tat peptide-mediated cellular delivery: back to basics

19
Tat peptide-mediated cellular delivery: back to basics Hilary Brooks, Bernard Lebleu, Eric Vive `s * De ´fenses Antivirales et Antitumorales, CNRS-UMR5124-Universite ´ de Montpellier II, CC086; 5, Place Euge `ne Bataillon, 34095 Montpellier Cedex 05, France Received 10 October 2004; accepted 27 October 2004 Available online 6 January 2005 Abstract Peptides are emerging as attractive drug delivery tools. The HIV Tat-derived peptide is a small basic peptide that has been successfully shown to deliver a large variety of cargoes, from small particles to proteins, peptides and nucleic acids. The dtransduction domainT or region conveying the cell penetrating properties appears to be confined to a small (9 amino acids) stretch of basic amino acids, with the sequence RKKRRQRRR [S. Ruben, A. Perkins, R. Purcell, K. Joung, R. Sia, R. Burghoff, W.A. Haseltine, C.A. Rosen, Structural and functional characterization of human immunodeficiency virus tat protein, J. Virol. 63 (1989) 1–8; S. Fawell, J. Seery, Y. Daikh, C. Moore, L.L. Chen, B. Pepinsky, J. Barsoum, Tat-mediated delivery of heterologous proteins into cells, Proc. Natl. Acad. Sci. U. S. A. 91 (1994) 664–668; E. Vives, P. Brodin, B. Lebleu, A truncated HIV-1 Tat protein basic domain rapidly translocates through the plasma membrane and accumulates in the cell nucleus, J. Biol. Chem. 272 (1997) 16010–16017; S. Futaki, T. Suzuki, W. Ohashi, T. Yagami, S. Tanaka, K. Ueda, Y. Sugiura, Arginine-rich peptides. An abundant source of membrane-permeable peptides having potential as carriers for intracellular protein delivery, J. Biol. Chem. 276 (2001) 5836–5840.]. The mechanism by which the Tat peptide adheres to, and crosses, the plasma membrane of cells is currently a topic of heated discussion in the literature, with varied findings being reported. This review aims to bring together some of those findings. Peptide interactions at the cell surface, and possible mechanisms of entry, will be discussed together with the effects of modifying the basic sequence and attaching a cargo. D 2004 Elsevier B.V. All rights reserved. Keywords: TAT; Uptake; Cell delivery; CPP Contents 1. Introduction..................................................... 560 2. Implication of the basic cluster ........................................... 560 3. Influence of other components surrounding the basic domain ........................... 561 0169-409X/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.addr.2004.12.001 * Corresponding author. Tel.: +33 467 16 33 06; fax: +33 467 16 33 01. E-mail address: [email protected] (E. Vive `s). Advanced Drug Delivery Reviews 57 (2005) 559 – 577 www.elsevier.com/locate/addr

Upload: hilary-brooks

Post on 03-Sep-2016

212 views

Category:

Documents


0 download

TRANSCRIPT

www.elsevier.com/locate/addr

Advanced Drug Delivery Rev

Tat peptide-mediated cellular delivery: back to basics

Hilary Brooks, Bernard Lebleu, Eric Vives*

Defenses Antivirales et Antitumorales, CNRS-UMR5124-Universite de Montpellier II, CC086; 5, Place Eugene Bataillon,

34095 Montpellier Cedex 05, France

Received 10 October 2004; accepted 27 October 2004

Available online 6 January 2005

Abstract

Peptides are emerging as attractive drug delivery tools. The HIV Tat-derived peptide is a small basic peptide that has been

successfully shown to deliver a large variety of cargoes, from small particles to proteins, peptides and nucleic acids. The

dtransduction domainT or region conveying the cell penetrating properties appears to be confined to a small (9 amino acids)

stretch of basic amino acids, with the sequence RKKRRQRRR [S. Ruben, A. Perkins, R. Purcell, K. Joung, R. Sia, R. Burghoff,

W.A. Haseltine, C.A. Rosen, Structural and functional characterization of human immunodeficiency virus tat protein, J. Virol.

63 (1989) 1–8; S. Fawell, J. Seery, Y. Daikh, C. Moore, L.L. Chen, B. Pepinsky, J. Barsoum, Tat-mediated delivery of

heterologous proteins into cells, Proc. Natl. Acad. Sci. U. S. A. 91 (1994) 664–668; E. Vives, P. Brodin, B. Lebleu, A truncated

HIV-1 Tat protein basic domain rapidly translocates through the plasma membrane and accumulates in the cell nucleus, J. Biol.

Chem. 272 (1997) 16010–16017; S. Futaki, T. Suzuki, W. Ohashi, T. Yagami, S. Tanaka, K. Ueda, Y. Sugiura, Arginine-rich

peptides. An abundant source of membrane-permeable peptides having potential as carriers for intracellular protein delivery, J.

Biol. Chem. 276 (2001) 5836–5840.]. The mechanism by which the Tat peptide adheres to, and crosses, the plasma membrane

of cells is currently a topic of heated discussion in the literature, with varied findings being reported. This review aims to bring

together some of those findings. Peptide interactions at the cell surface, and possible mechanisms of entry, will be discussed

together with the effects of modifying the basic sequence and attaching a cargo.

D 2004 Elsevier B.V. All rights reserved.

Keywords: TAT; Uptake; Cell delivery; CPP

Contents

1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 560

2. Implication of the basic cluster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 560

3. Influence of other components surrounding the basic domain . . . . . . . . . . . . . . . . . . . . . . . . . . . 561

0169-409X/$ - s

doi:10.1016/j.ad

* Correspondi

E-mail addres

iews 57 (2005) 559–577

ee front matter D 2004 Elsevier B.V. All rights reserved.

dr.2004.12.001

ng author. Tel.: +33 467 16 33 06; fax: +33 467 16 33 01.

s: [email protected] (E. Vives).

