Protamine Sulfate Enhances Lipid-Mediated Gene Transfer

Home , Protamine, Protamine sulfate

FL Sorgi1,2, S Bhattacharya1 and L Huang1 1The Laboratory of Drug Targeting, Department of Pharmacology, The University of Pittsburgh School of Medicine, Pittsburgh, PA, USA

A polycationic peptide, protamine sulfate, USP, has been USP as a condensation agent was found to be superior to shown to be able to condense plasmid DNA efficiently for poly-L-lysine as well as to various other types of protamine. delivery into several different types of cells in vitro by sev- These differences among various salt forms of protamine eral different types of cationic liposomes. The monovalent appear to be attributable to structural differences between cationic liposomal formulations (DC-Chol and lipofectin) the protamines and not due to differences in the net charge exhibited increased transfection activities comparable to of the molecule. The appearance of lysine residues within that seen with the multivalent cationic liposome formu- the protamine molecule correlate with a reduction in bind- lation, lipofectamine. This suggests that lipofectamine’s ing affinity to plasmid DNA, as well as an observed loss in superior in vitro activity arises from its ability to condense transfection-enhancing activity. This finding sheds light on DNA efficiently and that protamine’s primary role is that of the structural requirements of condensation agents for use a condensation agent, although it also possesses several in gene transfer protocols. Furthermore, protamine sulfate, amino acid sequences resembling that of a nuclear localiz- USP is an FDA-approved compound with a documented ation signal. While the use of polycations to condense DNA safety profile and could be readily used as an adjuvant to has been previously reported, the use of protamine sulfate, a human gene therapy protocol.

Keywords: liposome; gene therapy; nonviral vector; DC-Chol; condensation

Introduction in muscle.16 These improvements in vector efficiency translated into a 100 000-fold increase in lung expression The field of nonviral vector-mediated gene therapy, parti- over the first generation cationic lipid–DNA vectors cularly by using cationic liposomes has made great developed in the late 1980s. While these improvements 1 strides from their initial report by Felgner et al in 1987 are impressive and are approaching the efficiency (as to their use in the world’s first in vivo human gene ther- total gene expression) of the adenovirus, they are still at 2 apy clinical trial by Nabel et al in 1992. Cationic lipids least four orders of magnitude behind the adenovirus in have been shown to be a viable alternative to viral vector- terms of the transgene expression for each copy of trans- mediated gene delivery and have demonstrated an excel- gene administered. Furthermore, the time required for 2–4 lent safety profile. They are, however, hampered by development of these advances is slow. Lacking well reports of minimal efficiency as well as a poor fundamen- defined structure/function relationships and organ speci- tal understanding of the basic mechanisms of DNA com- ficity, these advances have occurred through the gross plexation, delivery and expression. While some search for screening of many lipid analogs and plasmid constructs. the ‘magic lipid’, other investigators have been studying While improvements in vector design are needed, they the structure of the DNA–lipid complex, examining the only address the issue of entry into the cell and cellular trafficking to identify barriers to efficient expression once inside the nucleus. Recent work has indi- expression, understanding the lipid coating/uncoating cated that these steps are relatively efficient (60% of the process, producing a stable ‘one-vial’ formulation, or delivered dose of DNA was found within the cell in 6 h) identifying components which may overcome known with modest room for improvement. The most critical 5–15 obstacles to transgene expression. step, which appears to be rate limiting and needs to be Recently, much work has focused on efforts to increase addressed, is the efficient delivery of the gene from the the level of transfection efficiency. New lipid formu- cytoplasm into the nucleus.10,13 It is this step in which the lations have demonstrated a 1000-fold increase in in vitro virus is very efficient and the cationic lipid vehicle transfection activity while new plasmid constructs have appears to be hampered. For the cationic lipid vehicle to resulted in a 3300-fold increase in transgene expression attain the level of gene expression as seen with the adenovirus, the process of efficient delivery of plasmid from the cytoplasm into the nucleus needs large-scale Correspondence: L Huang, University of Pittsburgh School of Medicine, improvement. W1351 Biomedical Science Tower, Pittsburgh, PA 15261, USA 2Present address: Advanced Therapies, Inc, 371 Bel Marin Keys Blvd, It has been hypothesized that the condensation of DNA Suite 210, Novato, CA 94949, USA into a toroidal-like structure is essential for the protection Received 4 