-lenvegl construct, with the exception that a plasmid containing the native rTHY-1 gene was used as 15 template. Supernatants of 35S labeled cells were harvested 72 hours post transfection, precipitated with a mouse monoclonal antibody OX7 against rTHY-l and anti mouse IgG sepharose, and run on a 12~ reducing SDS-PAGE.
THE procedures used in this experiment are described in 20 greater detail below.
Expression levels of rTHY-lenvegl were decreased in comrA~i~on to a similar construct with wild-type rTHY-1 as the fusion partner, but were still considerably higher than rTHY-lenv. Accordingly, both parts of the 25 fusion protein influenced expression levels. The addition of rTHY-lenv did not restrict expression to an equal level as seen for rTHY-lenv alone. Thus, regulation by rev appears to be ineffective if protein expression is not almost completely suppressed.
30 Codon preference in HIV-1 enveloPe qenes Direct c-_ A~ison between codon usage frequency of HIV envelope and highly expressed human genes reveals a striking difference for all twenty amino acids. one simple measure of the statistical significance of this 35 codon preference is the finding that among the nine amino W O 97/11086 PCT~US96/15088 acids with two fold codon degeneracy, the favored third residue is A or U in all nine. The probability that all - nine of two equiprobable choices will be the same is approximately 0.004, and hence by any conventional 5 measure the third residue choice cannot be considered random. Further evidence of a skewed codon preference is found among the more degenerate codons, where a strong selection for triplets bearing adenine can be seen. This contrasts with the pattern for highly expressed genes, 10 which favor codons bearing C, or less commonly G, in the third position of codons with three or more fold degeneracy.
The systematic exchange of native codons with codons of highly expressed human genes dramatically 15 increased expression of gpl20. A quantitative analysis by ELISA showed that expression of the synthetic gene was at least 25 fold higher in comparison to native gpl20 after transient transfection into human 293 cells. The concentration levels in the ELISA experiment shown were 20 rather low. Since an ELISA was used for quantification which is based on gpl20 binding to CD4, only native, non-denatured material was detected. This may explain the apparent low expression. Measurement of cytoplasmic mRNA
levels demonstrated that the difference in protein 25 expression is due to translational differences and not mRNA stability.
Retroviruses in general do not show a similar preference towards A and T as found for HIV. But if this family was divided into two subgroups, lentiviruses and 30 non-lentiviral retroviruses, a similar preference to A
and, less frequently, T, was detected at the third codon position for lentiviruses. Thus, the availing evidence suggests that lentiviruses retain a characteristic pattern of envelope codons not because of an inherent 35 advantage to the reverse transcription or replication of WO 97/110~6 PCTAUS96/15088 such residues, but rather for some reason peculiar to the physiology of that class of viruses. The major difference between lentiviruses and non-complex retroviruses are additional regulatory and non-5 essentially accessory genes in lentiviruses, as alreadymentioned. Thus, one simple explanation for the restriction of envelope expression might be that an important regulatory ~ch~n;~m of one of these additional molecules is based on it. In fact, it is known that one 10 of these proteins, rev, which most likely has homologues in all lentiviruses. Thus codon usage in viral mRNA is used to create a class of transcripts which is susceptible to the stimulatory action of rev. This hypothesis was proved using a similar strategy as above, 15 but this time codon usage was changed into the inverse direction. Codon usage of a highly expressed cellular gene was substituted with the most fre~uently used codons in the HIV envelope. As assumed, expression levels were considerably lower in comparison to the native molecule, 20 almost two orders of magnitude when analyzed by immunofluorescence of the surface expressed molecule (see 4.7). If rev was coexpressed in trans and a RRE element was present in cis only a slight induction was found for the surface molecule. However, if THY-1 was expressed as 25 a secreted molecule, the induction by rev was much more prominent, supporting the above hypothesis. This can probably be explained by accumulation of secreted protein in the supernatant, which considerably amplifies the rev effect. If rev only induces a minor increase for surface 30 molecules in general, induction of HIV envelope by rev cannot have the purpose of an increased surface abundance, but rather of an increased intracellular gpl60 level. It is completely unclear at the moment why this should be the case.

