Novel Biotinylated Plasmid Expression Vectors Retain Biological

Aug 1, 1996 - Function and Can Bind Streptavidin. Patrick Leahy,*,† Gordon G. Carmichael,‡ and Edward F. Rossomando†. Department of BioStructure...
0 downloads 0 Views 212KB Size
Download PDF
Bioconjugate Chem. 1996, 7, 545−551

545

ARTICLES Novel Biotinylated Plasmid Expression Vectors Retain Biological Function and Can Bind Streptavidin Patrick Leahy,*,† Gordon G. Carmichael,‡ and Edward F. Rossomando† Department of BioStructure and Function and Department of Microbiology, University of Connecticut Health Center, Farmington, Connecticut 06030. Received August 3, 1995X

A new method for coupling proteins to plasmid expression vectors is presented. Biotin was covalently attached to a plasmid expression vector containing a chloramphenicol acetyltransferase (CAT) gene. The specific label was one biotin per 100 bp. An electrophoretic mobility shift assay showed that the plasmid was capable of binding multiple streptavidin molecules. When transfected into mouse fibroblasts, the biotinylated plasmid retained 40% of the native plasmid’s biological activity, as determined by CAT assay, and was not affected by the binding of streptavidin. The method allows for attachment of any protein to plasmid DNA expression vector while retaining biological function. Hybrid plasmids in which the transcription cassettes were kept free of biotin label were constructed by digesting biotinylated and unbiotinylated plasmids at sites outside the transcription cassette and re-ligating the digestion products. Electron microscopy studies show that the ligation products formed large tangled assemblages of plasmid DNA. When equimolar (with respect to gene number) amounts of these large hybrid biotinylated plasmids were transfected into mouse fibroblasts by means of calcium phosphate precipitation, an increase in CAT expression 25-fold greater than that of original biotinylated plasmid was observed. Slot-blot analysis of total DNA extracted from transfected cells shows that this enhanced activity was not due to increased transfection efficiency. Receptor-mediated delivery could not be shown when a complex comprising biotinylated asialoglycoprotien/streptavidin/biotinylated CAT expression vector was placed in media containing Hep G2 cells.

INTRODUCTION

In the past decade, the field of receptor-mediated gene delivery as a means of achieving gene therapies has attracted a lot of research interest. The strategy takes advantage of the affinity of cell surface receptors for their natural ligands to effect delivery of genes to target tissues in vivo (1). Generally, the bioconjugate molecules used to achieve this goal have two components: a gene expression vector and a ligand (usually a protein) with specificity for binding receptors on target cells. The chemistry most often used to join the DNA and ligands is based on the electrostatic attraction between positively charged polylysine and negatively charged DNA (2, 3). The ligand is first covalently bound to polylysine, and the resulting protein conjugate is allowed to interact electrostatically with the nucleic acid component. The complex formed is then used to transfect specific cells in vivo by means of receptor-mediated endocytosis. A number of problems inherent to the use of polylysine have emerged: widspread heterogeneity of the protein conjugate (4), tedious quality control measures (5), insolubility (6), and variability in the efficacy of complexes to effect delivery of genes to tissues posessing the receptors (2, 7). In view of these difficulties, it was * Address correspondence to this author at the Department of Biochemistry, Case Western Reserve University, 10900 Euclid Ave., Cleveland, OH 44106 [telephone (216) 368-3634; fax (216) 368-4544; e-mail [email protected]]. † Department of BioStructure and Function. ‡ Department of Microbiology. X Abstract published in Advance ACS Abstracts, August 1, 1996.

S1043-1802(96)00044-4 CCC: $12.00

decided to explore alternative methods for attaching protein ligands to expression vectors. There are a number of methods that allow for the covalent attachment of proteins to nucleic acids (8-11), but these are concerned almost exclusively with the crosslinking of intimate RNA-protein associations and are not useful as general coupling protocols. A more general method that allows for the covalent attachment of any protein to nucleic acids has been reported (12) and describes the covalent coupling of R2-macroglobulin to a plasmid DNA chloramphenicol acetyltransferase (CAT) expression vector. However, the authors do not show any data to indicate that the modified expression vector could function in a biological setting. Recently it was speculated that plasmid expression vectors with covalently attached exogenous material would be transcriptionally inactive or give rise to altered transcription (13, 14). While this is somewhat intuitive, there are no experimental results to date to support such speculation. In the present study we revisit this problem and ask the question whether or not it is possible to covalently attach exogenous material to an expression vector and retain transcriptional activity. To answer this question, biotin was covalently bound to plasmid expression vectors by means of photoactivation (15). The biotinylated plasmid was then tested for two important functions: its ability to bind streptavidin and, following transfection into cells, its ability to function as a template encoding functioning CAT enzyme. EXPERIMENTAL PROCEDURES