H. Brooks et al. / Advanced Drug Delivery Reviews 57 (2005) 559–577560

4. Influence of the cargo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 562

4.1. Influence of Tat peptide exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 563

4.2. Influence of the hydrophobicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 564

4.3. Influence of the chemical linkage between Tat and the cargo . . . . . . . . . . . . . . . . . . . . . . . . 565

4.4. Influence of the Tat peptide density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 567

4.5. Influence of the serum. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 568

5. Tat and cell surface interactions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 568

6. Possible mechanisms of internalisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 571

7. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 572

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 573

1. Introduction

Over the last decade, several publications revealed

a massive improvement in the cellular delivery of

various biologically active molecules upon their

attachment to a peptide derived from the HIV-1 Tat

protein. This peptide can be reduced to a cluster of

basic amino acids containing 6 arginine and 2 lysine

residues within a linear sequence of 9 amino acids.

Because of the high content of arginine residues within

the Tat sequence, various homopolymers of arginine

have also been investigated to study the mechanism of

entry of various cargoes. Very similar results were

obtained with these simple polymers of arginine in

terms of transduction efficiency and apparent mecha-

nism of entry compared to the Tat peptide [4].

Among others, we recently reevaluated the mecha-

nism of entry of the Tat peptide [5] and highlighted

various problems related with the FACS quantification

and the fixation procedure prior to microscopy obser-

vations. Despite the possibility that the uptake of various

entities previously described in the literature could have

been artifactual or overestimated, it is unlikely that the

efficiency of the Tat-mediated uptake could be disputed,

due to the high number of examples of biological

activity which have been provided upon Tat peptide-

mediated cellular delivery of peptides, proteins or

nucleic acids (to name but a few) [6–8]. In the large

majority of the experiments, the chimera concentration

used for obtaining the expected biological responses was

not outrageously different to those used to assess the cell

uptake of the Tat peptide itself. For instance, most of the

fused compounds are active at concentrations in the 100-

nM range ([9] for oligonucleotides, [10] for peptides,

and [11] for protein delivery), whereas fluorescence

microscopy or FACS quantification of the uptake were

usually performed with 1–10 AM of the peptide [3–

5,12]. Despite a possible dose effect causing variations

in the efficiency of the uptake, and potentially the

cellular pathway induced in uptake, it has been assumed

that the entry mechanism of the Tat peptide and of Tat-

carrying chimera is similar. However, little has been

done to unambiguously answer this question by

comparing the uptake efficiency of two different entities

under the same cellular and experimental conditions. It

might be very important to consider the influence of the

physicochemical character of the cargo.

Despite the high number of biological applications

using these peptides, and principally the Tat peptide,

the precise mechanism of entry still appears controver-

sial and certainly requires further investigations. Con-

tradictory results are still often obtained. They could

result from experimental variations in, for example, the

diversity of the Tat peptide sequence used to promote

the translocating activity, the wide variety of cell lines

studied, the differing protocols applied to investigate

the entry mechanism or the high diversity of cargoes,

all of which might well influence the behavior of the

Tat peptide during the cellular entry process. This

review is aimed at giving an up-to-date statement of

various parameters possibly influencing the observed

results during the investigations about the translocating

properties of the Tat peptide and its attached cargoes.

2. Implication of the basic cluster

The initial work showing the ability of a Tat

protein-derived peptide to deliver heterologous mol-

ecules into cells was provided 10 years ago by

H. Brooks et al. / Advanced Drug Delivery Reviews 57 (2005) 559–577 561

Fawell et al. [2]. In this study, the covalent coupling

of a 36 amino acid peptide (Tat37–72) to large

proteins, such as h-galactosidase, RNAse or perox-

idase, was performed through a heterolinker. Four to

five peptide molecules per protein unit were suffi-

cient to mediate the cellular delivery of the

covalently bound protein [2]. The delivery was

assessed by monitoring the corresponding activity

of the delivered protein, therefore, demonstrating the

effectiveness of the uptake process and also, more

importantly, the cell viability.

This cell-penetrating peptide (CPP) encompassed a

highly cationic cluster composed of 6 arginine and 2

lysine residues in the very middle of the peptide

sequence and an a-helical structure on the N-terminal

part [13]. Around the same time, another peptide able

to translocate through the plasma membrane had been

discovered [14]. This peptide was derived from the

third domain of the Antennapedia homeodomain from

Drosophila melanogaster known to bind DNA

sequences and to activate various genes. Interestingly,

this peptide also contains a high number of cationic

amino acids and structural data showed an overall a-

helical structure. Since the N-terminal portion of the

Tat sequence used for introducing large proteins into

cells was shown to adopt an a-helical conformation,

and the cationic cluster was shown to adopt an

extended structure [13], a structure–activity relation-

ship study was performed to delineate which feature

within the Tat peptide was responsible for the cell

membrane translocating property [3]. Several peptides

carrying either partial or total deletion within the a-

helical structure and a peptide harboring a deletion

within the cationic domain of the original peptide

were synthesized [3]. It was shown, primarily by

fluorescence microscopy, that the main determinant

responsible for the translocating activity was the

cationic cluster of amino acids and that deletion of

arginine led to an apparent non-translocating peptide.

The a-helical structure was shown to induce signifi-

cant toxicity in HeLa cells, as assessed by MTT

toxicity testing [3]. The toxicity of this region has

been recently confirmed in an independent study [15].