To test whether small subtotal elements of a gene are sufficient to restrict expression and render it rev-dependent rTHYlenv:immunoglobulin fusion proteins were generated, in which only about one third of the total 5 gene had the envelope codon usage. Expression levels of this construct were on an intermediate level, indicating that the rTHY-lenv negative sequence element is not dominant over the immunoglobulin part. This fusion protein was not or only slightly rev-responsive, 10 indicating that only genes almost completely suppressed can be rev-responsive.
Another characteristic feature that was found in the codon frequency tables is a striking under representation of CpG triplets. In a comparative study 15 of codon usage in E. coli, yeast, drosophila and primates it was shown that in a high number of analyzed primate genes the 8 least used codons contain all codons with the CpG dinucleotide sequence. Avoidance of codons cont~; n; ng this dinucleotide motif was also found in the 20 sequence of other retroviruses. It seems plausible that the reason for under representation of CpG-bearing triplets has something to do with avoidance of gene silencing by methylation of CpG cytosines. The expected number of CpG dinucleotides for HIV as a whole is about 25 one fifth that expected on the basis of the base composition. This might indicate that the possibility of high expression is restored, and that the gene in fact has to be highly expressed at some point during viral pathogenesis.
The results presented herein clearly indicate that codon preference has a severe effect on protein levels, and suggest that translational elongation is controlling ~ ~lian gene expression. However, other factors may play a role. First, abundance of not maximally loaded 35 mRNA's in eukaryotic cells indicates that initiation is W O 97/11086 PCTfUS96/15088 rate limiting for translation in at least some cases, since otherwise all transcripts would be completely covered by ribosomes. Furthermore, if ribosome stalling and subsequent mRNA degradation were the ?ch~n;~ , 5 suppression by rare codons could most likely not be reversed by any regulatory ?ch~n; ! like the one presented herein. one possible explanation for the influence of both initiation and elongation on translational activity is that the rate of initiation, or 10 access to ribosomes, is controlled in part by cues distributed throughout the RNA, such that the lentiviral codons predispose the RNA to a~ ate in a pool of poorly initiated RNAs. However, this limitation need not be kinetic; for example, the choice of codons could 15 influence the probability that a given translation product, once initiated, is properly completed. Under this ?~h~n;~, abundance of less favored codons would incur a significant cumulative probability of failure to complete the nascent polypeptide chain. The sequestered 20 RNA would then be lent an improved rate of initiation by the action of rev. Since adenine residues are abundant in rev-responsive transcripts, it could be that RNA
adenine methylation mediates this translational suppression.
25 Detailed Procedures The following procedures were used in the above-described experiments.
5e~uence AnalYsis Sequence analyses employed the software developed 30 by the University of Wisconsin Computer Group.
Plasmid constructions Plasmid constructions employed the following methods. Vectors and insert DNA was digested at a concentration of 0.5 ~g/10 ~l in the appropriate 35 restriction buffer for 1 - 4 hours (total reaction volume CA 0223l394 l998-03-09 O 97/11086 PCT~US96/15088 approximately 30 ,ul). Digested vector was treated with 10% (v/v) of 1 ~g/ml calf intestine alkaline phosphatase for 30 min prior to gel electrophoresis. Both vector and insert digests (5 to 10 ~l each) were run on a 1.5% low melting agarose gel with TAE buffer. Gel slices containing bands of interest were transferred into a 1.5 ml reaction tube, melted at 65~C and directly added to the ligation without removal of the agarose. Ligations were typically done in a total volume of 25 ~ll in lx Low Buffer lx Ligation Additions with 200-400 U of ligase, 1 ,ul of vector, and 4 ,ul of insert. When necessary, 5' overhanging ends were filled by adding 1/10 volume of 250 ,uM dNTPs and 2--5U of Klenow polymerase to heat inactivated or phenol extracted digests and incubating for approximately 20 min at room temperature. When necessary, 3' overhanging ends were filled by adding 1/10 volume of 2.5 mM dNTPs and 5-10 U of T4 DNA polymerase to heat inactivated or phenol extracted digests, followed by incubation at 37~C for 30 min. The following buffers were used in these reactions: lOx Low buffer (60 mM Tris HCl, pH 7.5, 60 mM MgCl2, 50 mM NaCl, 4 mg/ml BSA, 70 mM
~-mercaptoethanol, 0.02% NaN3); lOx Medium buffer (60 mM
Tris HCl, pH 7.5, 60 mM MgCl2, 50 mM NaCl, 4 mg/ml BSA, 70 mM ,t~--mercaptoethanol, 0. 02% NaN3); lOx High buffer (60 mM Tris HCl, pH 7.5, 60 mM MgCl2, 50 mM NaCl, 4 mg/ml BSA, 70 mM ,~--mercaptoethanol, 0. 02% NaN3); lOx Ligation additions (1 mM ATP, 20 mM DTT, 1 mg/ml BSA, 10 mM
spermidine); 50x TAE (2 M Tris acetate, 50 mM EDTA).
Oli~onucleotide sYnthesis and ~urification Oligonucleotides were produced on a Milligen 8750 synthesizer (Millipore). The columns were eluted with 1 ml of 30% ammonium hydroxide, and the eluted oligonucleotides were deblocked at 55~C for 6 to 12 hours. After deblockiong, 150 ~l of oligonucleotide were precipitated with lOx volume of unsaturated n--butanol in 1.5 ml reaction tubes, followed by centrifugation at 15,000 rpm in a microfuge. The pellet was washed with 70% ethanol and resuspended in 50 ~l of H20. The concentration was determined by measuring the optical 5 density at 260 nm in a dilution of 1:333 (1 OD260 = 30 ~g/ml).
The following oligonucleotides were used for construction of the synthetic gpl20 gene (all sequences shown in this text are in 5' to 3' direction).
oligo 1 forward (Nhel): cgc ggg cta gcc acc gag aag ctg (SEQ ID N0:1).
oligo 1: acc gag aag ctg tgg gtg acc gtg tac tac ggc gtg ccc gtg tgg aag ag ag gcc acc acc acc ctg ttc tgc gcc agc gac gcc aag gcg tac gac acc gag gtg cac aac gtg 15 tgg gcc acc cag gcg tgc gtg ccc acc gac ccc aac ccc cag gag gtg gag ctc gtg aac gtg acc gag aac ttc aac at (SEQ
ID N0:2).
oligo 1 reverse: cca cca tgt tgt tct tcc aca tgt tga agt tct c (SEQ ID NO~3)-.
oligo 2 forward: gac cga gaa ctt caa cat gtg gaa gaa caa cat (SEQ ID N0:4) oligo 2: tgg aag aac aac atg gtg gag cag atg cat gag gac atc atc agc ctg tgg gac cag agc ctg aag ccc tgc gtg aag ctg acc cc ctg tgc gtg acc tg aac tgc acc gac ctg 25 agg aac acc acc aac acc aac ac agc acc gcc aac aac aac agc aac agc gag ggc acc atc aag ggc ggc gag atg (SEQ ID
N0:5).
oligo 2 reverse (Pstl): gtt gaa gct gca gtt ctt cat ctc gcc gcc ctt (SEQ ID NO:6).
oligo 3 forward (Pstl): gaa gaa ctg cag ctt caa cat cac cac cag c (SEQ ID N0:7).
oligo 3: aac atc acc acc agc atc cgc gac aag atg cag aag gag tac gcc ctg ctg tac aag ctg gat atc gtg agc atc gac aac gac agc acc agc tac cgc ctg atc tcc tgc aac 35 acc agc gtg atc acc cag gcc tgc ccc aag atc agc ttc gag W O 97/11086 PCT~US96/15088 ccc atc ccc atc cac tac tgc gcc ccc gcc ggc ttc gcc (SEQ
ID N0:8).
oligo 3 reverse: gaa ctt ctt gtc ggc ggc gaa gcc ggc ggg (SEQ ID N0:9).
oligo 4 forward: gcg ccc ccg ccg gct tcg cca tcc tga agt gca acg aca aga agt tc (SEQ ID N0:10) oligo 4: gcc gac aag aag ttc agc ggc aag ggc agc tgc aag aac gtg agc acc gtg cag tgc acc cac ggc atc cgg ccg gtg gtg agc acc cag ctc ctg ctg aac ggc agc ctg 10 gcc gag gag gag gtg gtg atc cgc agc gag aac ttc acc gac aac gcc aag acc atc atc gtg cac ctg aat gag agc gtg cag atc (SEQ ID N0:11) oligo 4 reverse (Mlul): agt tgg gac gcg tgc agt tga tct gca cgc tct c (SEQ ID N0:12).
oligo 5 forward (Mlul): gag agc gtg cag atc aac tgc acg cgt ccc (SEQ ID N0:13).
oligo 5: aac tgc acg cgt ccc aac tac aac aag cgc aag cgc atc cac atc ggc ccc ggg cgc gcc ttc tac acc acc aag aac atc atc ggc acc atc ctc cag gcc cac tgc aac atc 20 tct aga (SEQ ID N0:14) .
oligo 5 reverse: gtc gtt cca ctt ggc tct aga gat gtt gca (SEQ ID N0:15).
oligo 6 forward: gca aca tct cta gag cca agt gga acg ac (SEQ ID N0:16).
oligo 6: gcc aag tgg aac gac acc ctg cgc cag atc gtg agc aag ctg aag gag cag ttc aag aac aag acc atc gtg ttc ac cag agc agc ggc ggc gac ccc gag atc gtg atg cac agc ttc aac tgc ggc ggc (SEQ ID N0:17).
oligo 6 reverse (EcoR1): gca gta gaa gaa ttc gcc~0 gcc gca gtt ga (SEQ ID N0:18).
oligo 7 forward (EcoR1): tca act gcg gcg gcg aat tct tct act gc (SEQ ID N0:19).
oligo 7: ggc gaa ttc ttc tac tgc aac acc agc ccc ctg ttc aac agc acc tgg aac ggc aac aac acc tgg aac aac 35 acc acc ggc agc aac aac aat att acc ctc cag tgc aag atc aag cag atc atc aac atg tgg cag gag gtg ggc aag gcc atg tac gcc ccc ccc atc gag ggc cag atc cgg tgc agc agc (SEQ
ID NO:20) oligo 7 reverse: gca gac cgg tga tgt tgc tgc tgc 5 acc gga tct ggc cct c (SEQ ID NO:21).
oligo 8 forward: cga ggg cca gat ccg gtg cag cag caa cat cac cgg tct g (SEQ ID N0:22).
oligo 8: aac atc acc ggt ctg ctg ctg acc cgc gac ggc ggc aag gac acc gac acc aac gac acc gaa atc ttc cgc 10 ccc ggc ggc ggc gac atg cgc gac aac tgg aga tct gag ctg tac aag tac aag gtg gtg acg atc gag ccc ctg ggc gtg gcc ccc acc aag gcc aag cgc cgc gtg gtg cag cgc gag aag cgc (SEQ ID N0:23).
oligo 8 reverse (Notl): cgc ggg cgg ccg ctt tag 15 cgc ttc tcg cgc tgc acc ac (SEQ ID N0:24).
The following oligonucleotides were used for the construction of the ratTHY-lenv gene.
oligo 1 forward (BamHl/Hind3): cgc ggg gga tcc aag ctt acc atg att cca gta ata agt (SEQ ID N0:25).
oligo 1: atg aat cca gta ata agt ata aca tta tta tta agt gta tta caa atg agt aga gga caa aga gta ata agt tta aca gca tct tta gta aat caa aat ttg aga tta gat tgt aga cat gaa aat aat aca aat ttg cca ata caa cat gaa ttt tca tta acg (SEQ ID N0:26).
oligo 1 reverse (EcoRl/Mlul): cgc ggg gaa ttc acg cgt taa tga aaa ttc atg ttg (SEQ ID N0:27).
oligo 2 forward (BamHl/Mlul): cgc gga tcc acg cgt gaa aaa aaa aaa cat (SEQ ID N0:28).
oligo 2: cgt gaa aaa aaa aaa cat gta tta agt gga 30 aca tta gga gta cca gaa cat aca tat aga agt aga gta aat ttg ttt agt gat aga ttc ata aaa gta tta aca tta gca aat ttt aca aca aaa gat gaa gga gat tat atg tgt gag (SEQ ID
N0:29).
oligo 2 reverse (EcoRl/Sacl): cgc gaa ttc gag ctc 35 aca cat ata atc tcc (SEQ ID N0:30).