February 1997; accepted 9 May 1997 of DNA from enzymatic degradation as well as to facili- Protamine sulfate enhances gene transfer FL Sorgi et al 962 tate entry of the DNA into the nucleus.17–23 Freeze-frac- known.17–22 In order to condense DNA before its associ- ture electron microscopy of DNA–cationic liposome com- ation with the cationic lipid, increasing amounts of prota- plexes has revealed the presence of elongated lipid- mine sulfate, USP or PLL were added to 1 ␮g of DNA coated structures. These structures have been shown to followed by subsequent addition of 7.5 nmol of DC-Chol be lipid-coated DNA and have been seen with several liposomes. As seen in Figure 1a and b, increasing monovalent cationic liposome formulations.9 However, amounts of PLL resulted in an increase in transfection, multivalent cationic liposome formulations, upon com- reaching a constant level of luciferase gene expression at plexation with DNA did not result in the production of approximately 1 ␮g of poly-l-lysine per microgram of elongated structures, rather they produced small DNA. Poly-l-lysine was able to increase the expression (approximately 20 nm), highly condensed lipidic struc- 3.5- to 10-fold (in 293 and CHO cells, respectively) over tures.24 It is felt that this efficient condensation or packag- DNA–liposome complexes without polycation. These ing of plasmid DNA accounts for the observation that the data are in agreement with previously published find- polyvalent liposome formulations generally result in a ings.15 Increasing amounts of protamine sulfate, USP also more efficient rate of in vitro gene transfer than the mono- resulted in an increase in transfection, reaching a constant valent cationic liposome formulations. level of luciferase gene expression at approximately 2 ␮g Recently, several investigators have been using a cat- of protamine sulfate, USP per microgram of DNA. How- ionic polymer, poly-l-lysine (PLL) in an effort to con- ever, the level of expression achieved was seven- to 45- dense DNA into an artificial virus-like structure to facili- fold (in 293 and CHO cells, respectively) over DNA–lipo- tate entry into the cell.15,25–31 However, PLL has several some complexes without polycation, a two- to four-fold qualities which could potentially render it a poor candi- increase over the levels seen with poly-l-lysine. date for human use. Firstly, PLL is a synthetic polymer which is not well defined. The polymerization reaction is Effect on transfection activity of different types of difficult to control resulting in a wide range of molecular protamine weights (the PLL used in this work has a molecular The results shown in Figure 1a and b do not agree with weight range of 18 000–19 200). Further, PLL is not ident- that of Gao and Huang15 in which the transfection ified as a ‘generally regarded as safe (GRAS)’ compound enhancement of activity of PLL was several-fold higher and much testing will be needed before this substance is than that of a protamine. However, this earlier work was ready to be introduced into humans. With these obvious done with a different salt form of protamine, ie prota- hurdles in mind, there may exist other cationic substances which have not been adequately explored that may be capable of producing a similar condensation effect, without the unwanted characteristics associated with PLL. One such cationic compound could be protamine. Pro- tamines are small peptides (MW 4000–4250) which are very basic due to their high arginine content. They are naturally occurring substances found only in sperm and are purified from the mature testes of fish, usually salmon. Protamine’s role in sperm is to bind with DNA, assist in forming a compact structure, and deliver the DNA to the nucleus of the egg after fertilization. This unique role overcomes a major obstacle in gene therapy by nonviral vectors, the efficient delivery of DNA from the cytoplasm into the nucleus.10–13 Furthermore, protamine sulfate is a USP compound, FDA-approved and is used clinically as an antidote to heparin-induced anti- coagulation. Protamine sulfate has also been complexed with insulin (also known as NPH) and serves as a long- acting delivery system which is administered to patients on a daily basis. Due to the extensive history of adminis- tration of NPH insulin to diabetics, issues of toxicity and immunogenicity are minimal. For these reasons, we hypothesize that protamine sulfate may be a safer and more appropriate alternative to PLL for condensation, as well as the delivery of plasmid DNA to the nucleus. In this report, we describe the ability of several different protamine species in enhancing the transfection activity of cationic lipids. Figure 1 Comparison of the ability of protamine sulfate, USP and poly- Results l-lysine to increase transfection activity in (a) CHO cells and (b) 293 cells. Varying amounts of protamine sulfate USP (b) or poly-l-lysine ć ␮ Comparison of the ability of protamine sulfate USP and ( ) were added to 1 g of pUK21-CMV-LUC DNA prior to complexing with 7.5 nmol of DC-chol liposomes per well. Transfection, luciferase and