These studies, however, found no toxicity for the

cationic cluster [3,4,15], in contrast to a previous

study showing a neurotoxic activity induced by the

Tat basic peptide after intracerebroventricular (ICV)

injection in mice [16]. From these observations, the

translocating property of the initial peptide could be

reduced down to the sequence encompassing the

cationic cluster containing 8 basic charges within a 9

amino acid linear sequence. Despite the risk of

misleading results, due to fixation artifacts (see

below), this short Tat cationic peptide has been used

in several examples to mediate the cellular uptake of

biologically active molecules. At the same time,

evidence has been provided to show that the a-helical

structure within the Antennapedia peptide was dis-

pensable, since the insertion within the primary

sequence of two proline residues, an amino acid

known to disrupt a-helical structures, did not impair

the translocating activity of this peptide [17]. The

cationic content of these peptides, thus, appeared to be

responsible for their ability to be taken up by cells. In

the same study [17], tryptophan residues were high-

lighted to play a crucial role in the translocating

property of the Antennapedia peptide [17]. No

tryptophan residue, or other aromatic amino acids, is

present within the minimal Tat peptide primary

sequence shown to be able to enter cells. Solely the

work of Thoren et al., using Tat peptide with a

terminal tryptophan, continues to describe uptake at 4

8C [18]. Despite this latter difference and some

controversial data regarding the actual mechanism of

entry (see below), both peptides were initially shown

to be taken up by cells very rapidly (within minutes)

and at temperatures known to inhibit active transport.

These findings stimulated a very high interest in the

possibility of using such peptides to carry various

drugs into cells.

3. Influence of other components surrounding the

basic domain

Since different versions of the Tat peptide have

been used to promote cellular delivery of many types

of cargo, it appears necessary to consider a putative

effect of their molecular nature. Various results,

sometimes contradictory, about the efficacy of the

uptake or about the internalisation pathway have been

obtained during the last years. It appears now widely

accepted that the cationic nature of the Tat peptide

alone promotes the cellular delivery of very different

entities in terms of molecular size, structure and

overall physicochemical properties. The common

H. Brooks et al. / Advanced Drug Delivery Reviews 57 (2005) 559–577562

sequence determinant in most of the studies per-

formed with the Tat peptide is the GRKKRRQRRR

sequence. The C-terminal part of this peptide has been

extended by various sequences or coupled to different

moieties (see Table 1). These include amino acids

present in the native Tat protein sequence, such as the

PPQ sequence [3,19–23], but also longer cargoes in

order to induce their own cellular uptake ([9,11,24]

and Table 1). The N-terminal part of this minimal Tat

peptide has also been extended by various amino acid

sequences from either the native Tat protein sequence

or again attached to different entities [15,21,22,24–

27]. All of these Tat-cargo derivatives, except the one

used in the work from Koppelhus and collaborators

[23], showed the ability to be taken up, since

fluorochromes or covalently bound molecules were

recovered in cells (see Table 1 for references). From

these different applications, no clear influence of the

additional moiety attached either to the N-terminal or

to the C-terminal end of the bcoreQ Tat peptide could

be related to a variable ability of the basic Tat peptide

to mediate cellular uptake. It should, however, be

pointed out that comparison is rather difficult since

most of these studies were performed on different cell

lines following different experimental protocols, as is

discussed below.

Table 1

Tat-derived peptides used for the cellular delivery of various cargoes

N-terminal end Core peptide C-terminal end Reference

GRKKRRQRRR -GYK(FITC)C [28]

GRKKRRQRRR -PPQ [19,22]

GRKKRRQRRR -G [25]

GRKKRRQRRR -PPQC [9,20,29]

GRKKRRQRRR -GYK(FITC) [30]

GRKKRRQRRR -Cre protein [11]

MYG- GRKKRRQRRR -G [27]

GST protein- GRKKRRQRRR -GFP protein [24]

SGYG GRKKRRQRRR -C [31]

MLGISY- GRKKRRQRRR -PPQT [21]

CY- GRKKRRQRRR [22,25]

CGISY- GRKKRRQRRR [15]

CFITKALGISY- GRKKRRQ [15]

C- GRKKRRQRRR [23]

Biotin-Y GRKKRRQRRR [26]

MLGISY- GRKKRRQRRR [21]

His6- GRKKRRQRRR [32]

Y- GRKKRRQRRR [33,34]

Acetyl- GRKKRRQRRR [35]

4. Influence of the cargo

The more remarkable fact regarding the Tat

peptide’s ability as a vector system is the molecular

diversity of the btransducedQ entities, ranging from

small molecules of some hundreds of daltons to

massive structures with a diameter up to 200 nm

such as liposomes [36,37]. Until very recently, it

was believed that the translocating activity of the

Tat peptide could occur directly through the plasma

membrane following an inverted micelle formation

as earlier proposed for the Antennapedia-derived

peptide [17]. Such a mechanism was originally

proposed to explain the Antennapedia peptide trans-

location occurring even at a temperature (4 8C)known to inhibit all cellular energy-dependant path-

ways. Basically, ionic interactions between the

cationic charges of the peptide and anionic charges

of the membrane components (principally the

phosphate groups of the phospholipid heads) ini-

tiated the membrane adsorption of the peptide. Then,

phospholipid reorganization led to the scavenging of

the peptide inside a fully hydrophilic pocket, the

inverted micelle [17]. No direct evidence, however,

except in one single report [38], has been provided

so far to clearly demonstrate such a mechanism

within homogeneous artificial membrane systems for

either the Antennapedia or the Tat peptide. Never-

theless, the presence of numerous proteins anchored

in cell membrane or exposed at the cell membrane

surface, along with the variation in lipid composi-

tion within and between different cell types (e.g.

lipid rafts), could be required for such a mechanism

to function properly. It is difficult to imagine,

however, that an inverted micelle mechanism could

be applied to large molecules, such as proteins or

even much bigger structures. As an example,

ferromagnetic particles of about 45 nm in diameter

[22] were shown to be taken up by cells, as

demonstrated by the magnetic recovery of particle-

containing cells from a cell mixture. The particle

diameter was about 15-fold larger than the entire

thickness of membrane (30 angstroms or 3 nm).