Expression Vector. Our expression vector was a 3871 bp plasmid. The transcription cassette occupies 997 © 1996 American Chemical Society

546 Bioconjugate Chem., Vol. 7, No. 5, 1996

bp (25.7% of total) and comprised an RNA polymerase II promoter, a CAT gene, and an SV40 polyadenylation signal. Two NspI restriction sites flank the transcription cassette, and a unique BsaI restriction site also lies outside the transcription cassette. These restriction sites play an important role in making hybrid plasmid constructs (see below). Biotinylation of Plasmid Expression Vector. Intact plasmid (500 µg) was prepared in distilled deionized water at a concentration of 1 µg/µL. In a darkroom, 0.5 mg of Immunopure photobiotin (Pierce Chemical Co.) was reconstituted in 500 µL of distilled deionized water. Equal volumes of DNA and photobiotin were mixed in a 1.5 mL polypropylene microfuge tube. The open-capped sample was then subjected to eight camera flashes from a distance of approximately 15 cm above. The sample was made up to 15 mL with TE buffer (10 mM Tris-Cl, pH 7.4, 1 mM EDTA) and was washed repeatedly with TE buffer in a Centriprep 10 microconcentrator (MW cutoff ) 10 000; Amicon, Lexington, MA) to remove unbound photobiotin. Quantification of Biotinylation. The degree of biotinylation of the plasmid was quantified by means of a fluorescence assay. Fluorescein-labeled avidin (200 µg, 3.3 nmol; Avidin Neutralite-FL, Molecular Probes Inc., Junction City, OR) was added to 100 µg (38 pmol) of biotinylated plasmid. The reaction components were allowed to interact for 1 h at 37 °C. The free avidin was separated from the bound on a 10 mL size exclusion chromatography column using Bio-Rad A 1.5 gel equilibrated with TE buffer. Ten fractions of 0.5 mL were collected. The first DNA fraction eluted was quantified. Avidin content was gauged using a fluorescence spectrophotometer. A standard curve of fluorescein-labeled avidin was prepared (range: 0-200 ng/mL.). The excitation beam was set to 490 nm with an 8 nm slit, and the emitted light was detected at 520 nm. There were 40 biotins per plasmid, which represents approximately 1 biotin/100 bp of DNA Electrophoretic Mobility Shift Assay (EMSA). The ability of biotinylated plasmid to bind streptavidin was tested. Samples of biotinylated and unbiotinylated plasmid (2 µg; 0.79 pmol) were prepared in 6 µL of TE buffer (10 mM Tris-Cl, pH 7.5, 1 mM EDTA). For the biotinylated plasmid, this represented 31.8 pmol of biotin. A 1µL aliquot of a 10 mg/mL streptavidin stock (167 pmol, representing a 5.25 molar excess with respect to the biotin) was added to the samples. Samples were mixed and incubated at 37 °C for 1 h. Tracking dye was added, and the electrophoretic migration patterns of the samples were analyzed on 0.8% agarose gels in trisacetate-EDTA (TAE) buffer. Transfections. A modified calcium phosphate precipitation method (16) was used to transfect the biotinylated plasmid into mouse fibroblasts (NIH 3T3 cells). Four replicate 100 mm plates of 30-40% confluent cells were transfected with 10 µg of biotinylated plasmid. Controls plates were transfected with 10 µg of unbiotinylated plasmid or with buffer alone. The cells were given fresh medium at 24 h after transfection. Cells were harvested at 48 h and cell extracts prepared. When transfection with hybrid concatemer plasmids was performed, 10 µg of DNA was used. This represented an equimolar amount of gene cassettes. Preparation of CAT Extracts. Tissue culture plates were drained, and cells were harvested with scrapers in 1 mL of ice-cold phosphate-buffered saline (PBS: 10 mM phosphate, 150 mM NaCl). The cells were pelleted by centrifuging at 2500g for 5 min. The supernatant was removed, and the cells were resuspended and washed

Leahy et al.