poly-L-lysine to increase transfection activity protein assays were performed as noted in Materials and methods. Each The ability of polycations to increase the transfection data point represents the mean (with standard deviation) of triplicate activity of cationic lipid–DNA complexes is well samples and are normalized to protein content. Protamine sulfate enhances gene transfer FL Sorgi et al 963 mine-free base.15 It is important to determine if the primary differences are a 1.5-fold increase in the amount apparent discrepancy is due to differences in purity of of serine from the phosphate and free base to the sulfates; the two protamines, or if there was a difference in the and a slight decrease in the percentage of arginine salt form of protamine. Different salts of protamine present in the free base and phosphate in comparison to (chloride, free base, phosphate and sulfate), different the sulfates. However, the most striking difference is the sources of origin (salmon or herring), as well as different 16–32-fold increase in lysine content in the phosphate grades (II, III, X) of protamine sulfate were tested for their and free base of protamine. These differences in lysine ability to increase the transfection activity of DC-Chol content correlate in a linear fashion (r2 = 0.953, data not liposomes over that of a control (DC-Chol with DNA but shown) with the variations in activity observed in Figure no protamine). All of the tested protamines are derived 2. In all instances, the percentage of basic (arginine + from salmon with the exception of protamine sulfate lysine) residues remains relatively constant among the grade III which is from herring. Protamine sulfate grade protamines tested. Since the overall charge content of the II contains some insolubles as well as histones, grade III molecule remains unchanged, the differences in observed is essentially histone-free, and grade X is a white amorph- transfection enhancement could be a result of a difference ous powder which is also histone-free, as stated by the in DNA-binding affinity or the ability to effectively pack supplier (Sigma). DNA into a biologically active complex. As shown in Figure 2, the potentiation of gene expression by different types of protamine varied con- Displacement of intercalated ethidium bromide by siderably. Protamine phosphate and free base showed polycations: DNA-binding assay minimal if any increase in gene expression over that seen The addition of various polycations (protamine sulfate in the absence of any polycation. Poly-l-lysine did grade X, protamine phosphate, protamine chloride, pro- increase gene expression by 60-fold over control but the tamine free base, and poly-l-lysine) to the DNA-ethidium protamine sulfates showed an additional three- to five- bromide solution resulted in a rapid decrease in fluor- fold increase over poly-l-lysine. There was no difference escence (Figure 3). This loss in fluorescence can be attri- between salmon (USP, grades II and X) or herring (grade buted to the condensation of DNA and compaction of III) protamine sulfates, as well as no significant difference the DNA–polycation particle which could result in the between the various grades of protamine sulfate. This displacement of the intercalated ethidium bromide in the result also excludes the role of histones (present in prota- DNA by the polycations and/or quenching of the fluor- mine sulfate grade II) as a possible source of the increase escence as the particles tend to aggregate.31–33 Although in observed activity. all the protamines, as well as the poly-l-lysine, were able to condense DNA as shown by the size of the complex Amino acid analysis of various types of protamine (Table 2) as well as by the decreasing fluorescence, the Protamine phosphate, free base, chloride, sulfate (grade efficiency by which they condense DNA is different. The III), and sulfate USP (Lilly and Elkins-Sinn) were ana- condensation efficiency can be estimated from three para- lyzed for an amino acid composition, ie the relative per- meters: (1) the amount of positive charge of the poly- centages of amino acid residues present in each prota- cation required to reach the IC50 value, which can be mine species (Table 1). As can be seen, the amino acid defined as the positive to negative charge ratio (+/−)of composition of protamine-free base is very similar to that polycation to DNA at which 50% of the initial fluor- of the protamine phosphate. Also, there is great similarity escence (Fo) is quenched; (2) the residual fluorescence among the three different protamine sulfate samples. (FRES) at which the polycation can no further exclude ethi- However, there is some variation between the protamine dium bromide from the DNA; and (3) the charge ratio sulfates and the protamine-free base and phosphate. The (RMIN) at which the FRES occurs. The values and statistical significance of IC50, FRES and RMIN measured from Figure 3 for poly-l-lysine, protamine sulfate, protamine phosphate, protamine-free base and protamine chloride are shown in Table 2. Although the protamine sulfate and protamine-free