Although each particle carried 4 to 5 peptide

molecules at its surface, it seems very unlikely that

the membrane could reorganize itself completely

around the ferromagnetic particle as proposed for

the peptide itself.

H. Brooks et al. / Advanced Drug Delivery Reviews 57 (2005) 559–577 563

The cationic charges of the Tat peptide certainly

play a key role in the uptake process, since a single

deletion or substitution of basic charges induced a

reduction of the cell association of the peptide,

therefore, probably reducing the overall intracellular

uptake [3,39,40]. The guanidinium group of the

arginine side chain was shown to be more potent in

mediating cellular uptake than other cationic groups,

such as lysine, histidine or ornithine [41]. Along these

lines, very similar cell-penetrating properties were

obtained with simple homopolymers containing only

residues with guanidinium side chains ([4,40,42], see

below).

Since ionic interactions appear to initiate the

uptake process, it has to be recalled that the

improvement of the delivery of highly anionic

moieties, namely, nucleic acids, can be mediated

by the arginine-rich peptide derived from the Tat

protein, as has been recently documented [25]. It is

assumed that the complexation of nucleic acids to

cationic polymers allows the condensation of the

nucleic acid moiety and the subsequent release from

endosomes, as assessed for cationic lipids [43]. In

this study, the Tat peptide, with its high guanidi-

nium group content, has been attached to poly-

ethylene imine (PEI), an amine containing polymer,

and compared to PEI alone for the transfection

efficiency of a reporter gene. A marked increase, up

to 100-fold, of the cellular delivery of plasmids

already complexed with polycationic compounds

was recorded. Thus, the cationic charge of the Tat

peptide alone could not be considered responsible

for the improved translocating properties of these

nucleic acids, as it also markedly improved the

delivery of nucleic acids complexed with PEI,

which already carry a highly cationic charge due

to the high amount of amino groups. In conclusion,

the Tat guanidinium groups induced a massive

increase in the transfection efficiency of nucleic

acids delivered by complexation with the simple

amine-functionalized polymer, PEI. This latter

example also illustrates that particular features are

required to fulfill this translocating process. These

are, for instance, an appropriate charge ratio or the

number of Tat peptides present on each plasmid

particle. Therefore, possible factors expected to play

a role in the cell entry mechanism will be discussed

further.

4.1. Influence of Tat peptide exposure

In some cases, it cannot be excluded that the direct

environment of the Tat peptide once inserted into a

chimera and its overall exposure within this structure

could influence the behavior of the translocating

process. As an example, a shorter version than the

36 amino acid Tat peptide used originally to mediate

the cellular delivery of proteins was also evaluated in

the same study [2]. A weaker translocating activity

was reported for such covalently bound chimeras.

Thus, the level of exposure of the Tat basic peptide at

the molecular surface of the transported protein could

be closely related with the level of the observed

cellular uptake. Along these lines, it is possible that a

steric hindrance of the short cationic peptide pre-

vented interactions with cellular components promot-

ing the uptake, since the Tat peptide coupling was

mediated randomly on large proteins, such as h-galactosidase (120 kDa) [2]. In longer versions of the

bound peptide, the cationic cluster is potentially better

exposed and, therefore, more accessible to cell surface

structures.

The influence of the exposure of the short Tat basic

peptide has been elegantly demonstrated for the

cellular delivery of liposomes [37]. In this study,

liposomes were functionalized by lipids coupled to the

Tat peptide following various degrees of exposure

depending on the spacer length. Only liposomes with

an appropriate exposure of the peptide were taken up

by cells [37]. Other studies, including one showing in

vivo delivery of proteins fused to an 11 amino acid Tat

peptide containing the cationic cluster, provided

evidence of the importance of the exposure of the

Tat peptide to fulfill the cellular delivery of the

chimera [44]. A fully unfolded fusion protein,

believed to expose the highly hydrophilic Tat

sequence to the fluidic environment, was recovered

in various tissues, including lungs, heart, spleen,

kidneys, and also the brain, after intraperitoneal

injection [44]. It was later discussed that the protein

had to be unfolded to be efficiently taken up by cells

[45,46]. Although the reasons of this requirement are

still not fully understood, this work was certainly

influential in stimulating the widespread use of the Tat

peptide to mediate the cellular delivery of various

entities into several cell types, both in vitro and in

vivo (for a review, see Ref. [6]).

H. Brooks et al. / Advanced Drug Delivery Reviews 57 (2005) 559–577564

In another example, two very different mecha-

nisms of Tat peptide-mediated cell entry (caveolae

versus macropinocytosis) were recently proposed

following strongly convincing studies [11,24]. The

cargoes used in these studies were both proteins:

One was made up of the Tat peptide fused to the

GST tag protein at the N-terminal end and to the

GFP reporter protein at the C-terminal end [24],

while the second was composed of the Tat peptide

fused to the Cre-recombinase at its C-terminal end

[11]. Although the studies were again performed on

different cell lines (HeLa and CHO cells for the

GST-Tat-GFP and 3T3 cells for the Tat-Cre con-

struct), it is conceivable that the exposure, and thus

the accessibility of the peptide, could be different in

both constructs, because of differences in their

folding, due to the physicochemical properties of

the cargo itself. This could lead subsequently to a

different ability to follow one or the other cell entry

pathway. Along these lines, some toxins have been

shown to enter cells through caveolae, whereas other

pathogens exclusively used alternative routes without

the involvement of identified specific cellular recep-

tors able to trigger one or the other entry route [47].