again in 4 mL of PBS. The cells were pelleted again and the supernatants removed. The cells were lysed by resuspending the pellet in 0.5 mL of 0.25 M Tris, pH 7.4/ 0.5%Triton X-100 and placing on ice for 10 min. The lysate was centrifuged for 2 min at 10000g at 4 °C and then placed in a water bath at 65 °C for 15 min. Samples were centrifuged again for 5 min at 10000g at 4 °C to repellet the lysate debris. The resulting supernatants were the CAT extracts and were reserved on ice. Aliquots of the CAT extracts were reserved for determination of protein concentration using the BCA method (Pierce). Assay for CAT Activity. CAT reaction buffer was prepared just prior to use: 12.5 mM chloramphenicol, 25mM Tris-Cl, pH 7.4, 0.25 µCi of [3H-acetyl]acetyl-CoA (200 mCi/mmol) or 25 µM acetyl-CoA. Since chloramphenicol is reconstituted in ethanol, the buffer also contained 10% ethanol. A 50 µL aliquot of the CAT reaction buffer was placed in a 7 mL scintillation vial. The assay was started by adding 75 µL of CAT extract to this buffer to give a final reaction volume of 125 µL. This reaction mixture was overlaid with 3 mL of waterimmiscible “Econofluor” scintillant (NEN, Boston, MA). Samples were immediately placed in a liquid scintillation counter. Radiolabeled acylated forms of chloramphenicol diffuse rapidly into the upper phase of scintillant. The radioactivity (c.p.m.) was read at regular intervals over a 2-3 h period. CAT activity was determined from the slopes of the early (linear) portions of these time curves. All CAT activities were normalized for protein concentration. Plasmid Restriction Digests and Ligations. Restriction enzyme digests were carried out according to the manufacturer’s instructions. After digestion with NspI, the enzyme was inactivated by incubation at 70 °C for 15 min. Without further treatment, the digests were used in a series of ligation reactions carried out in volumes of 100 µL, incubated at 16 °C overnight. The ligation reaction components were 100 µg of NspIdigested plasmid (152 pmol-ends), 50 mM Tris-Cl, pH 7.4, 10 mM MgCl2, 1 mM ATP, 1 mM DTT, and 2 units of T4 DNA ligase. These high concentrations of pmol-ends and ligase are conducive to intermolecular ligations. Ligation was confirmed by comparing electrophoretic banding patterns of the DNA on 0.8% agarose TAE gels before and after the ligation procedure. Electron Microscopy. Samples of the CAT plasmid ligation products were prepared for examination by transmission electron microscopy according to the method of Laemmli (17). Formvar-carbon-coated specimen grids were glow-discharged and coated with 1 µg/mL poly(Llysine) (MW > 300 000; Sigma Chemical Co.). The grids were then floated on 30 µL drops of sample on parafilm for 30 min, rinsed for a total of 1 min on 4 drops of distilled water, and then stained on a drop of 2% aqueous uranyl acetate for 1 min. The excess uranyl acetate was wicked off each grid with filter paper, leaving the grid wet with a small amount of stain. After drying, the grids were observed and photographed at 60 kV in a JEOL 100CX TEM. Magnification was calibrated with a diffraction grating replica (2180 lines/mm). DNA Slot-Blot Analysis. Transfection efficiency was determined by extracting total DNA from lysates of cells that had been transfected either with native plasmid or with a large-sized ligation product derivative of the plasmid. Total DNA was extracted according to a standard phenol/chloroform-based protocol (18). DNA samples (10 µg) and hybridization buffer were prepared according to the instructions of the manufacturers of Genescreen Plus hybridization membrane (DuPont, Boston, MA). Samples were loaded onto the membrane using

Biotinylated Plasmid Expression Vectors

Bioconjugate Chem., Vol. 7, No. 5, 1996 547

Figure 2. Comparison of biological activity of biotinylated and native plasmid. Replicate 100 mm plates of 30-40% confluent mouse fibroblast (3T3) cells were transfected with 10 µg of either biotinylated CAT plasmid or unbiotinylated CAT plasmid using a calcium phosphate precipitation method. A control group were transfected with buffer alone (n ) 4 in all cases). Cells were harvested 48 h after transfection, and the CAT activities of the extracts were determined (n ) 4 in all cases). The results were normalized for protein content.