base have apparently the same IC50 value, the protamine- free base shows a lower affinity of binding as evident from the larger values of the RMIN and the FRES. The extent to which the fluorescence drops is possibly a direct indication of the extent of the polycation binding to the DNA, which in turn, is indicative of the stability of the complex. The composition of the complex between the polycation and the DNA which imparts the maximum stability can be estimated from the FRES value. A low FRES value indicates the efficiency by which the polycation has condensed the DNA, excluding the intercalated ethidium bromide. By comparing the FRES, RMIN and IC50 values, it Figure 2 Effect on transfection activity of different types of protamine in is possible to classify the polycations used in this study CHO cells. Two micrograms of varying types of protamine or 1 ␮g poly- l ␮ in terms of their ability and efficiency to condense DNA -lysine were added to 1 g of pUK21-CMV-LUC DNA prior to com- and form a complex. The protamine phosphate and the plexing with 7.5 nmol of DC-Chol liposomes per well. Transfection, luciferase and protein assays were performed as noted in Materials and protamine-free base seem to form the poorest complexes methods. Each data point represents the mean (with standard deviation) with DNA as evident from Table 2, indicating the highest of quadruplicate samples and the data are normalized to protein content. values in all three categories. Protamine chloride can be Protamine sulfate enhances gene transfer FL Sorgi et al 964 Table 1 Amino acid composition of various types of protamine (expressed as percentage of composition)

Amino acid Free base Phosphate Chloride Sulfate Sulfate, USP Sulfate, USP Predicteda (grade III) (Lilly) (Elkins-Sinn)

Aspartate 0.23 0.55 0 0.12 0.50 0.03 0 Threonine 0.10 0.35 0 0 0 0 0 Serine 5.95 6.60 8.91 8.46 8.95 8.23 12.5 Glutamate 0.07 0.01 0 0 0.09 0 0 Glycine 6.69 7.07 7.04 6.56 7.30 6.36 6.25 Alanine 1.44 1.76 1.49 0.94 1.73 1.32 0 Valine 4.69 4.57 4.61 4.93 4.43 4.31 6.25 Methionine 0.56 0.77 0.70 0.64 0.75 0.64 0 Isoleucine 1.26 1.23 1.33 0.81 1.31 1.20 0 Leucine 0.17 0.20 0 0 0.22 0 0 Histadine 0.10 0.11 0 0 0 0 0 Lysine 8.14 8.84 1.49 0.47 0.21 0.23 0 Arginine 61.82 59.61 70.37 68.25 65.90 69.27 65.63 Proline 8.79 8.35 8.60 8.81 8.27 8.42 9.38

aPredicted values are calculated from a published sequence of salmon sperm protamine.36

placed in the intermediate category with an intermediate IC50 value but a large FRES, lessening the importance of the small RMIN value. The most efficient and stable complex formation was displayed by the protamine sulfate which exhibited a low RMIN and the lowest values for l IC50 and FRES. The poly- -lysine, however, was efficient at condensation and forming stable complexes, but had slightly higher values than protamine sulfate in all aspects of forming a stable complex with DNA. Condensation of DNA–lipid complexes by polycations The particle size of the DNA–polycation–DC-chol liposome complex was measured by dynamic light scattering (Table 2). The complex formed by DNA and DC-Chol liposomes in the absense of a polycation had a mean ± Figure 3 Binding behavior of polycations to DNA, studied by ethidium diameter of 637 84 nm. All forms of protamine and PLL bromide assay. Polycations used were poly-l-lysine (í), protamine sulfate were capable of condensing the complex to approxi- (), protamine phosphate (ć), protamine free base (̅) and protamine mately 110–120 nm in diameter. There was not a signifi- chloride (b). The charge ratios were calculated assuming 3.33 nmoles of cant difference in the particle size of complexes contain- phosphate negative charge per microgram of DNA, 4.10 nmoles of positive ing different polycations. charge (based on 21 arginine residues) per microgram of protamines (MW 5.12 kDa) and 4.79 nmoles of positive charge (based on 92 lysine residues The ability of protamine sulfate, USP to potentiate per molecule) per microgram of poly-l-lysine (MW 19.2 kDa). Each data transfection mediated by different cationic liposome point represents the mean of five replicate measurements. The dashed lines formulations indicate the calculated values of IC50,FRES and RMIN for protamine phosphate (shown as an example). In an effort to show that this potentiation of transfection was unrelated to the lipid formulation utilized, five dif-

Table 2 Condensation efﬁciency and size of the condensed plasmid DNA by various polycations

a b c ± d Polycation IC50 FRES RMIN Size (nm) s.d.

Protamine sulfate grade X 0.831 2.8 2.073,5 118 ± 17 Protamine chloride 0.88 8.6 1.825 114 ± 6 Protamine phosphate 1.20 10.04 3.302 141 ± 6 Protamine free base 0.831 10.44 3.252 111 ± 5 Poly-l-lysine 0.98 4.3 2.563 111 ± 17

a + − The charge ratio ( / ) of polycation to DNA at which 50% of the initial ﬂuorescence (Fo) is quenched. bThe residual value of ﬂuorescence indicating the extent of condensation. c + − The charge ratio ( / ) of polycation to DNA at which the FRES occurs. dLipid and DNA in the absence of a polycation results in a complex with a size of 637 ± 84 nm.