Biochemical evidence also highlighted that each

construct followed its own individual entry pathway.

Real-time microscopy experiments were performed

to follow the entry route of the GFP-cargo, the

kinetics of which implicated the caveolae pathway

[24]. In addition, it was assumed that the GFP

protein had to be trapped in a neutral environment to

maintain its fluorescent property. No reduction of the

fluorescence activity was recorded, therefore, con-

firming indirectly that the GFP-cargo was taken up

by cells through the caveolae pathway, which is not

acidified during the course of intracellular trafficking

[48]. On the other hand, the Tat-Cre biological

response was strongly increased upon co-incubation

with a Tat peptide fused to the fusogenic sequence

derived from the Influenza hemagluttinin protein,

known to promote membrane fusion once exposed to

an acidic environment [49]. This observation, there-

fore provided an additional argument that the Tat-Cre

protein was taken up by a pathway undergoing

acidification [11], while the GST-Tat-GFP construct

appeared to be taken up though a mechanism of

entry with a stable neutral pH environment [24].

Again, the different cell entry behaviors of these

similar constructs, both are fusion proteins, are likely

due to unknown parameters that require further

investigation.

In addition, the orientation of the coupling of

peptide to cargo has been investigated in a number of

studies. For instance, morpholino oligonucleotides

(PMO) were coupled to the Tat peptide either at their

5V or 3V end [50]. A higher antisense activity was

recorded when the Tat peptide was attached to the 5Vend. It is believed that the addition of bulky moieties

to the 3V end of an oligonucleotide could decrease the

expected biological activity, because of probable

steric interferences affecting PMO/mRNA binding

[50]. Any possible direct effect of this orientation on

the uptake itself was, however, not investigated in this

study.

4.2. Influence of the hydrophobicity

The influence of the structure attached to the Tat

peptide should also be further considered. For

instance, in early studies of the Tat peptide trans-

locating properties, it was shown that the attachment

of a biotin group at the N-terminal end of the longer

version of the Tat peptide could lead to a 6-fold

increase of the cellular uptake [51]. In this study, the

improvement of the cellular delivery of the biotiny-

lated peptide was assessed by the increase of the

biological activity mediated by the cargo-bound

RNAse [51]. This associated biological effect allowed

the avoidance of any risk of artifactual results, such as

increased extracellular association or relocalisation

upon fixation, as recently described [5]. Therefore, the

attachment of such a small hydrophobic group to the

Tat peptide could significantly modify the trans-

locating efficiency of the cationic Tat peptide. The

improvement of the cellular uptake upon increase of

the hydrophobicity has also been indirectly shown

after variation of the methylene content of the side

chain of h-amino acids of arginine [40]. In this study,

although performed on an arginine homopolymer, a

stronger cell associated signal was observed by FACS

analysis when 4 to 6 methylene groups were present

between the a-carbon and the distal guanidinium

group. Conversely, a reduction of the side chain

length (down to 2 methylene groups) showed a

weaker signal compared to the native side chain

length containing 3 methylene groups [40]. Addition-

H. Brooks et al. / Advanced Drug Delivery Reviews 57 (2005) 559–577 565

ally, the insertion of aminocaproic acid (aca) groups

within the peptide backbone showed a stronger cell

associated signal than the corresponding heptahomo-

polymers of arginine [42]. Aca groups allow spacing

along the peptidic backbone between the a-carbon, to

which the arginine side chain is attached, but also

confer a highly hydrophobic character because of their

five methylene groups. Although not directly related

to the Tat peptide, the improvement of the cell

association of a homopolymer of arginine upon

attachment of a stearyl moiety has also been

documented [52]. Very recently, two phenylalanine

amino acids have been inserted at the C-terminal end

of arginine homopolymers [50]. These homopolymers

were compared directly to the Tat peptide with regard

to their expression of luciferase mediated by a positive

readout of the antisense action provided by a

phosphorothioate oligonucleotide [53]. Despite the

absence of a direct comparison of the positive effect

of these two phenylalanine residues on the trans-

fection efficiency of either Tat or arginine polymers in

this study, it is noteworthy to compare the biological

response induced by the native Tat sequence and by

the hexapolymer of arginine to which the two

phenylalanine were attached. They were found to be

very similar [50], whereas without the addition of

these two extra phenylalanine residues, Tat was

shown to be more efficiently taken up when compared

to a simple hexapolymer of arginine [54]. It thus

appears that the attachment of hydrophobic residues,

or the inclusion of hydrophobic patches in a poly-

cationic cell-penetrating peptide, such as Tat or

arginine polymers, could improve their overall uptake.

Whether this apparent increase of the measured signal

resulted from an effective improvement of the trans-

locating process itself, an increase of the initial cell

binding or the use of an additional entry pathway

upon peptide modification could not be fully ascer-

tained and, again, could be the subject of further

investigations.

The hydrophobicity of the peptide does not appear

to be the only single characteristic promoting peptide

uptake. In a closely related example, the substitution

of two tryptophan residues by two phenylalanine

residues within the Antennapedia peptide [17] led to

the complete loss of translocating properties, although

phenylalanine residues show a relatively higher

hydrophobicity than tryptophan residues [55].

On the other hand, the Tat peptide does not contain

any hydrophobic amino acids and is conversely very

hydrophilic because of the presence of 8 ionic charges

on the side chains and the two N- and C-terminal ionic

groups. The translocating process has also been

assessed for hydrophilic cargoes. For instance, the

covalent coupling of the Tat peptide to a phosphor-

othioate oligonucleotide with a high number of

anionic charges, still resulted in the uptake of the

nucleic acid cargo [9]. These results appear quite

convincing since they were obtained after trypsin

treatment of the cells prior to analysis either by FACS

or by fluorescent microscopy, therefore, allowing the

removal of the nonspecific membrane-bound peptides

as detailed in another section of this chapter [5].