Figure 1. Electrophoresis mobility shift assay confirming the presence of biotin on the plasmid expression vector. Samples of the CAT plasmid (1 µg) were subjected to electrophoresis on an 0.8% agarose TAE (tris-acetate-EDTA) gel. All samples were preincubated for 15 min at 37 °C in the presence or absence of excess (10 µg) streptavidin. Lane 1, unbiotinylated plasmid; lane 2, unbiotinylated plasmid with streptavidin; lane 3, 1 kb size markers (BRL); lane 4, biotinylated plasmid; lane 5, biotinylated plasmid with streptavidin. The location of the 1 kb size marker band is indicated by the black triangle.

a Hybri-slot manifold apparatus (BRL). 32P-Labeled CAT ribozyme probe was generated according to methods we have already published in detail elsewhere (19). Prehybridization was for 1 h at 55 °C, and hybridization was allowed to run overnight at 55 °C. Posthybridization washes were carried out in accordance with the instructions of the manufacturers of the membrane. The membrane was monitored between washes with a Geiger counter and, when background signal was sufficiently reduced, was then exposed to Kodak film for 2 h. RESULTS

Characterization of the Biotinylated Plasmid. A 3871 bp plasmid expression vector for the CAT gene was biotinylated and the unbound photobiotin removed. The specific biotin label was determined by means of a fluorescence assay. The molar ratio of biotin to plasmid was 40. This represents approximately 1 biotin/100 bp of DNA. The Biotinylated Expression Vector Binds Streptavidin. The ability of the plasmid to bind streptavidin was confirmed using an EMSA. Biotinylated and unbiotinylated plasmids were incubated in the presence of a molar excess of streptavidin. The electrophoretic migration patterns of the DNA were analyzed on 0.8% agarose gels. There was a 100% shift in the banding pattern of the biotinylated plasmid, indicating that all plasmid molecules possessed biotin label. The mobility shift was attributed to the binding of multiple streptavidin molecules (Figure 1, lane 5). The mobility of the native plasmid was not affected by the presence of streptavidin (Figure 1, lane 2). Biotinylated Expression Vector Remains Biologically Active. The ability of the biotinylated expression

vector to function within mammalian cells as a template encoding fully active CAT enzyme was tested. The biotinylated plasmid was transfected into mouse fibroblasts (NIH 3T3 cells) using a calcium phosphate precipitation method. Control groups were transfected with native plasmid or with buffer alone. Cells were assayed for CAT activity 2 days later. The biotinylated plasmid was biologically active, retaining approximately 40% of its original activity (Figure 2). It is improbable that this amount of CAT activity could be attributed to a fraction of the plasmid population being unbiotinylated since the EMSA (Figure 1) shows that all of the plasmids possessed biotin. Further, with 40 randomly attached biotins per plasmid, and a transcription cassette occupying 25.7% of the plasmid, the percentage of expression vectors having biotin-free transcription cassettes is (1-0.257)40 × 100% ) 0.0007%. Therefore, the results show that eucaryotic RNA polymerase II was capable of reading through those parts of the duplex DNA bearing the covalently attached biotins. More generally, it is concluded that expression vectors, with multiple covalently attached biotins, are able to continue functioning as templates encoding biologically active proteins. Biotinylated Expression Vector Remains Biologically Active When Complexed with Streptavidin. To attach proteins, these biotinylated plasmids must be complexed with streptavidin-conjugated ligands. An important question then arises. Is transcription possible from biotinylated template when streptavidin (60 kDa) is complexed? To answer this, our biotinylated expression vector was complexed with streptavidin prior to calcium phosphate transfection into 3T3 cells. The results show no significant difference in biological activity between biotinylated expression vectors either in the presence or in the absence of bound streptavidin (Figure 3). Construction of Plasmids with Biotin-Free Transcription Cassettes. Clearly, if one could keep the transcription cassette of the expression vector free of biotin, it should be possible to improve the level of expression from the plasmid. A scheme for the production of plasmid constructs with biotin-free transcription cassettes was designed (Figure 4). Biotinylated and unbiotinylated plasmids were digested separately with NspI to yield fragments of approximately 2 kb in size. Equimolar amounts of these fragments were cross-ligated

548 Bioconjugate Chem., Vol. 7, No. 5, 1996

Figure 3. Comparison of biological activity of biotinylated expression vector with and without complexed streptavidin. Biotinylated CAT plasmid was complexed with a 40 times molar excess of streptavidin. Replicate 100 mm plates of mouse fibroblasts were transfected with 10 µg of the complex using the calcium phosphate precipitation method. Control groups were transfected with unbiotinylated (native) plasmid and with uncomplexed biotinylated plasmid (n ) 4 in all cases). Cells were harvested 48 h after transfection, and the CAT activities of the extracts were determined (n ) 4 in all cases). The results were normalized for protein content.

Leahy et al.