W O 97/11086 PCTrUS96/15088 oligo 3 forward (BamH1/Sacl): cgc gga tcc gag ctc aga gta agt gga caa (SEQ ID N0:31).
~ oligo 3: ctc aga gta agt gga caa aat cca aca agt agt aat aaa aca ata aat gta ata aga gat aaa tta gta aaa 5 tgt ga gga ata agt tta tta gta caa aat aca agt tgg tta tta tta tta tta tta agt tta agt ttt tta caa gca aca gat ttt ata agt tta tga (SEQ ID N0:32).
oligo 3 reverse (EcoR1/Notl): cgc gaa ttc gcg gcc gct tca taa act tat aaa atc (SEQ ID N0:33).
10 PolYmerase Chain Reaction Short, overlapping 15 to 25 mer oligonucleotides annealing at both ends were used to amplify the long oligonuclotides by polymerase chain reaction (PCR).
Typical PCR conditions were: 35 cycles, 55~C annealing 15 temperature, 0.2 sec extension time. PCR products were gel purified, phenol extracted, and used in a subsequent PCR to generate longer fragments consisting of two adjacent small fragments. These longer fragments were cloned into a CDM7-derived plasmid containing a leader 20 sequence of the CD5 surface molecule followed by a Nhel/Pstl/Mlul/EcoR1/BamH1 polylinker.
The following solutions were used in these reactions: lOx PCR buffer (500 mM KCl, lO0 mM Tris HCl, pH 7.5, 8 mM MgCl2, 2 mM each dNTP). The final buffer 25 was complemented with 10% DMS0 to increase fidelity of the Taq polymerase.
Small scale DNA pre~aration Transformed bacteria were grown in 3 ml LB
cultures for more than 6 hours or overnight.
30 Approximately 1.5 ml of each culture was poured into 1.5 ml microfuge tubes, spun for 20 seconds to pellet cells and resuspended in 200 ~l of solution I. Subsequently 400 ~l of solution II and 300 ~l of solution III were added. The microfuge tubes were capped, mixed and spun 35 for > 30 sec. Supernatants were transferred into fresh tubes and phenol extracted once. DNA was precipitated by filling the tubes with isopropanol, mixing, and spinning in a microfuge for > 2 min. The pellets were rinsed in 70 % ethanol and resuspended in 50 ~l dH20 contA;n;ng 10 5 ~l of RNAse A. The following media and solutions were used in these procedures: LB medium (1.0 % NaCl, 0.5%
yeast extract, 1.0~ trypton); solution I (10 mM EDTA pH
8.0); solution II (0.2 M NaOH, 1.0% SDS); solution III
(2.5 M KOAc, 2.5 M glacial aceatic acid); phenol (pH
10 adjusted to 6.0, overlaid with TE); TE (10 mM Tris HCl, pH 7.5, 1 mM EDTA pH 8.0).
Larqe scale DNA ~reparation One liter cultures of transformed bacteria were grown 24 to 36 hours (MCl061p3 transformed with pCDM
15 derivatives) or 12 to 16 hours (MC1061 transformed with pUC derivatives) at 37~C in either M9 bacterial medium (pCDM derivatives) or LB (pUC derivatives). Bacteria were spun down in 1 liter bottles using a Beckman J6 centrifuge at 4,200 rpm for 20 min. The pellet was 20 resuspended in 40 ml of solution I. Subsequently, 80 ml of solution II and 40 ml of solution III were added and the bottles were shaken semivigorously until lumps of 2 to 3 mm size developed. -The bottle was spun at 4,200 rpm for 5 min and the supernatant was poured through 25 cheesecloth into a 250 ml bottle. Isopropanol was added to the top and the bottle was spun at 4,200 rpm for 10 min. The pellet was resuspended in 4.1 ml of solution I
and added to 4.5 g of cesium chloride, 0.3 ml of 10 mg/ml ethidium bromide, and 0.1 ml of 1% Triton X100 solution.
30 The tubes were spun in a Beckman J2 high speed centrifuge at 10,000 rpm for 5 min. The supernatant was transferred into Beckman Quick Seal ultracentrifuge tubes, which were then sealed and spun in a Beclcman ultracentrifuge using a NVT90 fixed angle rotor at 80,000 rpm for > 2.5 hours.
35 The band was extracted by visible light using a 1 ml W O 97/11086 PCT~US96/1~088 syringe and 20 gauge needle. An equal volume of dH2O was added to the extracted material. DNA was extracted once with n-butanol saturated with 1 M sodium chloride, followed by addition of an equal volume of 10 M ammonium acetate/ 1 mM EDTA. The material was poured into a 13 ml snap tube which was tehn filled to the top with absolute ethanol, mixed, and spun in a Beckman J2 centrifuge at 10,000 rpm for 10 min. The pellet was rinsed with 70%
ethanol and resuspended in 0. 5 to 1 ml of H2O. The DNA
10 concentration was determined by measuring the optical density at 260 nm in a dilution of 1:200 (1 OD260 = 50 ~g/ml).
The following media and buffers were used in these procedures: M9 bacterial medium (10 g M9 salts, 10 g I A~A~;nO acids (hydrolyzed), 10 ml M9 additions, 7.5 ,ug/ml tetracycline (500 ,ul of a 15 mg/ml stock solution), 12.5 ~g/ml ampicillin (125 ~1 of a 10 mg/ml stock solution); M9 additions (10 mM CaC12, 100 mM MgSO4, 200 ,ug/ml thiamine, 70% glycerol); LB medium (1.0 % NaCl, 0.5 % yeast extract, 1.0 % trypton); Solution I (10 mM EDTA
pH 8.0); Solution II (0.2 M NaOH 1.0 % SDS); Solution III
(2.5 M KOAc 2.5 M HOAc) Se~uencinq Synthetic genes were sequenced by the Sanger dideoxynucleotide method. In brief, 20 to 50 ,ug double--stranded plasmid DNA were denatured in 0.5 M NaOH for 5 min. Subsequently the DNA was precipitated with 1/10 volume of sodium acetate (pH 5.2) and 2 volumes of ethanol and centrifuged for 5 min. The pellet was washed with 70% ethanol and resuspended at a concentration of 1 ~g/~l. The annealing reaction was carried out with 4 ~g of template DNA and 40 ng of primer in lx annealing buffer in a final volume of 10 ~1. The reaction was heated to 65~C and slowly cooled to 37~C. In a separate tube 1 ,ul of 0.1 M DTT, 2 fl.l of labeling mix, 0.75 ,Ul of W O 97/11086 PCT~US96/15088 dH20, 1 ~1 of [35S] dATP (10 uCi), and 0.25 ~1 of Sequenase~ (12 U/~l) were added for each reaction. Five ~1 of this mix were added to each annealed primer-template tube and incubated for 5 min at room 5 temperature. For each labeling reaction 2.5 ~1 of each of the 4 termination mixes were added on a Terasaki plate and prewarmed at 37~C. At the end of the incubation period 3.5 ~1 of labeling reaction were added to each of the 4 termination mixes. After 5 min, 4 ~1 of stop 10 solution were added to each reaction and the Terasaki plate was incubated at 80~C for 10 min in an oven. The sequencing reactions were run on 5% denaturing polyacrylamide gel. An acrylamide solution was prepared by adding 200 ml of lOx TBE buffer and 957 ml of dH20 to 15 100 g of acrylamide:bisacrylamide (29:1). 5%
polyacrylamide 46% urea and lx TBE gel was prepared by combining 38 ml of acrylamide solution and 28 g urea.
Polymerization was initiated by the addition of 400 ~1 of 10% ammonium peroxodisulfate and 60 ~1 of TEMED. Gels 20 were poured using silanized glass plates and sharktooth combs and run in lx TBE buffer at 60 to 100 W for 2 to 4 hours (depending on the region to be read). Gels were transferred to Whatman blotting paper, dried at 80~C for about 1 hour, and exposed to x-ray film at room 25 temperature. Typically exposure time was 12 hours. The following solutions were used in these procedures: 5x Annealing buffer (200 mM Tris HCl, pH 7.5, 100 mM MgCl2, 250 mM NaCl); Labelling Mix (7.5 ~M each dCTP, dGTP, and dTTP); Termination Mixes (80 ~M each dNTP, 50 mM NaCl, 8 30 ~M ddNTP (one each)); Stop solution (95% formamide, 20 mM
EDTA, 0.05 ~ bromphenol blue, 0.05 % xylencyanol); 5x TBE
(O.9 M Tris borate, 20 mM EDTA); Polyacrylamide solution (96.7 g polyacrylamide, 3.3 g bisacrylamide, 200 ml lx TBE, 957 ml dH20).