A paired two-tailed t test was performed to determine statistical significance for values reported for IC50, FRES and RMIN. All values have a significance of P Ͻ 0.001, unless otherwise indicated. 1 or 2 denotes no significant difference; 3 denotes P Ͻ 0.02; 4 or 5 denotes P Ͻ 0.005. Protamine sulfate enhances gene transfer FL Sorgi et al 965 ferent commercially available cationic liposome formu- amine (Figure 4e). This is probably because the spermine lations were tested and compared. All of the liposome headgroup of this lipid can effectively condense DNA formulations tested displayed an increase in the transfec- into a compact structure.18,24 Protamine sulfate can appar- tion activity (to varying degrees) upon the addition of ently do the same, thus improving the transfection activi- protamine sulfate to the transfection media. As shown in ties of other lipids. Figure 4a–e, protamine sulfate increased the transfection activity of DC-Chol 12–87-fold, CLONfectin five- to 20- Discussion fold, lipofectin 16–105-fold, DOTAP:DOPE liposomes 12– 222-fold, and lipofectamine up to four-fold. Protamine is known to be the major component in the While the extent of the increase in transfection activity sperm nucleus for condensing DNA. All protamines con- varied among the formulations tested, they all had two tain arginine and are strongly basic (isoelectric point = features in common. Firstly, the amount of lipid needed 11–12).34,35 Salmon sperm protamine is a protamine to achieve optimal transfection activity decreased as seen which has been sequenced and shown to contain the fol- by a shift to the left in the optimal ratio of DNA to lipid. lowing 32-amino acid sequence: PRO ARG ARG ARG This is probably due to the neutralization of the negative ARG SER SER SER ARG PRO VAL ARG ARG ARG ARG charge of the phosphate groups of DNA by the arginine ARG PRO ARG VAL SER ARG ARG ARG ARG ARG groups of protamine sulfate. Thus, less cationic lipid is ARG GLY GLY ARG ARG ARG ARG.36 Nearly two- needed to neutralize the DNA–protamine sulfate com- thirds of this sequence (21 of 32 residues) is composed of plexes. This is significant as most of the cellular toxicity arginine, found clustered in four distinct regions, contain- seen upon transfection has been attributed to the cationic ing four to six arginine repeats. The protamine molecule lipid. By reducing the amount of lipid necessary to see has the ability to change its structure from a random coil adequate transfection activity, a reduction in lipid- in solution to a structure containing four ␣-helical regions induced cellular toxicity is observed. in the presence of nucleic acids. These four ␣-helical Secondly, although the extent of potentiation in activity regions having a length of between six and 10 residues, upon addition of protamine sulfate varies among the each joined by a flexible joint at residues 9, 16, and 26, various liposome formulations, the level of protamine followed by proline or glycine which are well known enhanced transfection activity mediated by each lipid helix breakers.34 In this confirmation, protamine can align was similar. In the absence of protamine sulfate, the best with the major groove of double helical DNA of the B transfection activity in CHO cells was seen with lipofect- form. Further, protamine may wind about a single double-strand or may cross-link with adjacent strands, resulting in cross-linking, condensation, and stabilization of the DNA into a highly compact structure.36 The addition of a polycation, such as polylysine, to the transfection media has been shown to enhance cationic lipid-mediated transfections.15 This is thought to occur due to electrostatic interactions between the polycation and DNA, resulting in a charge neutralization of the complex and the formation of a condensed structure. This condensed structure, due to its diminished size, may be more readily endocytosed by the cell, resulting in the increased levels of transgene expression. It is known that polycations with a charge exceeding +2 are sufficient to induce condensation, therefore it was not expected to find significant differences in transfection activity among different polycations, let alone the different salts of protamine, as they were assumed to be identical structures which differed only by their source, purity and coun- terion.18 An analysis of the amino acid compositions of these various protamines revealed the appearance of lysine residues within the protamines which exhibited minimal activity. The data seem to suggest that the lysine residues were the cause of the loss of activity. The appearance of lysine was at the expense of arginine, with the amount of lysine and arginine within each protamine molecule remaining constant. As both lysine and arginine are cationic, the net charge of the various protamines remained constant and should be equally effective in Figure 4 The ability of protamine sulfate, USP to potentiate transfection binding to DNA and producing similarly condensed mediated by different cationic liposome formulations. Two micrograms of structures, assuming that this effect was mediated solely protamine sulfate, USP was added to 1 ␮g of DNA before complexation due to charge, rather than structural composition. with varying amounts of (a) DC-Chol liposomes; (b) clonfectin liposomes; There are two possible explanations for the above (c) lipofectin; (d) DOTAP:DOPE (1:1 m/m) liposomes; or (e) lipofectam- observations. First, that the exchange of one or several ine liposomes. Results are shown for the transfection activity of the complex of the DNA:cationic lipids formulation alone () or in the presence arginine residues for lysine interferes with the binding or of protamine sulfate (í) and are depicted as the luciferase activity nor- affinity of protamine to DNA. It is possible that the lysine malized to cellular protein content. could interrupt the ␣-helical structure, resulting in a Protamine sulfate enhances gene transfer FL Sorgi et al 966 diminished affinity of the binding of the protamine to the propranol precipitation and treatment by high salt to DNA. This hypothesis is strongly supported by the remove high molecular weight RNA. The remaining amino acid analysis, transfection and binding assays supernatant was subject to size exclusion chromatogra- which all indicate a loss in activity/binding correlating phy to separate low molecular RNA from plasmid DNA. with the appearance of lysine residues within the prota- The DNA was shown to be absent of RNA or bacterial mine molecule. DNA as analyzed by gel electrophoresis and had an The second and possibly parallel explanation is that A260/280 ratio between 1.80 and 1.90 as determined spec- protamine contains a nuclear localization signal (NLS) trophotometrically. Detailed growth conditions and puri- which specifically directs the complex to the nucleus. It fication procedure will be published elsewhere. is known that nuclear localization signals are generally characterized as containing short (four to eight residues) Liposomes 38 runs of basic amino acids, often containing a proline or DC-Chol was synthesized as described. DOPE and serine residue upstream to provide a break in the ␣-heli- DOTAP were purchased from Avanti Polar Lipids cal strand. Further, a NLS is often so sequence specific (Alabaster, AL, USA). Lipofectin (DOTMA:DOPE) and that a single point mutation (such as a substitution of a lipofectamine (DOSPA: DOPE) were obtained from Gibco lysine for an arginine residue) is often sufficient to result BRL (Grand Island, NY, USA), while clonfectin was in complete loss of nuclear targeting.37 From the general obtained from Clontech (Palo Alto, CA, USA). All com- characteristics found among NLSs, there exist four such mercially available liposome formulations were used as potential regions within the protamine 32-mer which per the manufacturers’ instructions. DOTAP liposomes could potentially serve as a nuclear localization signal containing DOPE (1:1, m/m) were prepared by sonic- (amino acids 2–5, 12–16, 21–26 and 29–32). The existence ation. DC-Chol liposomes were produced by microfluid- of a NLS within the protamine molecule could account ization at a 3:2 molar ratio of DC-chol to DOPE at a con- ␮ for the increased activity in comparison to other cationic centration of 2 mol/ml (1.2 mg/ml of total lipid) as 39,40 polymers (such as PLL). Likewise, the disruption of the previously described. NLS by the incorporation of a lysine residue could also Polycations account for the loss of activity as seen with the prota- Poly-l-lysine (MW 18000–19 200), protamine-free base mine-free base and phosphate salt. (grade IV), protamine phosphate (grade X), protamine Apart from the aforementioned advantages protamine chloride (grade V), protamine sulfate (grade II), prota- has over the homomorphic polycationic polymers, protamine sulfate (grade III) and protamine sulfate (grade X) mine sulfate is a USP compound and approved for were all obtained from Sigma Chemical (St Louis, MO, human use by the US FDA. Protamine sulfate USP is USA). Protamine sulfate, USP was obtained from Elkins- marketed as a drug product as an antidote for heparin Sinn (Cherry Hill, NY, USA) and Eli Lilly (Indianapolis, overdoses. Protamine combines avidly with heparin for- IN, USA). The analysis of the amino acid composition of ming a stable salt, resulting in the loss of activity of both the various protamines was performed by the Protein compounds. This could be advantageous and exploited and Peptide Core Facility at the University of Pittsburgh. as it has been suggested13 that heparin may play a role in neutralizing or displacing lipids from DNA–lipid com- Transfection of mammalian cells in vitro plexes, resulting in diminished, or loss of, activity. Prota- Chinese hamster ovarian (CHO) cells or human embry- mine has been administered i.v. to humans at a dose of onic kidney 293 cells were seeded into a 48-well plate 600–800 mg with minimal toxicity. Its duration of action ° and grown in a 37 C incubator in a 5% CO2 atmosphere. is 2 h and it has an LD50 in mice of 100 mg/kg. Protamine The cells (approximately 60% confluent culture) were sulfate USP is also an ingredient in neutral protamine exposed to DNA complexed to DC-Chol liposomes with hagedorn (NPH) insulin. Protamine is complexed with or without a polycation such as protamine or PLL and insulin to increase its half-life and duration of action. Pro- incubated for 5.5–6 h after which time the transfection tamine sulfate is nonantigenic, due to the lack of aromatic