Figure 5. Characterization of the DNA constructs. Samples of DNA constructs (2 µg) were electrophoresed on an 0.8% agarose TAE gel together with 1 kb size markers (BRL). Lane 1, 1 kb size markers (BRL); lane 2, homopolymer 1; lane 3, homopolymer 2; lane 4, heteropolymer 1. The location of the 1 kb size band is indicated by the black triangle

Figure 6. Biological activity of the new plasmid ligation products. Replicate 100 mm plates of mouse fibroblasts were transfected with 10 µg of heteropolymer 1, homopolymer 1, and homopolymer 2, respectively. A control group was transfected with 10 µg (an equimolar amount with respect to gene number) of native CAT plasmid. Cells were harvested 48 h after transfection, and the CAT activities of the extracts were determined (n ) 4 in all cases). The results were normalized for protein content.

Figure 4. Scheme for the production of biotinylated plasmid expression vectors with biotin-free transcription cassettes. Native and biotinylated CAT plasmid samples are digested with NspI (Amersham). Equimolar amounts of biotinylated and unbiotinylated fragments are cross-ligated to produce very large concatemers termed “heteropolymer 1”. Heteropolymer 1 contains biotin label and many biotin-free transcription cassettes on the same molecule. Homopolymer 1 and homopolymer 2 are produced under similar ligation conditions using only unbiotinylated or biotinylated fragments, respectively.

at high concentration (1 µg/µL DNA or greater) to facilitate intermolecular ligation (20). The electrophoretic migration pattern of cross-ligation products were analyzed on an 0.8% agarose TAE gel. A single broad band migrated to the rear of the largest (12 kb) of the linear DNA size markers (Figure 5A, lane 4). It was not possible to say whether the products were linear or

circular. The ligation product was termed “heteropolymer 1” to indicate its composition of both biotinylated and unbiotinylated fragments and to reflect its large size. Calcium phosphate transfection of 10 µg of heteropolymer 1 into 3T3 cells gave unanticipated results. The biological activity was 10 times greater than that achieved with a similar mass (equimolar number of genes) of the native plasmid (Figure 6). This represents an even greater improvement (25-fold) on the biological activity of the original biotinylated plasmid (Figure 2). Characterizing the Enhanced Biological Activity. To ascertain whether the increased activity found in heteropolymer 1 was attributable to the presence of biotin, which could possibly interfere with normal cellular exonuclease degradative activity, or to its size and shape, which are known to affect degradation of nucleic acids (4, 21), two separate ligations of unbiotinylated and biotinylated NspI-generated fragments were made to produce “homopolymer 1” and “homopolymer 2”, respectively (Figure 4). The two product sizes were similar to each other and to heteropolymer 1 (Figure 5A, lanes 2

Bioconjugate Chem., Vol. 7, No. 5, 1996 549

Biotinylated Plasmid Expression Vectors

Figure 8. Slot-blot analysis of 10 µg total DNA extracted from cells transfected with native plasmid and cells transfected with homopolymer 1. Lane 1, positive control (5 ng of native CAT plasmid) ; lane 2, mock transfection; lane 3, cells transfected with native plasmid; lane 4, cells transfected with homopolymer 1; lane 5, positive control (5 ng homopolymer 1).

Figure 7. Electron micrograph of homopolymer 1. The native plasmid was digested with NspI. The enzyme was heat inactivated, and the digestion products were used in an overnight ligation reaction. See Experimental Procedures for sample preparation.

and 3). When equimolar amounts (10 µg) of these concatemer products and native plasmid were transfected into 3T3 cells, homopolymer 1 showed the same enhanced activity as heteropolymer 1. Homopolymer 2 (with biotin) had only a fraction of the biological activity shown by homopolymer 1. Since homopolymer 1 lacked biotin but possessed the enhanced activity, it was concluded that enhanced activity was attributable to the size and/or shape of the DNA. The same size enhancement of biological activity was observed when one compared the increased CAT activity of homopolymer 2 (Figure 6) with that of its precursor, biotinylated plasmid (Figure 2). Again there was an approximate 10-fold increase in biological activity. We also transfected plates of mouse fibroblasts with native and concatemer plasmids using lipofection (data not shown). We did not see the enhanced activity. We feel this may be due to the inherent problems of this particular transfection system: the product literature cautions the user about finding the proper molar ratio of DNA and lipofectin reagent (BRL). Electron microscopy studies were undertaken to see whether any three-dimensional properties were evident. The DNA appeared as tangled aggregates of plasmid DNA with no other discernible features (Figure 7). The question arose as to whether the increased activity seen in cells transfected with large-sized ligation products is attributable to an increase in transfection efficiency or to increased levels of transcription from the large polymeric form of the plasmid. To address this question, slot-blot analysis of total DNA extracts of lysates of cells that had been transfected with native plasmid and cells transfected with homopolymer 1 were carried out to quantify the amount of plasmid in the transfected cells. These cells had already been assayed for CAT activity, and the results were 17 507 and 90 093 cpm min-1 (mg of protein)-1, respectively (a 5-fold difference). Our results show there was no difference in transfection efficiency between native plasmid and homopolymer 1 (Figure 8, lanes 3 and 4). By using two positive controls, we also show that the hybridization efficiency of the probe is the same for both forms of the plasmid (Figure 8, lanes 1 and 5). Since transfection efficiency was ruled out, the 5-fold increase in CAT activity observed with cells transfected with homopolymer 1 must be due to another process such as an increase in transcription rate or some differential intracellular distribution of the transfected product leading to greater accumulation of the transfected concatemer plasmid DNA in the nucleus.