4.3. Influence of the chemical linkage between Tat and

the cargo

Surprisingly, the nature of the linkage between the

Tat peptide and its cargo has not been deeply

investigated. Indeed, this could be highly important

if we consider first the requirement of an efficient

exposure of the Tat peptide when bound to the cargo

to any cell component involved in the translocating

process (see above), and, secondly, the intracellular

activity of the cargo itself which has to be unaffected

by the nature of its chemical coupling to the Tat

peptide. For instance, the Tat peptide could impair the

biological response by reducing the affinity of the

cargo to the targeted material. Along these lines, little

in vitro evaluation of the biological activity of the

cargo moiety prior to and after Tat attachment has

been provided, nor has detection of the subcellular

localisation of the cargo following Tat-mediated

internalisation been investigated thoroughly, despite

a very important number of Tat-mediated deliveries of

various cargoes (see Ref. [6] for a review). Most of

the studies were devoted to gaining substantial

biological activity triggered by the cargo once coupled

to the CPP without considering these biochemical

features. Moreover, most of these biological responses

have been recorded with substantially high doses of

extracellular chimera (100 nM or more). Considering

that an intracellular concentration of the cargo is

expected [56], this could reflect either a very weak

activity of the cargo moiety, or an inappropriate

location within the cell upon Tat peptide attachment,

H. Brooks et al. / Advanced Drug Delivery Reviews 57 (2005) 559–577566

thus scavenging the active drug. Therefore, a labile

bond between the Tat carrier and the desired cargo is

expected to perform better.

The most convenient bond formation for this

purpose is that of a disulfide bridge between the

CPP and the cargo. Considering the preparation of

these chimeras, this strategy first allows the separate

synthesis, purification and characterisation of both

entities. Secondly, several strategies are available for

promoting the oriented formation of the heterodimer

entity upon activation of one of the sulfhydryl

functions only, such as the incorporation of a

thionitropyridine either on the peptides or an oligo-

nucleotide cargo [57–59]. Thirdly, the reduction of the

disulfide bridge once the cargo reaches the cytoplas-

mic compartment is expected to induce its release,

therefore, preventing putative negative effects of the

cell penetrating peptide. The efficiency of the intra-

cellular reduction of the disulfide bridge between a

CPP and its cargo has been demonstrated by using a

quenched fluorescent construct whose fluorescence

was activated upon cytoplasmic disulfide bond

reduction [56].

In some particular examples, however, an increase

in the biological response is expected upon attachment

of a cationic moiety in general, and particularly upon

attachment of the Tat peptide. This is the case for the

Tat-mediated delivery of antisense oligonucleotides.

The expected activity of an oligonucleotide delivered

in this manner results from a complementary binding

to its RNA/DNA target sequence. This leads to either

steric hindrance of transcription or splicing, or the

digestion of mRNA by RNAse H, and, thus, a

modulation of the corresponding target protein

expression (for a review, see Ref. [60]). The positive

effect on oligonucleotide binding to a complementary

oligonucleotide sequence upon their coupling to

cationic sequences has been widely investigated

[61,62]. Because of the highly cationic nature of the

Tat peptide, the improvement of binding affinity and

kinetics of a Tat-oligonucleotide chimera is also

expected and should be considered an important issue

for Tat-mediated oligonucleotide delivery. This

increase in kinetics has been recently documented

for the Tat peptide coupled to an oligonucleotide

using the BioCore technology [9]. Therefore, in some

cases, a stable link between the Tat peptide and its

oligonucleotide cargo might be preferred in the

interest of augmenting the hybridisation of comple-

mentary sequences of nucleic acids. The placement of

the peptide with respect to the cargo should also be

taken into consideration [50].

A comparison of a stable versus a labile bond has

been recently provided [50]. Although this study used

arginine homopolymers (9 arginine residues) as a cell-

penetrating peptide for delivering morpholino

oligomers (PMO), similar data could be expected

with the Tat peptide considering their very close

behavior with regards to inducing cellular uptake of

various cargoes. Interestingly, it was shown that when

both constructs were labeled with fluorescein, the

uptake was more efficient for the stable link construct

compared to the labile bond construct, but both gave

nearly identical antisense activity in a dose-dependant

manner [50]. The following comment was made to

explain such differences: A reduction of the disulfide

bridge could occur at the membrane level of the cell

by gluthatione [63], therefore, reducing the overall

pool of internalised PMO. To explain the identical

biological effect obtained for both constructs, it was

proposed that the stable nature of the maleimide linker

could induce the scavenging of the PMO to anionic

cellular components rather than the targeted mRNA

because of the positive charges of the CPP [50]. These

results took into account the risk of a nonspecific

binding at the cell surface membrane since a trypsin

treatment was performed prior to FACS analysis as

previously demonstrated [5].

Therefore, the nature of the chemical bond between

a CPP, such as the Tat peptide, and its cargo should be

probably more deeply evaluated in each application to

optimize the level of the expected biological response,

particularly in the case of biologically active oligo-

nucleotides and peptides.