medium was replaced with fresh F-12 medium sup- 34 amino acids and lack of a rigid structure. This lack of plemented with 10% fetal bovine serum (FBS). Typically, antigenicity is apparent as NPH insulin is administered 8 ␮g of protamine (or 4 ␮g of PLL) were mixed with 4 to diabetics on a daily basis for the lifespan of the patient. ␮g of DNA in a 1 ml volume of HBSS in a 4-ml tube. A In conclusion, it has been shown that protamine sulfate second 4-ml tube containing 30 nmol of DC-Chol lipo- is an effective adjuvant for in vitro cationic lipid-mediated somes in 1-ml of HBSS was added to the 1-ml sample of gene transfer. Due to its excellent safety profile, its use DNA–polycation and incubated at room temperature for in vivo as a systemic transfection vector should be rigor- 15 min. After mixing, 0.5-ml aliquots (containing 1 ␮g ously evaluated. DNA, 2 ␮g protamine or 1 ␮g PLL and 7.5 nmol of DC- Chol liposomes) were added to each well of the 48-well Materials and methods plate. The cells were further incubated for a total of 35.5– 38.5 h before being assayed for luciferase activity. Each data point represents the mean (with standard deviation) Plasmid DNA of three or four replicates and is normalized to protein The plasmid DNA (pUK21-CMV-Luc) used for all experi- content. While the results of transfection might vary from ments consisted of a luciferase (LUC) reporter gene time to time, the reproducibility among different repli- driven by the human cytomegalovirus (CMV) immedi- cates within a given experiment was excellent. ate–early promoter and cloned into a pUK21 backbone. Plasmid DNA was prepared by growing bacterial stocks Luciferase assay in a glycerol enriched terrific broth (TB) medium. The The medium was aspirated and the cells were washed bacteria were subject to detergent and alkaline lysis, once with a 0.9% NaCl solution. The cells were lysed with Protamine sulfate enhances gene transfer FL Sorgi et al 967 100 ␮l of lysis buffer (2 mm EDTA, 100 mm Tris, 0.05% 6 Reimer DL et al. Formation of novel hydrophobic complexes Triton X-100) per well and subjected to one freeze–thaw between cationic lipids and plasmid DNA. Biochemistry 1995; 34: cycle. The cell lysates were collected, briefly centrifuged 12877–12883. to pellet the cellular debris, and 10 ␮l of the supernatant 7 Wong FMP, Reimer DL, Bally MB. Cationic lipid binding to DNA: characterization of complex formation. Biochemistry 1996; was used for the luciferase assay. The samples were 35: 5756–5763. loaded into an AutoLumat LB953 (Berthold, Gaithers- ␮ 8 Mahato RI et al. Physiochemical and pharmacokinetic character- burg, MD, USA) luminometer, 100 l of the luciferase istics of plasmid DNA–cationic liposome complexes. J Pharm Sci substrate (Promega, Madison, WI, USA) was added to the 1995; 84: 1267–1271. cell lysate, and the relative light units (RLU) of each 9 Sternberg B, Sorgi FL, Huang L. New structures in complex for- sample were counted for 20 s. Protein content of the mation between DNA and cationic liposomes visualized by supernatant was determined by Coomassie Blue Plus freeze-fracture electron microscopy. FEBS Lett 1994; 356: 361– Protein Reagent (Pierce, Rockford, IL, USA). 366. 10 Zabner J et al. Cellular and molecular barriers to gene transfer by a cationic lipid. J Biol Chem 1995; 270: 18997–19007. Measurement of fluorescence 11 Friend DS, Papahadjopoulos D, Debs RJ. Endocytosis and intra- Fluorescence studies were performed using a Perkin cellular processing accompanying transfection mediated by cat- Elmer (Norwalk, CT, USA) luminescence spectrometer ionic liposomes. Biochem Biophys Acta 1996; 1278: 41–50. LS50B at an excitation wavelength of 516 nm (slit width 12 Wrobel I, Collins D. Fusion of cationic liposomes with mam- 6 nm) and emission wavelength of 598 nm (slit width 10 malian cells occurs after endocytosis. Biochim Biophys Acta 1996; nm). Briefly, 15 ␮g of pCMV-Luc plasmid DNA was 1235: 296–304. added to a 2955 ␮l sample of 20 mm Tris-HCl buffer (pH 13 Xu Y, Szoka FC. Mechanism of DNA release from cationic liposome–DNA complexes used in cell transfection. Biochemistry 7.4) in a fluorescence cuvette. Ethidium bromide (1.5 ␮g) 1996; 35: 5616–5623. was added to this DNA solution and a baseline fluor- 14 Hofland HEJ, Shephard L, Sullivan SM. Formation of stable cat- escence (Fo) was determined. Various amounts of polyca- ionic lipid–DNA complexes for gene transfer. Proc Natl Acad Sci tions (protamines sulfate grade X, chloride, phosphate, USA 1996; 93: 7305–7309. free base, and poly-l-lysine) were added to the above 15 Gao X, Huang L. Potentiation of cationic liposome-mediated mixture and the fluorescence was measured (n = 5) after gene delivery by polycations. Biochemistry 1996; 35: 1027–1036. each addition. All the fluorescence data were corrected 16 Felgner PL. Improvements in cationic liposomes for in vivo gene for dilution as a result of the addition of the polycation transfer. Hum Gene Ther 1996; 7: 1791–1793. solution, normalized to the fluorescence in the absence of 17 Baeza I et al. Electron microscopy and biochemical properties of any polycation (F ), which was assigned a value of 100. polyamine-compacted DNA. Biochemistry 1987; 26: 6387–6392. o 18 Widom J, Baldwin RL. Cation-induced toroidal condensation of 3+ DNA. Studies with Co (NH3)6. J Mol Biol 1980; 144: 431–453. Measurement of particle size 19 Duguid JG, Bloomfield VA. Electrostatic effects on the stability Particle size measurements were performed using a Zeta- of condensed DNA in the presence of divalent cations. Biophys Sizer 4 system (Malvern Instruments, South Borough, J 1996; 70: 2838–2846. MA, USA). DNA (10 ␮g) was condensed with 20 ␮gof 20 Andreasson B, Nordenskio¨ld L, Schultz J. Interactions of spermi- a protamine sample or 17.1 ␮g of PLL such that the con- dine and methylspermidine with DNA studied by nuclear mag- densed DNA contained an equivalent net charge. This netic resonance self-diffusion measurements. Biophys J 1996; 70: 2847–2856. was further complexed with 80 nmol of DC-Chol lipo- = 21 Gosule LC, Schellman JA. DNA condensation with polyamines somes. The data are represented as the mean (n 5) I. Spectroscopic studies. J Mol Biol 1978; 121: 311–326. diameter with standard deviation. 22 Chattoraj DK, Gosule LC, Schellman JA. DNA condensation with polyamines II. Electron microscopic studies. J Mol Biol 1978; 121: 327–337. Acknowledgements 23 Kabanov AV, Kabanov VA. DNA complexes with polycations for the delivery of genetic material into cells. Bioconj Chem 1995; This work was partially supported by NIH grants CA 6: 7–20. 59327, CA 64654, CA 71731, DK 44935, HL 50256 and the 24 Sorgi FL, Sternberg B, Huang L. Interactions of DNA with lipo- Cystic Fibrosis Foundation Z 967. somes containing different types of cationic amphiphiles. Biochem Biophys Acta (Submitted). 25 Stankovics J et al. Overexpression of human methylmalonyl CoA References mutase in mice after in vivo gene transfer with asialoglycoprot- ein–polylysine–DNA complexes. Hum Gene Ther 1994; 5: 1095– 1 Felgner PL et al. Lipofection: a highly efficient, lipid-mediated 1104. DNA transfection procedure. Proc Natl Acad Sci USA 1987; 84: 26 Erbacher P, Roche AC, Monsigny M, Midoux P. Glycosylated 7413–7417. polylysine–DNA complexes: gene transfer efficiency in relation 2 Nabel GJ et al. Direct gene transfer with DNA–liposome com- to the size and the sugar substitution level of glycosylated poly- plexes in melanoma: expression, biologic activity, and lack of lysines and with the plasmid size. Bioconj Chem 1995; 6: 401–410. toxicity in humans. Proc Natl Acad Sci USA 1993; 90: 11307– 27 Ross GF et al. Surfactant protein A-polylysine conjugates for 11311. delivery of DNA to airway cells in culture. Hum Gene Ther 1995; 3 Stewart MJ et al. Gene transfer in vivo with DNA–liposome com- 6: 31–40. plexes: safety and acute toxicity in mice. Hum Gene Ther 1992; 28 Wagner E, Cotton M, Foisner R, Birnstiel ML. Transferrin–poly- 3: 267–275. cation–DNA complexes: the effect of polycations on the struc- 4 Caplen NJ et al. Liposome-mediated CFTR gene transfer to the ture of the complex and DNA delivery to cells. Proc Natl Acad nasal epithelium of patients with cystic fibrosis. Nature Med Sci USA 1991; 88: 4255–4259. 1995; 1: 39–46. 29 Wagner E et al. Influenza virus hemagglutinin HA-2 N-terminal 5 Gershon H, Ghirlando R, Guttman SB, Minsky A. Mode of for- fusogenic peptides augment gene transfer by transferrin–poly- mation and structural features of DNA–cationic liposome com- lysine–DNA complexes: toward a synthetic virus-like gene plexes used for transfection. Biochemistry 1993; 32: 7143–7151. transfer vehicle. Proc Natl Acad Sci USA 1992; 89: 7934–7938. Protamine sulfate enhances gene transfer FL Sorgi et al 968 30 Taxman DJ, Lee ES, Wojchowski DM. Receptor-targeted trans- 36 Warrant RW, Kim S-H. ␣-Helix–double helix interaction shown fection using stable maleimido-transferrin/thio-poly-l-lysine in the structure of a protamine-transfer RNA complex and a conjugates. Anal Biochem 1993; 213: 97–103. nucleoprotamine model. Nature 1978; 271: 130–135. 31 Wolfert MA, Seymore LW. Atomic force microscopy analysis of 37 Peterson BR, Sun LJ, Verdine GL. A critical arginine residue the influence of the molecular weight of poly(L)lysine on the mediates cooperativity in the contact interface between tran- size of polyelectrolyte complexes formed with DNA. Gene Ther- scription factors NFAT and AP-1. Proc Natl Acad Sci USA 1996; apy 1996; 3: 269–273. 93: 13671–13676. 32 Wolfert MA et al. Characterization of vectors for gene therapy 38 Gao X, Huang L. A novel cationic liposome reagent for efficient formed by self-assembly of DNA with synthetic block co-poly- transfection in mammalian cells. Biochem Biophys Res Commun mers. Hum Gene Ther 1996; 7: 2123–2133. 1991; 179: 280–285. 33 Gershon H, Ghirlando R, Guttman SB, Minsky A. Mode of for- 39 Caplen NJ et al. In vitro liposome-mediated DNA transfection of mation and structural features of DNA–cationic liposome com- epithelial cell lines using the cationic liposome DC-Chol/DOPE. plexes used for transfection. Biochemistry 1993; 32: 7143–7151. Gene Therapy 1995; 2: 603–613. 34 Ando T, Yamasaki M, Suzuki K. Protamines: Isolation, Characteriz- 40 Sorgi FL, Huang L. Large-scale production of DC-Chol lipo- ation, Structure and Function. Springer-Verlag: New York, 1973, somes by microfluidization. Int J Pharm 1996; 144: 131–139. pp 3–87. 35 Alavi N, Lianos E, Andres G, Bentzel CJ. Effect of protamine on the permeability and structure of rat peritoneum. Kidney Int 1982; 21: 44–53.