Receptor-Mediated Delivery Trials. Some trials were undertaken to assess whether streptavidin bridges linking biotinylated protein ligands and biotinylated CAT expression plasmids could be used to transfect cells by receptor-mediated endocytosis (data not shown). Biotinylated (ratio 1:1) asialo-orosomucoid (AsOR)was added to the streptavidin-biotin-CAT plasmid complex. The theoretical final AsOR/streptavidin/plasmid molar ratio was 120:40:1. The complex (equivalent to 10 µg of plasmid DNA) was added to 100 mm plates of 30-50% confluent Hep G2 cells. The latter cell type is known to express an abundance of receptors for AsOR on the cell surface. We could not show any delivery. CAT activity was never greater than that found in control plates which received no DNA. DISCUSSION

The first report of a general method for covalent attachment of any protein to nucleic acids appeared in 1983 (12). R2-Macroglobulin was covalently linked to a plasmid expression vector at a protein/DNA molar ratio of 2:1. While the authors claimed that there was no apparent damage to the structure of the DNA, they did not present any data pertaining to this conjugate’s ability to function in a biological setting. More recently, it has been speculated that covalent coupling of ligands to the DNA would cause transcriptional inactivation or be mutagenic (13, 14). Again, no data were presented to support these speculations. In this study we have shown that it is possible to get substantial amounts of biological activity (40% of the native activity) from a DNA template with covalently bound biotin. Our data suggest that RNA polymerase II is able to accommodate the presence of these biotins on the plasmid template during transcription. We have also shown that biotinylated plasmids that have been complexed with streptavidin remain biologically functional when transfected into mouse fibroblasts. It is possible that RNA polymerase II can displace biotinbound streptavidin from a biotinylated template. In a related report, it has been shown that bacteriophage RNA polymerase had the strength to displace up to 80% of linear DNA template fragments that were attached to streptavidin by means of an end-labeled biotin (22). The ability to bind streptavidin also implies that any biotinylated protein may be added to the conjugate to form a tripartite complex: biotinylated DNA-streptavidinbiotinylated protein. Making use of avidin-biotin technology to attach protein ligands allows great flexibility in the number of ligands one may attach to the plasmid. Current methods for biotinylation of DNA allow for a great deal of control over the number of ligand binding sites (biotins) that can be placed on a plasmid. We labeled our plasmid expression vector by means of photobiotinylation (Pierce Chemical), which resulted in the covalent attachment of approximately 40 biotins per plasmid. Streptavidin has a valency of four, which means the unoccupied sites may