One of the most intriguing applications observed in

the field of Tat-cargo coupling was the improvement

of the cellular delivery of an adenovirus simply upon

the mixture of cells with a Tat peptide solution for 30

min prior to exposing the cells to the adenovirus (4 h)

[19]. Interactions between Tat (and also the Anten-

napedia peptide) with cellular coat proteins or lipids

of the cell membrane were proposed to improve the

effective surface concentration of the adenovirus. It is

noteworthy that high concentrations (above 100 AM)

of either Tat or Ant were required to show an

improvement of the Adenovirus transfection effi-

H. Brooks et al. / Advanced Drug Delivery Reviews 57 (2005) 559–577 567

ciency [19]. Conversely, a simple mixture of the Tat

peptide either with a peptide [39] or an oligonucleo-

tide [64] do not lead to cargo internalisation. A simple

mixture with plasmid DNA was, however, sufficient

to increase the DNA transfection rate [65].

4.4. Influence of the Tat peptide density

The Tat peptide density also appears to be an

important issue in explaining the cellular delivery of

very large structures, such as particles, liposomes or

phages [21,28,36,66]. The requirement of 4 to 5 Tat

molecules to promote the uptake of native h-galactosidase [2] has been already mentioned (see

above). Although the cellular uptake of chimeras with

lower loading degrees was not investigated in this

latter study, it was shown in another report that one

single Tat peptide was sufficient to allow the cellular

delivery of an unfolded fusion construct of the same

protein [44]. The possible differences in the exposure

of the Tat peptide within both constructs have been

already discussed. Other studies, however, showed

that several Tat peptides (in some cases up to several

hundreds) attached at the surface of large particles

were required to promote efficient cellular delivery

[21,28,36,66]. Along these lines, another study was

carried out, in which six to seven Tat peptides were

coupled to large (45 nm diameter) ferromagnetic

particles, thus promoting their efficient delivery.

Again, a direct evaluation of the translocating

property in accordance with the number of attached

Tat peptides was not considered in this work [66] and

has not been further investigated. Considering the size

of such a structure, it is difficult to believe that a direct

passage through the plasma membrane was possible

when mediated only by six to seven Tat peptides

exposed at the particle surface. They do, however,

seem sufficient to induce an efficient cellular delivery,

since it was possible to recover cells through magnetic

collection [66]. Since no stringent washes were

performed prior to cell recovery, it cannot be excluded

that the ferromagnetic particles were simply strongly

bound to the cell surface. Confocal microscopy

pictures provided in this study could not discriminate

the cellular compartimentation of the particle since it

was performed after cell fixation and permeabilization

prior to observation. Whether the fixation step could

induce the cellular translocation of such a massive

structure as demonstrated for fusion proteins [67,68]

remains unknown.

Linear repeats from one to four Tat sequences have

also been used to increase the PEI-mediated trans-

fection of plasmids [25]. Important differences in the

transfection rate between the different complexes

indicated that the total content of guanidinium groups

per complex appeared to be important, since dimeric or

tetrameric repeats of linear Tat peptides complexed to

PEI improved the overall transfection efficiency [25].

Conversely, the use of a much longer homopolymer of

arginine (with a molecular weight ranging from 5000

to 15,000 Da, corresponding to a linear length of 50 to

150 residues) led to a minor increase in the transfection

efficiency of the PEI-plasmid [25]. Along these lines,

homopolymers containing 16 arginines were also

shown to be poorly taken up by cells when compared

to shorter homopolymer sequences [4]. On the other

hand, cellular plasmid delivery, after complexation to

branched-Tat peptide constructs, was shown to be

much more effective with larger peptide repeats than

the corresponding plasmid complexed with lower

branched-Tat structures [65]. Basically, branched

structure containing 8 Tat peptides showed a better

transfection efficacy than those containing 4 Tat

peptides, and even better results than those containing

2 or 1 Tat peptide [65]. Therefore, these data indicate

that the Tat peptide sequence requires an appropriate

number of arginines to efficiently translocate into cells

and that a given cargo requires an optimal number of

Tat peptides to be efficiently taken up by cells.

Whether these results could be influenced by differ-

ences in the experimental protocols, the cell type used

or the physicochemical properties of these cargo

molecules still remains to be tested.

The investigation of the influence of the Tat

peptide density around these massive structures will

likely highlight important features about the mecha-

nism of their translocation. One work, which, in

parallel experiments, investigated the influence of the

number of attached Tat peptides to the cargo, by

derivatizing Fab fragments with either 1.1 or 1.6 Tat

peptide molecules [69]. The Tat peptide used in this

study corresponded to a slightly different version of

the Tat peptide (Tat 37–62 encompassing also the

basic region), but appears relevant enough to be

discussed in this chapter. The cargo substituted with

more peptide was shown to be more efficiently taken

H. Brooks et al. / Advanced Drug Delivery Reviews 57 (2005) 559–577568

up by cells. However, when evaluated in vivo, this

chimera showed a weaker cell selectivity of the Fab

fragment per se. This result very likely reflects a

strong nonspecific cell association of the higher

loaded chimera directed by the highly cationic Tat

peptide, whereas the Tat peptide loading balance for

the weaker derivatized chimera still allowed the Fab

fragment to preferentially reach its target cells [69].

Along these lines, tumor growth inhibition induced

by Tat-Liposomes loaded with doxorubicin was also

evaluated in vivo on BALB/c mice [22]. About 50%

reduction of the tumor size was observed compared to

the liposome control . The benefit of the Tat peptide

was, however, not that important since bnudeQ doxor-ubicin-loaded liposomes (without Tat peptide) showed

identical, or better, tumor reduction [22]. This data also

indicates that the Tat peptide interferes with the

delivery of the liposome, probably by adhering locally

to other cells prior to reaching the targeted cells.