CA 0223l394 l998-03-09 W O 97/11086 PCT~US96/15088 RNA isolation Cytoplasmic RNA was isolated from calcium phosphate transfected 293T cells 36 hours post transfection and from vaccinia infected Hela cells 16 ~ 5 hours post infection essentially as described by Gilman.
(Gilman Preparation of cytoplasmic RNA from tissue culture cells. In Current Protocols in Molecular Biology, Ausubel et al., eds., Wiley & Sons, New York, 1992). Briefly, cells were lysed in 400 ,ul lysis buffer, 10 nuclei were spun out, and SDS and proteinase K were added to 0.2% and 0.2 mg/ml respectively. The cytoplasmic extracts were incubated at 37~C for 20 min, phenol/chloroform extracted twice, and precipitated. The RNA was dissolved in 100 ~l buffer I and incubated at 15 37~C for 20 min. The reaction was stopped by adding 25 ~l stop buffer and precipitated again.
The following solutions were used in this procedure: Lysis Buffer (TRUSTEE containing with 50 mM
Tris pH 8.0, 100 mM NaCl, 5 mM MgCl2, 0.5~ NP40); Buffer I (TRUSTEE buffer with 10 mM MgC12, 1 mM DTT, 0.5 U/,ul placental RNAse inhibitor, 0.1 U/~l RNAse free DNAse I);
Stop buffer (50 mM EDTA 1.5 M NaOAc 1.0 % SDS).
Slot blot analvsis For slot blot analysis 10 ~g of cytoplasmic RNA
25 was dissolved in 50 ~l dH2O to which 150 ~l of 10x SSC/18% formaldehyde were added. The solubilized RNA was then incubated at 65~C for 15 min and spotted onto with a slot blot apparatus. ~lo~tively labeled probes of 1.5 kb gpl20IIIb and syngpl~0mn f~a~ments were used for 30 hybridization. Each of thc two fragment~ was random labeled in a 50 ~1 reaction with 10 ~1 of 5x oligo-labeling buffer, 8 ~1 of 2.5 mg/ml BSA, 4 ~l of ~[32p]_ dCTP (20 uCi/~l; 6000 Ci/mmol), and 5 U of Klenow ~ragment. After 1 to 3 hours incubation at 37~C 100 ~l 35 of TRUSTEE were added and unincorporated o'[32P]-dCTP was eliminated using G50 spin column. Activity was measured in a ~k~n beta-counter, and equal specific activities were used for hybridization. Membranes were pre-hybridized for 2 hours and hybridized for 12 to 24 hours 5 at 42~C with 0.5 x 106 cpm probe per ml hybridization fluid. The membrane was washed twice (5 min) with washing buffer I at room temperature, for one hour in washing buffer II at 65~C, and then exposed to x-ray film. Similar results were obtained using a 1.1 kb 10 Notl/Sfil fragment of pCDM7 cont~;n;ng the 3 untranslated region. Control hybridizations were done in parallel with a random-labeled human beta-actin probe. RNA
expression was quantitated by s~-~nn;ng the hybridized nitrocellulose membranes with a Magnetic Dynamics 15 phosphorimager.
The following solutions were used in this procedure:
5x Oligo--labeling buffer (250 mM Tris HCl, pH 8.0, 25 mM
MgC12, 5 mM ~-mercaptoethanol, 2 mM dATP, 2 mM dGTP, mM
20 d~TP, 1 M Hepes pH 6.6, 1 mg/ml hexanucleotides [dNTP]6);
Hybridization Solution (_ M sodium phosphate, 250 mM
NaCl, 7% SDS, 1 mM EDTA, 5% dextrane sulfate, 50%
formamide, 100 ~g/ml denatured salmon sperm DNA); Washing buffer I (2x SSC, 25 0.1% SDS); W~h;ng buffer II (0.5x SSC, 0.1 % SDS); 20x SSC (3 M NaCl, 0.3 M Na3citrate, pH adjusted to 7.0).
Vaccinia recombination Vaccinia recombination used a modification of the of the method described by Romeo and Seed (Romeo and 30 Seed, Cell, 64: 1037, 1991). Briefly, CV1 cells at 70 to 90% confluency were infected with 1 to 3 ~l of a wild-type vaccinia stock WR (2 x lo8 pfu/ml) for 1 hour in culture medium without calf serum. After 24 hours, the cells were transfected by calcium phosphate with 25 ~g 35 TKG plasmid DNA per dish. A~ter an additional 24 to 48 hours the cells were scraped off the plate, spun down, and resuspended in a volume of 1 ml. After 3 freeze/thaw cycles trypsin was added to 0.05 mg/ml and lysates were incubated for 20 min. A dilution series of 5 10, 1 and 0.1 ~l of this lysate was used to infect small dishes (6 cm) of CVl cells, that had been pretreated with 12.5 ,~g/ml mycophenolic acid, 0.25 mg/ml xanthin and 1. 36 mg/ml hypoxanthine for 6 hours. Infected cells were cultured for 2 to 3 days, and subsequently stained with 10 the monoclonal antibody NEA9 301 against gpl20 and an alkaline phosphatase conjugated secondary antibody.
Cells were incubated with 0. 33 mg/ml NBT and 0.16 mg/ml BCIP in AP-buffer and finally overlaid with 1% agarose in PBS. Positive plaques were picked and resuspended in 15 100 ~l Tris pH 9Ø The plaque purification was repeated once. To produce high titer stocks the infection was slowly scaled up. Finally, one large plate of Hela cells was infected with half of the virus of the previous round. Infected cells were detached in 3 ml of PBS, 20 lysed with a Dounce homogenizer and cleared from larger debris by centrifugation. VPE-8 recombinant vaccinia stocks were kindly provided by the AIDS repository, Rockville, MD, and express HIV-1 IIIB gpl20 under the 7.5 ;~e~ early/late promoter (Earl et al., J. Virol., 65: 31, 25 1991). In all experiments with recombinant vaccina cells were infected at a multiplicity of infection of at least 10 .
The following solution was used in this procedure:
AP buffer !loO mM Tris HCl pH 9.5 100 mM NaCl 5 mM
30 MgCl2) Cell culture The monkey kidney-carcinoma cell lines CVl and Cos7, the human kidney carcinoma cell line 293T, and the human cervix carcinoma cell line Hela were obtained from 3 5 the American Tissue Typing Collection and were maintained in supplemented IMDM. They were kept on 10 cm tissue culture plates and typically split 1:5 to 1:20 every 3 to 4 days. The following medium was used in this procedure:
5 Supplemented IMDM (90% Iscove's modified Dlllh~-o Medium, 10~ calf serum, iron-complemented, heat inactivated 30 min 56~C, 0.3 mg/ml L-glutamine, 25 ~g/ml gentamycin 0.5 mM ~-mercaptoethanol (pH adjusted with 5 M NaOH, 0.5 ml)).
10 Transfection Calcium phosphate transfection of 293T cells was performed by slowly adding and under vortexing 10 ~g plasmid DNA in 250 ~1 0.25 M CaC12 to the same volume of 2x HEBS buffer while vortexing. After incubation for 10 15 to 30 min at room temperature the DNA precipitate was added to a small dish of 50 to 70% confluent cells. In cotransfection experiments with rev, cells were transfected with 10 ~g gpl20IIIb, gpl20IIIbrre, syngpl20mnrre or rTHY-lenveglrre and 10 ~g of pCMVrev or 20 CDM7 plasmid DNA.
The following solutions were used in this procedure: 2x HEBS buffer (280 mM NaCl, 10 mM KCl, 1.5 mM
sterile filtered); 0.25 mM CaC12 (autoclaved).
Immuno~reci~itation After 48 to 60 hours medium was ~c-h~nged and cells were ;ncl~h~ted for additional 12 hours in Cys/Met-free medium con~;n;ng 200 ~Ci of 35S-translabel.
Supernatants were harvested and spun for 15 min at 3000 rpm to remove debris. After addition of protease 30 inhibitors leupeptin, aprotinin and PMSF to 2.5 ~g/ml, 50 ~g/ml, 100 ~g/ml respectively, 1 ml of supernatant was incubated with either 10 ~1 of packed protein A sepharose alone (rTHY-lenveglrre) or with protein A sepharose and 3 ~g of a purified CD4/immunoglobulin fusion protein (kindly provided by Behring) (all gpl20 constructs) at 4~C for 12 hours on a rotator. Subsequently the protein A beads were washed 5 times for 5 to 15 min each time.
After the final wash 10 ~l of loading buffer containing was added, samples were boiled for 3 min and applied on 5 7% (all gpl20 constructs) or 10% (rTHY-lenveglrre) SDS
polyacrylamide gels (TRIS pH 8.8 buffer in the resolving, TRIS pH 6.8 buffer in the stacking gel, TRIS-glycin running buffer, Maniatis et al. 1989). Gels were fixed in 10% acetic acid and 10 % methanol, incubated with 10 Amplify for 20 min, dried and exposed for 12 hours.
The following buffers and solutions were used in this procedure: Wash buffer (100 mM Tris, pH 7.5, 150 mM
NaCl, 5 mM CaCl2, 1% NP-40); 5x Running Buffer (125 mM
Tris, 1.25 M Glycin, 0.5% SDS); Loading buffer (10 %
15 glycerol, 4% SDS, 4% ~-mercaptoethanol, 0.02 % bromphenol blue).
Immunofluorescence 293T cells were transfected by calcium phosphate coprecipitation and analyzed for surface THY-l expression 20 after 3 days. After detachment with 1 mM EDTA/PBS, cells were stained with the monoclonal antibody OX-7 in a dilution of 1:250 at 40C for 20 min, washed with PBS and subsequently incubated with a 1:500 dilution of a FITC-conjugated goat anti-mouse immunoglobulin antiserum.
25 Cells were washed again, resuspended in 0.5 ml of a fixing solution, and analyzed on a EPICS XL
cytofluorometer (Coulter).
The following solutions were used in this procedure:
30 PBS (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na2HP04, 1.4 mM
KH2P04, pH adjusted to 7.4); Fixing solution (2%
formaldehyde in PBS).
ELISA
The concentration of gpl20 in culture supernatants 35 was determined using CD4-coated ELISA plates and goat anti-gpl20 antisera in the soluble phase. Supernatants of 293T cells transfected by calcium phosphate were harvested after 4 days, spun at 3000 rpm for 10 min to remove debris and incubated for 12 hours at 40C on the 5 plates. After 6 washes with PBS 100 ~l of goat anti-gpl20 antisera diluted 1:200 were added for 2 hours. The plates were washed again and incubated for 2 hours with a peroxidase-conjugated rabbit anti-goat IgG antiserum 1:1000. Subsequently the plates were washed and 10 incubated for 30 min with 100 ~l of substrate solution cont~; n; ng 2 mg/ml o-phenylenediamine in sodium citrate buffer. The reaction was finally stopped with 100 ~l of 4 M sulfuric acid. Plates were read at 490 nm with a Coulter microplate reader. Purified recombinant 15 gpl20IIIb was used as a control. The following buffers and solutions were used in this procedure: Wash buffer (0.1% NP40 in PBS); Substrate solution (2 mg/ml o-phenylene~;~m;ne in sodium citrate buffer).
Green Fluorescent Protein The efficacy of codon replacement for gpl20 suggests that replacing non-preferred codons with less preferred codons or preferred codons (and replacing less preferred codons with preferred codons) will increase expression in ~ ~lian cells of other proteins, e.g., 25 other eukaryotic proteins.
The green fluorescent protein (GFP) of the jellyfish Aequorea victoria (Ward, Photochem. Photobiol.
4:1, 1979; Prasher et al., Gene 111:229, 1992; Cody et al., Biochem. 32:1212, 1993) has attracted attention 30 recently for its possible utility as a marker or reporter for transfection and lineage studies (Chalfie et al., Science 263:802, 1994).
~ m; n~tion of a codon usage table constructed from the native coding se~uence of GFP showed that the 35 GFP codons favored either A or U in the third position.

W O 97/11086 PCTrUS96/15088 The bias in this case favors A less than does the bias of gpl20, but is substantial. A synthetic gene was created in which the natural GFP sequence was re-engineered in much the same manner as for gpl20. In addition, the 5 translation initiation sequence of GFP was replaced with sequences corresponding to the translational initiation consensus. The expression of the resulting protein was contrasted with that of the wild type sequence, similarly engineered to bear an optimized translational initiation 10 consensus (FIG. 10, panel B and FIG. 10, panel C). In addition, the effect of inclusion of the mutation Ser 65-Thr, reported to improve excitation efficiency of GFP
at 490 nm and hence preferred for fluorescence microscopy (Heim et al., Nature 373:663,1995), was ~A ;ned (FIG.
15 10, panel D). Codon engineering conferred a significant increase in expression efficiency (an concomitant percentage of cells apparently positive for transfection), and the combination of the Ser 65-Thr mutation and codon optimization resulted in a DNA segment 20 encoding a highly visible mammalian marker protein (FIG.
10, panel D).
The above-described synthetic green fluorescent protein coding sequence was assembled in a similar manner as for gpl20 from six fragments of approximately 120 bp 25 each, using a strategy for assembly that relied on the ability of the restriction enzymes BsaI and BbsI to cleave outside of their r~ogn;tion sequence. Long oligonucleotides were synthesized which con~A;n~
portions of the coding sequence ~or GFP embedded in 30 flanking sequences encoding EcoRI and BsaI at one end, and BamHI and BbsI at the other end. Thus, each oligonucleotide has the configuration EcoRI/BsaI/GFP
fragment/BbsI/BamHI. The restriction site ends generated by the BsaI and BbsI sites were designed to yield 35 compatible ends that could be used to join adjacent GFP

W O 97/11086 PCT~US96tlS088 fragments. Each of the compatible ends were designed to be uni~ue and non-selfcomplementary. The crude synthetic DNA segments were amplified by PCR, inserted between EcoRI and BamHI in pUC9, and sequenced. Subse~uently the 5 intact coding sequence was assembled in a six fragment ligation, using insert fragments prepared with BsaI and BbsI. Two of six plasmids resulting from the ligation bore an insert of correct size, and one contained the desired full length sequence. Mutation of Ser65 to Thr 10 was accomplished by st~n~A~d PCR based mutagenesis, using a primer that overlapped a unique BssSI site in the synthetic GFP.