550 Bioconjugate Chem., Vol. 7, No. 5, 1996

accommodate as many as 120 biotinylated ligands, making a ligand/DNA molar ratio of 120. Another method using M13 ssDNA, dNTPs-biotin, and Klenow to generate plasmid expression vectors with very high specific biotin label has been reported by us (23). We also showed that the loss of biological activity incurred by covalent derivatization of a plasmid expression vector may be overcome by the construction of large hybrid plasmids in which a large proportion of the transcription cassettes are kept unmodified. This method worked well for us; it could be used if warranted in the R2-macroglobulin situation mentioned earlier (12) and has general applicability in all situations when loss of activity of an expression vector is incurred. The increased transcription rate from the large ligation products was an unanticipated result. The increase may be due to the rearrangement of position-sensitive upstream activation sequences, promoters, or enhancers. There are no precedents for the assemblages presented here. However, Axelrod et al. (24) have shown that tandem repeats of promoters placed upstream of transcription start sites in plasmid expression vectors can produce nonlinear synergistic increases in transcription levels sometimes up to 4 orders of magnitude in excess of the levels in the native plasmid possessing a single promoter. It may be that the increased transcription levels presented here are due to some sort of synergistic effect produced by the chance arrangement of NspI-generated fragments and/ or the proximity of activation elements in the ligation products. Our failure to achieve receptor-mediated delivery is difficult to explain. Ligand/DNA molar ratios are unlikely to be the cause. Our ratio of 120:1 is well above what is considered the minimum requirement (10-15: 1) and comfortably within the range of recently published ratios for polylysine-based methods (25, 26). The failure to transfect may be due to the packaging of DNA. Polylysine, in addition to binding electrostatically to DNA, also causes condensation of the DNA into particles of definite size and shape (2). It may be that such condensation is a prerequisite for receptor-mediated delivery. Other applications for biotinylated expression vectors could be manifold. Nuclear localization signals (NLSs) comprising short, highly basic amino acid sequences have been implicated in the import of the SV40 large-T antigen (27) and another proteinsnucleoplasmin (28)sinto the nucleus. Recently, synthetic peptides containing putative NLSs, conjugated to bovine serum albumin (BSA), were shown to facilitate the translocation of BSA to the nucleus (29). By coupling these same sequences to an expression vector containing a sensitive reporter gene, one would have an assay system capable of providing important information on the relative strengths of NLSs. The cellular fate of transfected expression vectors could also be explored by conjugating biotinylated plasmids with commercially available fluorescein-labeled streptavidin or with antibodies. Subcellular fractionation and/ or in situ immunohistological stains might then be used to substitute for the more tedious in situ hybridizations. In conclusion, plasmid expression vectors with covalently attached biotins continue to function as templates for transcription in eucaryotic cells. By manipulation of biotinylated and unbiotinylated plasmid fragments, hybrid plasmid constructs of large size were made that are unique in possessing a combination of three valuable features: (1) unhindered transcription cassettes, (2) numerous streptavidin binding sites, and (3) enhanced biological activity.

Leahy et al. ACKNOWLEDGMENT

This research was supported by a grant from the Patrick and Catherine Welden Donaghue Medical Research Foundation (93-089). We gratefully acknowledge the assistance given by Dr. Jonathan Clive at the Office of Biostatistical Consultation at the University of Connecticut Health Center. The electron microscopy was performed by Ms. Christine Pearson of the Central Electron Microscope Facility, University of Connecticut Health Center. LITERATURE CITED (1) Mulligan, R. (1993) The basic science of gene therapy. Science 260, 926-932. (2) Perales, J. C., Ferkol, T., Beegan, H., Ratnoff, O. D., and Hanson, R. W. (1994) Gene transfer in vivo: sustained expression and regulation of genes introduced into liver by receptor targeted uptake. Proc. Natl. Acad. Sci. U.S.A. 91, 4086-4090. (3) Chowdhury, N. R., Wu, C. H., Wu, G. Y., Yerneni, P. C., Bommineni, V. R., and Chowdhuri, J. R. (1993) Fate of DNA targeted to the liver by asialoglycoprotein receptor mediated endocytosis in vivo. J. Biol. Chem. 268, 11265-11271. (4) Ferkol, T., Lindberg, G. L., Chen, J., Perales, J. C., Crawford, D. R., Ratnoff, O. D., and Hanson, R. W. (1993) Regulation of phosphoenolpyruvate carboxykinase/humanfactor IX gene introduced into the livers of adult rats by receptor mediated gene transfer. FASEB J. 7, 1081-1091. (5) McKee, T. D., DeRome, M. E., Wu, G. Y., and Findeis, M. A. (1994) Preparation of asialoorosomucoid-polylysine conjugates. Bioconjugate Chem. 5, 306-311. (6) Wu, G. Y., and Wu, C. H. (1988) Evidence for targeted gene delivery to Hep G2 hepatoma cells in vitro. Biochemistry 27, 887-892. (7) Curiel, D. T., Agarwal, S., Wagner, E., and Cotton, M. (1991) Adenovirus enhancement of transferrin-polylysine-mediated gene delivery. Proc. Natl. Acad. Sci. U.S.A. 88, 8850-8854. (8) Rinke, J., Yuki, A., and Brimacombe, R. (1976) Studies on the environment of protein S7 within the 30-S subunit of Escherichia coli ribosomes. Eur. J. Biochem. 64, 77-89. (9) Sperling, J., and Sperling, R. (1978) Photochemical crosslinking of histones to DNA in nucleosomes. Nucleic Acids Res. 8, 2755-2773. (10) Vanin, E. F., Burkhard, S. J., and Kaiser, I. I. (1981) p-Azidophenylglyoxal: a heterobifunctional photosensitive reagent. FEBS Lett. 124, 89-92. (11) Baumert, H. G., Skold, S.-E., and Kurland, C. G. (1978) RNA-protein neighbourhoods of the ribosome obtained by crosslinking. Eur. J. Biochem. 89, 353-359. (12) Cheng, S.-Y., Merlino, G. T., and Pastan, I. H. (1983) A versatile method for the coupling of protein to DNA: synthesis of R2-macroglobulin-DNA conjugates. Nucleic Acids Res. 11, 659-669. (13) Wu, G. Y., and Wu, C. H. (1991) Delivery systems for gene therapy. Biotherapy 3, 87-95. (14) Chang, A. G. Y., and Wu, G. Y. (1994) Gene therapy: applications to the treatment of gastrointestinal and liver diseases. Gastroenterology 106, 1076-1084. (15) Forster, A. C., McInnes, J. L., Skingle, D. C., and Symons, R. H. (1985) Non-radioactive hybridization probes prepared by the chemical labelling of DNA and RNA with a novel reagent, photobiotin. Nucleic Acids Res. 13, 745-761. (16) Chen, C., and Okayama, H. (1987) High efficiency of transformation of mammalian cells by plasmid DNA. Mol. Cell. Biol. 7, 2745-2752. (17) Laemmli, U. K. (1975) Characterization of DNA condensates induced by poly (ethylene oxide) and polylysine. Proc. Natl. Acad. Sci. U.S.A. 72, 4288-4292. (18) Strauss, W. M. (1987) Preparation of genomic DNA from mammalian tissue. In Current Protocols in Molecular Biology, Section 2.2, Wiley-Interscience, New York. (19) Lichtler, A., Barrett, N. L., and Carmichael, G. G. (1992) Simple, inexpensive preparation of T1/T2 ribonuclease suitable for use in RNase protection experiments. BioTechniques 12, 231-232.

Bioconjugate Chem., Vol. 7, No. 5, 1996 551

Biotinylated Plasmid Expression Vectors (20) Struhl, K. (1989) Construction of hybrid DNA molecules. In Current Protocols in Molecular Biology, Section 3.16, Wiley-Interscience, New York. (21) Wagner, E., Cotten, M., Foisner, R., and Birnstiel, M. L. (1991) Transferrin-polycation-DNA complexes: the effect of polycations on the structure of the complex and DNA delivery to cells. Proc. Natl. Acad. Sci. U.S.A. 88, 4255-4259. (22) Fujita, K., and Silver, J. (1993) Surprising lability of biotinstreptavidin bond during transcription of biotinylated DNA bound to paramagnetic streptavidin beads. BioTechniques 14, 608-617. (23) Leahy, P., Carmichael, G. G., and Rossomando, E. F. (1995) Novel plasmid constructs for general and receptor mediated transfections. FASEB J. 9, A 1400. (24) Axelrod, N. J., Carmichael, G. G., and Farabaugh, P. J. (1990) Enhancer and promoter elements from simian virus 40 and polyomavirus can substitute for an upstream activation sequence in Saccharomyces cerevisiae. Mol. Cell. Biol. 10, 947-957. (25) Cotton, M., Wagner, E., and Birnstiel, M. L. (1993) Receptor mediated transport of DNA into eukaryotic cells. Methods Enzymol. 217, 618-644.

(26) Christiano, R. J., Smith, L. C., and Woo, S. L. C. (1993) Hepatic gene therapy: efficient gene delivcery and expression in primary hepatocytes utilizing a conjugated adenovirusDNA complex. Proc. Natl. Acad. Sci. U.S.A. 90, 1154811552. (27) Kalderon, D., Richardson, W. D., Markham, A. F., and Smith, A. E. (1984) Sequence requirements for nuclear location of simian virus 40 large-T antigen. Nature 311, 3338. (28) Dingwall, C., Sharnick, S. V., and Laskey, R. A. (1982) A polypeptide domain that specifies migration of nucleoplasmin into the nucleus. Cell 30, 449-458. (29) Bukrinsky, M. I., Haggerty, S., Dempsey, M. P., Sharova, N., Adzhubel, A., Spitz., L., Lewis, P., Goldfarb, D., Emerman, M., and Stevenson, M. (1993) A nuclear localization signal within HIV-1 matrix protein that governs infection of nondividing cells. Nature 365, 666-669.

BC960044Q