Codon optimization as a strateqY for imProved expression in mammalian cells The data presented here suggest that coding sequence re-engineering may have general utility for the improv~ snt of expression of mammalian and non- ~ ~lian eukaryotic genes in ~ ~lian cells. The results obtained here with three unrelated proteins: HIV gpl20, 20 the rat cell surface antigen Thy-l and green fluorescent protein fro~ Aequorea victoria, suggest that codon optimization may prove to be a ~ruitful strategy for improving the expression in mammalian cells of a wide variety of eukaryotic genes.
25 Use The synthetic genes of the invention are useful for expressing the a protein normally expressed in mammalian cells in cell culture (e.g. for ~. A~cial production of human proteins such as hGH, TPA, Factor 30 VII, and Factor IX). The synthetic genes of the invention are also useful for gene therapy.
Synthetic GFP genes can be used in any application in which a native GFP gene or other reporter gene can be used. A synthetic GFP gene which employs more preferred codons than the native GFP gene can be the basis of a highly sensitive reporter system. Such a system can be used, e.g., to analyze the influence of particular promoter elements or trans-acting f actors on gene expression. Thus, the synthetic GFP gene can be used in much the same fashion as other reporters, e.g., ~-galactosidase, has been used.

CA 02231394 l998-03-09 WO 97/ll086 PCT~US96/l5088 - ~6 -~Q~N~ LISTING
( 1 ) G~N~R~T~ INFORMATION:
(i) APPLICANT: THE ~.~N~R~T. HOSPITAL CORPORATION
(ii) TITLE OF 1NV~:N ~ lON: HIGH LEVEL EXPRESSION OF PROTEINS
(iii) NUMBER OF SEQUENCES: 40 (iv) CORRE~uN~:N~ AnDRT--eS:
,'A', ADDRESSEE: Fi~h & Richardson P.C.
IB) STREET: 225 Franklin Street ,'C, CITY: Boston ~D'I STATE: Mas~achu~etts (El ~OUN-~Y: U.S.A.
~FJ ZIP: 02110-2804 (v) COMPUTER R~An~RT.~ FORM:
'A'I MEDIUM TYPE: Floppy disk I'B, COMPUTER: IBM PC compatible ,'C, OPERATING SYSTEM: PC-DOS/MS-DOS
~DJ SOFTWARE: PatentIn Relea~e #1.0, Version ~1.30B
(Vi) UU~R~L APPLICATION DATA:
(A) APPLICATION NUMBER: PCT/US96/----(B) FILING DATE: -SEP-199fi (vii) PRIOR APPLICATION DATA:
(A) APPLICATION NUMBER:08~532,390 (B) FILING DATE: 22-SEP-1995 (vii) PRIOR APPLICATION DATA:
(A) APPLICATION NUMBER:08/324,243 (B) FILING DATE: l9-SEP-1994 (viii) ATTORNEY/AGENT INFORMATION:
(A) NAME: LECH, RAREN F.
(B) REGISTRATION NUMBER: 35,238 (C) ~-~:k~N~E/DOCRET NUMBER: 00786/294001 (ix) T~T-~CnM~JNICATION INFORMATION:
(A) TELEPHONE: (617) 542-5070 (B) TELEFAX: (617) 542-8906 (C) TELEX: 200154 (2) INFORMATION FOR SEQ ID NO:1:
(i) sb:yu~u~ CHARACTERISTICS:
~A', LENGTH: 24 base pairs ~B' TYPE: nucleic acid ,'C, STR~NDEDNESS: single ~D~ TOPOLOGY: linear (xi) ~Qu~N~ DESCRIPTION: SEQ ID NO:1:
CGCGGGCTAG crArC~-Ar-AA GCTG 24 (2) INFORMATION FOR SEQ ID NO:2:
(i) SEQUENCE CHARACTERISTICS:
lA' LENGTH: 196 base pairs IBJ TYPE: nucleic acid ,CJ STRANn~nNESS: single ~DJ TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO:2:
ACC~-A~-AA~-C TGTGGGTGAC CGTGTACTAC GGCGTGCCCG TGTGGAAGAG AGGCCACCAC 60 CACC~lLC TGCGCCAGCG ACGCCAAGGC GTAC~-A~ACC GAGGTGCACA ACGTGTGGGC 120 ~ACC~AGCG TGCGTGCCCA CC~-ArCC~AA CCCC~AGGAG GTGGAGCTCG TGAACGTGAC 180 C~A~-AACTTC AACATG 196 (2) INFORMATION FOR SEQ ID NO:3:
(i) SEQUENCE CHARACTERISTICS:
/A~I LENGTH: 34 base pairs.
I'B~ TYPE: nucleic acid ~C) ST~ANnF~nNESS: single ~Dl TOPOLOGY: linear (Xi) ~yU~N~ DESCRIPTION: SEQ ID NO:3:
C~ATGTT ~LL~l~C~AC ATGTTGAAGT TCTC 34 (2) INFORMATION FOR SEQ ID NO:4:
(i) ~:yU~N~: CHARACTERISTICS:
'A' LENGTH: 33 base pairs 'B'I TYPE: nucleic acid C, STR~NDEDNESS: ~ingle ~Dl TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO:4:
GA~C~A~AAC TTCAACATGT G~-AAGAA~A CAT 33 (2) INFORMATION FOR SEQ ID NO:5:
(i) SEQUENCE CHARACTERISTICS:
,'A' LENGTH: 192 base pairs I'BI TYPE: nucleic acid ,'C, STRANDEDNESS: ~ingle ,D~ TOPOLOGY: linear W O 97/11086 PCT~US96/15088 (Xi) ~k~U~N~ DESCRIPTION: SEQ ID NO:5:
TGr-~Ar-~ArA ACATGGTGGA GCAGATGCAT GAGGACATCA TCAGCCTGTG Gn~rrAr-Ar-C 60 CTGAAGCCCT GCGTGAAGCT r~CCCC~1~' GCGTGACCTG AACTG Q CCG ACCTGAGGAA 120 rACr~rr~AC ACr~r~r~ CACCGC Q AC AArA~r~r,CA ACAGCGAGGG Q C Q TCAAG 180 (2) INFORMATION FOR SEQ ID NO: 6:
( i ) ~kyUkN~k CHARACTERISTICS:
,~A~I LENGTH: 33 ba~e pairs ,8 TYPE: nucleic acid ~C, sTRANn~nNEss: single ~D,, TOPOLOGY: linear (xL) ~kyu~N~ DESCRIPTION: SEQ ID NO: 6:
GTTGAAGCTG Q ~L'~'~A TCTCGCCGCC CTT 33 (2) INFORMATION FOR SEQ ID NO:7:
(i) ~bYU~N~ CHARACTERISTICS:
(A~ LENGTH: 31 base pairs IB) TYPE: nucleic acid ,C, STRANDEDNESS: ~ingle ~D~ TOPOLOGY: linear (Xi) SEQUENCE D~SrRTPTION: SEQ ID NO: 7:
rAA~AACTGC AGCTTCAACA T Q CCACCAG C 31 (2) INFORMATION FOR SEQ ID NO:8:
( i ) ~kyUk_._k CHARACTERISTICS:
~Aj LENGTH: 195 ba~e pairq iBI TYPE: nucleic acid ,CJ sTRANn~nN~.cs: 8ingle ~D~ TOPOLOGY: linear (Xi) ~QU~N~ D~CrRTPTION: SEQ ID NO:8:
AACAT Q CCA CCAGCATCCG Cr~r~ArA~G r~AAr~Ar,T ACGCCCTGCT GTACAAGCTG 60 GATATCGTGA GCATCGACAA CGACAGCACC AGCTACCGCC TGA1~1C~1G r~ArA~rAnC 120 W O 97/11086PCT~US96/15088 (2) Iwr~O~ATION FOR SEQ ID NO:9:
(i) ~:~ur,~._r; CHARACTERISTICS:
(A'l LENGTH: 30 base pair~
IBI TYPE: nucleic acid ,CJ STRANn~nNESS: single ~DJ TOPOLOGY: linear (xi) ~r;~ul-.._r; D~-C~RTPTION: SEQ ID NO:9:
GAA~~ G TCGGCGGCGA AGCCGGCGGG 30 (2) INFORMATION FOR SEQ ID NO:l0 (i) SEQUENCE CHARACTERISTICS:
,'A~ LENGTH: 47 ba~e pair~
,'BI TYPE: nucleic acid ,C, STRANDEDNESS: ~ingle ~DJ TOPOLOGY: linear (xi) ~r;Qur;N~r,- DESCRIPTION: SEQ ID NO:lO:
GCGCCCCCGC CGGCTTCGCC ATCCTGAAGT GrAACr-ArAA GAAGTTC 47 (2) INFORMATION FOR SEQ ID NO:ll:
(i) ~r;~ur;N~ CHARACTERISTICS:
(A'l LENGTH: 198 ba~e pair~
,BI TYPE: nucleic acid ,C, ST~RANn~nNESS: ~ingle ~DJ TOPOLOGY: linear (xi) ~yu~:w~ DESCRIPTION: SEQ ID NO:ll:
GCCr-Ar-AAr-A AGTTCAGCGG CAAGGGCAGC TGcAAr-AA~G TGAGCACCGT GCAGTGCACC 60 CACGGCATCC GGCCG~lG~l GAG~ACCrAr, CTCCTGCTGA ACGGCAGCCT GGCCGAGGAG 120 GAG~-G~LGA TCCGCAGCGA GAACTTCACC r-ArAArGCCA Ar-ArrATCAT CGTGCACCTG 180 (2) INFORMATION FOR SEQ ID NO:12:
(i) SEQUENCE CHARACTERISTICS:
,'A', LENGTH: 34 ba~e pairs ,'BI TYPE: nucleic acid ,C, STRANn~nNESS: ~ingle ~DJ TOPOLOGY: linear (Xi) ~:yU~N~ DESCRIPTION: SEQ ID NO:12:
A~I~GGACG CGTGCAGTTG ATCTGCACGC TCTC 34 (2) INFORMATION FOR SEQ ID NO:13:
( i ) ~yU~N~ CHARACTERISTICS:
/A~I LENGTH: 30 ba~e pairs B I TYPE: nucleic acid TRANn~nNF!.eS: single ~DJ TOPOLOGY: linear (Xi) ~U~N~ DESCRIPTION: SEQ ID NO:13:

(2) INFORMATION FOR SEQ ID NO:14:
(i) ~:~u~ CHARACTERISTICS:
(AJ LENGTH: 120 base pairs ~Bl TYPE: nucleic acid ,C~ STRANDEDNESS: single ~Dl TOPOLOGY: linear (Xi) ~yU~N~ ~-erPTPTION: SEQ ID NO:14:
AACTGCACGC GTCCCAACTA rAArAAGCGC AAGCGCATCC ACATCGGCCC CGGGCGCGCC 60 TTCTArACrA crAA~-AArAT CATCGGCACC A~C~lC~AGG CCCACTGCAA CATCTCTAGA 120 (2) INFORMATION FOR SEQ ID NO:15:
($) ~yu~ CHA~ACTERISTICS:
/AI LENGTH: 30 base pairs IBI TYPE: nucle$c acid ,CJ STRANn~nNESS: ~ingle ~Dl TOPOLOGY: linear (xi) ~QU~N~ DESCRIPTION: SEQ ID NO:15:
C~ ~ ~C~AC TTGGCTCTAG AGAL~1G~A 30 (2) INFORMATION FOR SEQ ID NO:16:
( i ) ~U~N~ CHARACTERISTICS:
(Aj LENGTH: 29 ba~e pairs ~B TYPE: nucleic acid ,C STRANDEDNESS: single ~DJ TOPOLOGY: linear W O 97/11086 PCTrUS96/15088 (Xi) ~yU~:N~ DESCRIPTION: SEQ ID NO:16:
GCAACATCTC TAGAGCCAAG TG~-~AC~-~C 29 (2) INFORMATION FOR SEQ ID NO:17:
( i ) ~U~N~ CHARACTERISTICS: .
,AI LENGTH: 131 baqe pairq IBI TYPE: nucleic acid ,C, STRANDEDNESS: single ~DJ TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO:17:
GCCAAGTGGA ACr.~CCCT GCGCCAGATC GTGAGCAAGC TGAAGGAGCA GTTrA~-AAC 60 AA~C~ATCG TGTTCACCAG AGCAGCGGCG GCGACCCCGA GATCGTGATG CACAGCTTCA 120 ACTGCGGCGG C 13l (2) INFORNATION FOR SEQ ID NO:18:
(i) SEQUENCE CH~RACTERISTICS:
~Al LENGTH: 29 ba~e pairq IBI TYPE: nucleic acid ,C, STRANDEDNESS: ~ingle ~Dl TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO:18:

(2) INFORMATION FOR SEQ ID NO:l9:
( i ) Y~QU~.._~ CHARACTERISTICS:
(AJ LENGTH: 29 ba~e pair~
IBJ TYPE: nucleic acid ,C, STRANDEDNESS: single ~DI TOPOLOGY: linear (xi) SEQUENCE D~-C~RTPTION: SEQ ID NO:l9:

(2) INFORMATION FOR SEQ ID NO:20:
(i) ~:yu~N~ CH~RACTERISTICS:
(A) LENGTH: 195 ba~e pairq (B) TYPE: nucleic acid (C) STR~NDEDNESS: ~ingle W 0 97/11086 PCT~US96/15088 (D) TOPOLOGY: linear (xL) ~Qu~._~ DESCRIPTION: SEQ ID NO:20:
GGCGAATTCT TCTACTG Q A Q CCAGCCCC ~l~ll~AACA G Q CCTGGAA CGGrAA~ 60 ACCTGGAACA A~ACrACCGG CAG~AACAAC AATATTACCC TCCAGTGCAA GATCAAG Q G 120 ATCATCAACA TGTGGCAGGA GGTGGGCAAG GCCATGTACG CCCCCCC~T CGAGGGCCAG 180 AlCCG~lGCA G Q GC 195 (2) lN~hATION FOR SEQ ID NO:21t ( i ) ~QU~N~ CHARACTERISTICS:
~A', LENGTH: 40 base pairs , B I TYPE: nucleic acid ,C, sTRANn~nNEss: single ,D~ TOPOLOGY: linear ( Xi ) ~U~N~ DESCRIPTION: SEQ ID NO:2l:
G~-~CCGGT GATGTTGCTG CTG~CG~-A TCTGGCCCTC40 (2) INFORMATION FOR SEQ ID No:22:
(i) ~yu~. CHARACTERISTICS:
~A'I LENGTH: 40 base pairs 'B' TYPE: nucleic acid C, STRANn~nN~S: ~ingle ,DI TOPOLOGY: linear (Xi) ~U~N~ DESCRIPTION: SEQ ID NO:22:
CGAGGGC Q G A-lCCG~lGCA GCAGCAACAT CACCGGTCTG 40 (2) INFORMATION FOR SEQ ID NO:23:
( i ) ~U~N~ CHARACTERISTICS:
IAI LENGTH: 242 base pairs I B, TYPE: nucleic acid ,C, STR~NDEDNESS: single ~DJ TOPOLOGY: linear WO 97/11086 PCTrUS96/15088 (Xi) ~U~:N~ DESCRIPTION: SEQ ID NO:23:
AACAT Q CCG GTCTGCTGCT GCTGCTGACC CGGACGGCGG CAAGGACACC ~ C~A~G 60 A~A~C~AAT CTTCCGCGAC GGCGGCAAGG A~AC~C~ ~A~C~A~ATC TTCCGCCCCG 120 TCGAGCCCCT GGGCGTGGCC CC~A~-~r-G CCAAGCGCGC GGTGGTGCAG CGCr-~ GC 240 (2) INFORMATION FOR SEQ ID NO:24:
( i ) ~U~N~ CHARACTERISTICS:
I'A'I LENGTH: 38 ba~e pairs ,BI TYPE: nucleic acid ,,CJ STR~NDEDNESS: ~ingle ~,D~ TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO:24:
CGCGGGCGGC CGCTTTAGCG ~ GCGC TGCACCAC 38 (2) INFORMATION FOR SEQ ID NO:25:
( i ) ~yU~N~ CHARACTERISTICS:
~Aj n ~: 39 ba~e pair~
~B TYPE: nucleic acLd C, STRANDEDNESS: single ~DJ TOPOLOGY: linQar (xi) ~:yu~ DESCRIPTION: SEQ ID NO:25:
CGCGGGGGAT CCAAGCTTAC CATGATTCCA GTA~TAA~-T 39 (2) INFORMATION FOR SEQ ID NO:26:
(i) SEQUENCE CHARACTERISTICS:
,'A' LENGTH: 165 ba~e pairq B TYPE: nucleic acLd ,'C STRANn~nNEss: ~ingle ~D,, TOPOLOGY: linear ( Xi ) ~ ~:QU~N~ DESCRIPTION: SEQ ID NO:26:
ATGAATCCAG TAATAAr-TAT AACATTATTA TTAAGTGTAT TA~A~ATGAG TAGAGGACAA 60 AGAGTAATAA GTTTAACAGC ATCTTTAGTA AATCA~AATT TGAGATTAGA TTGTAr-Ar~T 120 ~AAAATAATA CAAATTTGCC AATA~T GAATTTTCAT TAACG 165 ~ 54 ~
(2) INFORMATION FOR SEQ ID NO:27:
(i) ~ur_._~ CHARACTERISTICS:
/Aj LENGTH: 36 ba~e pair~
~B~ TYPE: nucleLc acid ,C, STRANn~nNESS: ~ingle ~D) TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO:27:
CGC&&GGAAT TCACGCGTTA ATGAaaATTC ATGTTG 36 ~2) INFORMATION FOR SEQ ID NO:28:
(i) S~yU~N~: CHaRACTERISTICS:
/AJ LENGTH: 30 ba~e pairs BI TYPE: nucleic acid CJ STRaNDEDNESs: ~ingle ~Dl TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID No:28:
CGCGGATCCA CGCGT&AaAA AAAAAAA~A~ 30 (2) INFORMATION FOR SEQ ID NO:29:
(i) S~U~N~ CHaRACTERISTICS:
~Al LENGTH: 149 ba~e pair-BJ TYPE: nucleic acid ~CJ S~RANn~nNESS: ~ingle D, TOPOLOGY: linear (Xi) S~:yU~N~ DESCRIPTION: SEQ ID NO:29:
CGT~-AAAAAA AAAAA~A~GT ATTAAGTGGA ACATTAGGAG TA~A~AA~A TA~ATATA~A 60 AGTAGAGTAA ~' LlVlL ~AGT GATA~-ATTcA TAaAAGTATT AACATTAGCA AATTTTACAA 120 ~-ATGA AGr-Ar-A~TAT A~ vAG 149 (2) INFORMATION FOR SEQ ID N0:30:
(i) SEQUENCE CHaRACTERISTICS:
(A'l LENGTH: 30 ba~e pair~
~Bl TYPE: nucleic acid STRANn~n~R~s: ~ingle ~DJ TOPOLOGY: linQar WO 97/11086 PCT~US96/15088 (xi) ~:QU~N~ DESCRIPTION: SEQ ID NO:30:
CGCGAATTCG AGCTr~r~A TATAATCTCC 30 (2) INFORMATION FOR SEQ ID NO:31:
(i) ~U~N~ CHARACTERISTICS:
'A'I LENGTH: 30 ba~e pair~
,B TYPE: nucleic acid ~C, STRANDEDNESS: ~ingle ~DJ TOPOLOGY: linear (xi) ~yu~w~ DESCRIPTION: SEQ ID NO:31:

(2) INFORMATION FOR SEQ ID NO:32:
(i) SEQUENCE CHARACTERISTICS:
'A~ LENGTH: 170 ba~e pair~
IBI TYPE: nucleic acid ,C, STRANDEDNESS: single ~D, TOPOLOGY: linear (xi) ~Ub~ : DESCRIPTION: SEQ ID NO:32:
CTCAGAGTAA GTGr-ArAAAA Tcr-A~r-AAGT AGTA~TAAA~ r~T~A~TGT AATAAr~rAT 60 A~ATTAGTAA AATGTGAGGA ATAAGTTTAT TAGT~r~AA~ TACAAGTTGG TTATTATTAT 120 (2) INFORMATION FOR SEQ ID NO:33~
( i ) ~U~N~ CHARACTERISTICS:
lA) LENGTH: 36 ba~e pair~
~B] TYPE: nucleic acid ,'Cj STRANDEDNESS: ~ingle ~D, TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO:33:
CGCGAATTCG CGGCCGCTTC ~TA~Ar-TTAT A~AATC36 (2) INFORMATION FOR SEQ ID NO:34:
(i) ~U~N~ CHARACTERISTICS:
(A) LENGTH: 1632 ba~e pair~
(B) TYPE: nucleic acid (C) STRANDEDNESS: ~ingle CA 0223l394 l998-03-09 (D) TOPOLOGY: linear (xi) ~u~w~ DESCRIPTION: SEQ ID NO:34:
CTCGAGATCC ATTGTGCTCT AaAGGAGATA CCCGGC Q GA CACCCT Q CC TGCGGTGCCC 60 AGCTGCCCAG GCTGAGGCAA GAGAAGGCCA ~-AAA~A~GC CCATGGGGTC TCTGCAACCG 120 CTGGCCACCT TGTACCTGCT GGGGATGCTG GTCGCTTCCG TGCTAGCCAC Cr7A~-AAGCTG 180 TGGGTGACCG TGTACTACGG CGTGCCCGTG TGGAAGGAGG crArrArr~C C~l~ GC 240 GC Q GCGACG CCAAGGCGTA c~ArAr~cr-Ac. GTGrArAArG TGTGGGCCAC CCAGGCGTGC 300 GTGCCrArCG ACCC~AA~CC CCAGGAGGTG GAG~ C~GA ACGTGACCGA GAACTTCAAC 360 ATGTGGAAGA Ar-AA~ATGGT GGAGCAGATG CATGAGGACA TCATCAGCCT GTGGGAC Q G 420 AGCCTGAAGC CCTGCGTGAA GCTGAr,CCCC CTGTGCGTGA CCCTGAACTG rArC~-ACCTG 480 Arr,AArAr,rA Cr-AAr-Ar~rAA CAACAG Q cc~Gc~AArAArA ACAG QA Q G CGAGGG Q CC 540 AT Q AGGGCG GCGAGATGAA CAACTG Q GC TTCAA QT Q CrArrA~7Q T CCGcnArAAn 600 ATGCAGAAGG AGTACGCCCT GCTGTACAAG CTGGATATCG TGAGCATCGA rAArrArAr.C 660 ACCAGCTACC GCCTGATCTC CTGrAArArC AGCGTGAT Q CCCAGGCCTG GCCrAAr-ATC 720 AAr,rArAArA AGTTCAGCGG CAAGGG Q GC TGrAA~-AArG TGAG Q CCGT GCAGTGCACC 840 Q CGGCATCC GGCCG~L~.. GAGrACCrAr. CTCCTGCTGA ACGGCAGCCT GGCCGAGGAG 900 GAGC..G~.~GA TCCGCAGCGA GAACTT Q CC r-ArAArGC Q ArArrAT QT CGTG Q CCTG 960 AATGAGAGCG TG Q GAT Q A CTGCACGCGT CCCAaCTACA A QAGCGCAA GCG Q TCCAC 1020 ATCGGCCCCG GGCGCGCCTT CTArArrACC AA~AArA~Q TCGGCACCAT CCGCCAGGCC 1080 CACTGCAA Q TCTCTAGAGC CAAGTGGAAC r-ArArCCTGC GC Q GATCGT GAG Q AGCTG 1140 AAGGAG Q GT TrAA~-AArAA GAC QTCGTG TTr~Arr~r-A G QGCGGCGG CnAr~Ccn~n 1200 ATCGTGATGC ACAGCTTCAA CTGCGGCGGC GAA.l~..~. ACTGCAACAC Q GCCCC~.G 1260 TT Q ACAGCA CCTGGAACGG ~AArAArA~c TG~-AArAArA C Q CCGGCAG ~AArAAr~A~ 1320 ATGTACGCCC CCCCr~TCGA GGGC Q GATC CGGTG Q GCA GCAACAT Q C CC~ .GCTG 1440 CTGACCCGCG ACGGCGGCAA G~-ArACCr-Ar, ArrAArrArA ccr-AAATCTT CCGCCCCGGC 1500 GAGCCCCTGG GCGTGGCCCC ~ACrAAGGCC AAGCGCCGCG TGGTGCAGCG Cr-A~-AAr-CGC 1620 TAaAGcGGcC GC 1632 W 0 97111086 PCTAJS96/l5088 (2) INFORMATION FOR SEQ ID No:35:
(i) ~:QU~N~ C~ARACTERISTICS:
(A', LENGT~: 2481 ba~e pair~
~BI TYPE: nucleic acid ,CJ STRANDEDNESS: ~ingle ~DJ TOPOLOGY: linear (xi) ~QU~N~ DESCRIPTION: SEQ ID NO:3S:
ArcGAr-AAGc TGTGGGTGAC CGTGTACTAC GGCGTGCCCG TGTGGAAGGA GGCCACCACC 60 ACC~ GCGCCAGCGA CGCCAAGGCG TA~r-Ar-ArCG AGGTGCACAA C~aGGCC 120 GAGAACTTCA ACATGTGGAA r-AAr-AAr-ATG CTGGAGCAGA TGCATGAGGA CATCATCAGC 240 CTGTGGGACC AGAGCCTGAA GCCCTGCGTG AAGCTGACCC CC~LGCGT GACCCTGAAC 300 TG~ACC~-A~C TGAGGAACAC rACr-AA~A~C AA~AA~AGCA CCGCrAAr-AA CAACAGCAAC 360 AGCGAGGGCA CCATCAAGGG CGGCGAGATG AAr-AAr,TGCA GCTTCAA Q T CACCACCAGC 420 rArAAr,rArA GCACCAGCTA CCGCCTGATC TCCTGCAACA CCAGCGTGAT r-ACCr-AroGCC 540 TGCCCrAA~-A TCAGCTTCGA GCCCATCCCC ATCCACTACT GCGCCCCCGC CGGCTTCGCC 600 ATCCTGAAGT G~AAcr-Ar-AA GAAGTTCAGC GGCAAGGGCA GCTGr-AAr~A CGTr-Arr~r~ 660 GTGCAGTGCA CCCACGGCAT CCGGCCG~LG GTGAGCACCC AGCTCCTGCT GAACGGCAGC 720 CTGGCCGAGG AGGAGGTGGT GATCCGCAGC GAGAACTTCA CCr-ArAACGC rAAr-Ar,rATC 780 ATCGTGCACC TGAATGAGAG CGTGCAGATC AACTGCACGC GTCCCAACTA rAArAAGCGC 840 AAGCGCATCC A Q TCGGCCC CGGGCGCGCC TTCTArACrA crAArAAr-AT CATCGGCACC 900 ATCCGCCAGG CCCACTGCAA CATCTCTAGA GCCAAGTGGA ACr-Ar-ACCCT GCGCCAGATC 960 GTGAGCAAGC TGAAGGAGCA GTTrAAr-AAC AAr-ACrATCG TGTTrAArrA GAGCAGCGGC 1020 GGCr-ArCCCG AGATCGTGAT GCACAGCTTC AACTGCGGCG GCGAATTCTT CTACTGCAAC 1080 ACCAGCCCCC TGTTCAACAG CACCTGGAAC GGrAAr-AArA CCTGGAAr-AA r-ArrACCGGC 1140 ArrAArAArA ATATTACCCT CCAGTGCAAG ATCAAGCAGA TCATCAACAT GTGGCAGGAG 1200 GTGGGCAAGG CCATGTACGC CCCCCC~:ATC GAGGGCCAGA l''CGG~GCAG CAGCAACATC 1260 ACCG~.~.GC TGCTGACCCG CGACGGCGGC AAGr-Ar-ACCG A~Ar,rAAr,r-A ~-A~Cr-AAATC 1320 GTGGTGACGA TCGAGCCCCT GGGCGTGGCC cCrAr,rAAr-G CCAAGCGCCG C~lGCAG 1440 CGCrAr-AAroC GGGCCGCCAT CGGCGCCCTG TTCCTGGGCT TCCTGGGGGC GGCGGGCAGC 1500 W O 97/11086 PCTAUS96/l5088 - ~8 -GTGCAGCAGC Ar-AAr-AAr,CT CCTCCGCGCC ATCGAGGCCC AGCAGCATAT GCTCCAGCTC 1620 ACC~vGG GCATCAAGCA GCTCCAGGCC CGCGTGCTGG CCGTGGAGCG CTACCTGAAG 1680 CCCTGGAACG C~~ vGAG rAArAAGAr,C CTGr-A~r-ArA TCTGr-AAr-AA CATGACCTGG 1800 ATGCAGTGGG AGCGCGAGAT Cr-ATAAr,TAC ACCAGCCTGA TCTACAGCCT GCTGGAGAAG 1860 AGCCAr-ArCC AGCAGGAGAA r-AArr-Ar-CAG GAGCTGCTGG AGCTGGACAA ~GGGCGAGC 1920 CTGTGGAACT G~.~GACAT rAr,r~ArTGG ~v~ ACA TCAAaATCTT CATCATGATT 1980 CGCCCCr-A~-G GCATCGAGGA GGAGGGCGGC GAGCGCGACC GCr-ArAr,rAr- CGGCAGGCTC 2160 GTGCACGGCT TCCTGGCGAT CA~ GG~,lC GACCTCCGCA GC~v~lC~ GTTCAGCTAC 2220 r~rr~CCGCG ACCTGCTGCT GATCGCCGCC CGCATCGTGG AACTCCTAGG CCGCCGCGGC 2280 ATCGAGGTGC TCCAGAGGGC CGGGAGGGCG ATCCTGCACA TCCCr-ACCCG CATCCGCCAG 2460 (2) INFORMATION FOR SEQ ID NO:36:
(i) SEQUENCE CHaRACTERISTICS:
'A! LENGTH: 486 ba3e pair~
BJ TYPE: nucleic acid C, sTRANnRnNR-~S: ~ingle ~D~ TOPOLOGY: linear (xi) ~yu~ DESCRIPTION: SEQ ID NO:36:
ATGAATCCAG TAA~AAr,TAT AACATTATTA TTAAGTGTAT ~Ar~AATGAG TAGAGGACAA 60 AGAGTAATAA GTTTAACAGC ATGTTTAGTA AATCAAAATT TGAGATTAGA TTGTAr-~rAT 120 ~AAAATAATA CACCTTTGCC AA~ArAArAT GAATTTTCAT TAACGCGTGA AAAAAAAAAA 180 CATGTATTAA GTGGA~CATT AGGAGTACCA rAArATArAT A~Ar-AAr,TAG AGTAaATTTG 240 TTTAGTGATA GATTCATAAA AGTATTAACA TTAGCAA~TT TTArAAr-~AA AGATGAAGGA 300 GATTATATGT GTGAGCTCAG AGTAAGTGGA CAAaATCCAA CaAGTAGTAA TAAAAr.~ATA 360 AATGTAATA~ ~ArA~AAATT AGTAaAATGT GGArrr-AA~AA GTTTATTAGT A~AAAATA~A 420 A~.~G~l~AT TATTATTATT ATTAAGTTTA AGll L L ~ AC AAG~AAr~rA TTTTATAAGT 480 W O 97/11086 PCT~US96/15088 (2) INFORMATION FOR SEQ ID NO:37:
~i) SEQUENCE CHARACTERISTICS:
Aj LENGTH: 485 base pair~
B~ TYPE: nucleic acid ~CJ STRANDEDNESS: 8ingle D,~ TOPOLOGY: linear (Xi) ~U~N~ DESCRIPTION: SEQ ID NO:37:
ATGAACCrAr- TCATCAGCAT CA~ G CTTTCAGTCT TGCAGATGTC CCGAGGACAG 60 AGGGTGATCA GCCTGACAGC ~GC~ G A~rAr-AACCT TCGACTGGAC TGCCGTCATG 120 ArAATAAr~r CAACTTGCCC ATC Q GCATG AGTTCAGCCT r~Arccr~ArAr~ A~r.AAr.AAr,C 180 ACGTGCTGTC AGGCACCCTG GGG~l~CCCG AGCACACTTA CCGCTCCCGC GTCAACCTTT 240 TCAGTGACCG CTTTATCAAG ~lC~.ACTC TAGCCAACTT r.ArrArrAAr. GATGAGGGCG 300 ATGTGATCAG ArArAAr-CTG GTCAAGTGTG GTGGCATAAG CCTGCTGGTT rAAA~rAr,TT 420 CCTGGCTGCT GCTGCTCCTG ~l'lCC~. C~L ~ C~A AGCCACGGAC TTCAI,.~ C 480 (2) lN~OKhATION FOR SEQ ID NO:38:
(i) ~QU~N~ CHARACTERISTICS:
(A~ LENGTH: 33 ba~e pair~
~BJ TYPE: nucleic acid ~CJ STRANn~nN~S: 8ingle ~DJ TOPOLOGY: linear (ii) MnT~T~'CUT~ TYPE: DNA (genomic) (xi) SEQUENCE DESCRIPTION: SEQ ID NO:38:
CGCGGGGCTA GcGr-AAAr~Aq ~AATAAr,TTT AAC 33 (2) INFORMATION FOR SEQ ID NO:39:
(i) ~u~..CE CHARACTERISTICS:
~A'l LENGTH: 28 ba~e pair~
~Bl TYPE: nucleic acid ,C, STRANDEDNESS: single ~D,l TOPOLOGY: linear (ii) MOLECULE TYPE: DNA (genomic) (Xi) ~yU~N~ DESCRIPTION: SEQ ID NO:3g:

(2) INFORMATION FOR SEQ ID NO:40.
(i) ~yu~ CHARACTERISTICS:
A'I LENGTH: 762 base pair~
B~ TYPE: nucleic acid ,C, STRaNDEDNEss: ~ingle ,D, TOPOLOGY: linear (ii) M~T~T~'CUT~T~' TYPE DNA (9~~ i~) (xi) SEQUENCE DESCRIPTION: SEQ ID NO:40:
GAATTCACGC GTAAGCTTGC CGC~TG GTGAGCAAGG GCGAGGAGCT GTT~A~CGGG 60 GTGGTGCCCA TCCTGGTCGA GCTGGACGGC GACGTGAACG Gc~r~ArrTT CAGC~L~lCC 120 GGCAAGCTGC CCGTGCCCTG GCC~CCCTC GT~-~C~CCT TCAGCTACGG CGTGCAGTGC 240 TTCAGCCGCT ACCCC~-AC~ CATGAAGCAG CACGACTTCT TCAAGTCCGC CATGCCCGAA 300 GGCTACGTCC AGGAGCGCAC CA. ~ C AAG~-ACr-~G GCAACTACAA GACCCGCGCC 360 GAGGTGAAGT TCGAGGGCGA CAcc~la~G AACCGCATCG AGCTGAAGGG CATCGACTTC 420 AA~G~r~r~G GCAACATCCT GGGGCACAAG CTGGAGTACA ACTA~- C~A~AA~GTC 480 TATAT QTGG CC~-A~AGCA ~-~A~-~A~GGC ATCAAGGTGA ACTTCAAGAT CCGC~AAC 540 ATCr-~r~-~G GCAGCGTGCA GCTCGCCGAC CACTACCAGC Ar-~ccc CATCGGCGAC 600 GGCCCC~LaC TGCTGCCCGA ~AC~CTAC CTGAGCACCC AGTCCGCCCT GAGr-AAA~-Ac 660 CC~A~-A~-~ AGCGCGATCA CATGGTCCTG CTGGAGTTCG TGACCGCCGC CGGGATCACT 720 CACGGCATGG ACGAGCTGTA CAAGTA~AGC GGCCGCGGAT CC 762

Claims (12)

What is claimed is:

1. A synthetic gene encoding a protein normally expressed in a eukaryotic cell wherein at least one non-preferred or less preferred codon in the natural gene encoding said protein has been replaced by a preferred codon encoding the same amino acid.

2. The synthetic-gene of claim 1 wherein said synthetic gene is capable of expressing said eukaryotic protein at a level which is at least 110% of that expressed by said natural gene in an in vitro mammalian cell culture system under identical conditions.

3. The synthetic gene of claim 1 wherein said synthetic gene is capable of expressing said eukaryotic protein at a level which is at least 150% of that expressed by said natural gene in an in vitro cell culture system under identical conditions.

4. The synthetic gene of claim 1 wherein said synthetic gene is capable of expressing said eukaryotic protein at a level which is at least 200% of that expressed by said natural gene in an in vitro cell culture system under identical conditions.

5. The synthetic gene of claim 1 wherein said synthetic gene is capable of expressing said eukaryotic protein at a level which is at least 500% of that expressed by said natural gene in an in vitro cell culture system under identical conditions.

6. The synthetic gene of claim 1 wherein said synthetic gene is capable of expressing said eukaryotic protein at a level which is at least ten times that expressed by said natural gene in an in vitro cell culture system under identical conditions.

7. The synthetic gene of claim 1 wherein at least 10% of the codons in said natural gene are non-preferred codons.

8. The synthetic gene of claim 8 wherein at least 50% of the codons in said natural gene are non-preferred codons.

9. The synthetic gene of claim 1 wherein at least 50% of the non-preferred codons and less preferred codons present in said natural gene have been replaced by preferred codons.

10. The synthetic gene of claim 1 wherein at least 90% of the non-preferred codons and less preferred codons present in said natural gene have been replaced by preferred codons.

11. The synthetic gene of claim 1 wherein said protein is green fluorescent protein.

12. A method for preparing a synthetic gene encoding a protein normally expressed by eukaryotic cells, comprising identifying non-preferred and less-preferred codons in the natural gene encoding said protein and replacing one or more of said non-preferred and less-preferred codons with a preferred codon encoding the same amino acid as the replaced codon.