4.5. Influence of the serum

In contrast to the cationic lipid-mediated trans-

fection of oligonucleotides, serum has been shown to

unexpectedly augment the biological response when

oligonucleotides were delivered by Tat or Antenna-

pedia CPPs [70]. However, in later data using the

Kole system, where a delivered oligonucleotide

corrects an aberrant splice site resulting in expression

of a reporter gene [53], no serum effect was observed

[9]. In a recent study, serum was even shown to

decrease, but not to abolish, the biological effect of a

Tat-Cre fusion construct [11]. Unpublished data from

our group did not reveal a noticeable effect on the

cell-associated fluorescence when Tat peptide was

incubated with up to 50% serum. Although no trypsin

treatment was performed at that time prior FACS

analysis, a decreased signal should have been

observed if serum competes with the Tat peptide for

cell binding. To date, the observations of Tat peptide

uptake in the presence of serum remain promising for

future in vivo application.

5. Tat and cell surface interactions

Because of the highly cationic nature of the Tat

peptide, several anionic cellular candidates are avail-

able to influence the initial ionic cell surface

interactions. These interactions, or this binding to

the cell surface can, in part, be competitively inhibited

with heparin [11,24,31,71], along with heparin ana-

logues, such as PPS (pentosan polysulfate) [72], the

heparin-binding protein TSP (platelet thrombospon-

din-1) [73] and other soluble polyanions, such as

suramin, suramin derivatives [74] dextran sulfate [75]

and CS/DS chondroitin/dermatan sulfates [29].

Many initial studies of internalisation of the Tat

peptide, although now known to be false or

compromised in terms of internalisation, can reveal

interesting aspects of binding. Where no biological

activity is used as a control for effective entry and

delivery, externally bound peptide in many cases has

been confused with effective delivery. FACS analysis

is particularly susceptible to giving artificially high

fluorescent values. Given that standard wash techni-

ques in an isotonic buffer, such as the often used

PBS, leave residually bound peptide [2,5], a more

stringent treatment is required before analysis can be

performed.

The initial association of the Tat peptide with the

cell surface membrane occurs independently of

temperature, is resistant to mechanical washing with

isotonic buffers, such as PBS/EDTA [2], but is

sensitive to treatment with proteases, such as trypsin

[5]. Trypsinisation of externally bound peptide is,

therefore, an often preferred alternative to mechanical

washing. Yet again, to what extent this treatment or

washes with acidic buffers, high salt solutions or

competing substances, such as heparin, are effective

in disrupting ionic interactions has not been quanti-

tatively and comparatively studied for different cell

lines.

Both the full length GST-Tat protein and fusion

proteins containing the basic domain only require a

high ionic strength (1.6–1.3 M NaCl) to elute them

from bound heparin [74]. In contrast, the substitution

of six arginine residues within the basic domain using

alanine residues reduced the required ionic strength to

only 0.3 M NaCl, once again highlighting the

importance of these basic residues in the ionic binding

profile of the Tat peptide. With respect to the use of

high salt washes to eliminate excess peptide bound to

the surface of cells, a 2-M wash, as used in some

protocols [29], would be thought sufficient. Suzuki

and co-workers, however, reported that this was not

H. Brooks et al. / Advanced Drug Delivery Reviews 57 (2005) 559–577 569

the case in their study [20]. Washing the cells with a

high salt buffer [20 mM HEPES containing 2 M NaCl

(pH 7.4)] produced little difference in the amount of

peptides bound to the cell surface compared with PBS

alone [20].

In cases where the internalisation of the peptide has

not been distinguished from strong extracellular

attachment, the data might still be useful with respect

to the analysis of binding. In one example, the uptake

of Tat peptide was examined in a range of cells [31]

using FACS analysis after washing the cells simply

with EDTA, thereby looking at EDTA resistant bound

peptide (along with any internalised). Competing

anionic compounds, such as heparin and dextran

sulfate, caused a significant decrease (60–70%) in

cell-surface association of the peptide, whereas other

glycosaminoglycans, such as chondroitin sulfate (CS)

A, B, and C and hyaluronic acid, had no effect. If

anything, it was noted that chondroitin sulfate A

increased Tat-cell association. A similar study looking

at GST-Tat-GFP in CHO cells examined only the

trypsin resistant fraction of whole protein-GFP and

yet came to similar conclusions with regard to the type

of glycosaminoglycans (GAG) interacting with Tat.

Internalisation of the Tat protein was inhibited by HS,

but not by the chondroitin sulfates [76]. Suzuki et al.,

however, using HeLa cells washed only with PBS and

then fixed with acetone:methanol, found that the basic

peptide (Tat48–60) was inhibited in its binding to

HeLa cells by all GAGs [20]. The uptake was

significantly reduced in the presence of heparin sulfate

or chondroitin sulfates A, B, and C, as well as by pre-

treatment of the cells with the anti-heparin sulfate

antibody or heparinase III [20]. Looking at the

biological activity of an effectively delivered Cre

recombinase, it was shown that heparin was able to

confer a total inhibition at low concentrations (2.5 Ag/ml), followed by CS-B at only slightly higher amounts

[11]. CS-C, however, was only able to inhibit

recombination by 80% at 20-fold higher concentra-

tions of the inhibitor and CS-A showed only 40%

inhibition at all concentrations tested. This study

effectively showed the competition for cell surface

binding of the Tat-Cre construct with the free GAG

and the specificity of this cell surface interaction. It

was observed very early on for the full-length Tat

protein that uptake and Tat-promoted transactivation

of HIV-1 gene expression could be blocked by soluble

polyanions (heparin and dextran sulfate) [75]. Sur-

prisingly, the same study could not show an effect for

trypsinisation or heparinase treatment.

The Tat protein binds to cell surface heparan

sulfate (HS) and heparin [77], and consistent with

its heparin-binding properties, Tat can be purified to

homogeneity by heparin-affinity chromatography

[78]. The group of Presta and co-workers first

examined the role of the basic domain in the binding

of Tat to heparin. They found that neutralization of the

positive charges in the basic domain of Tat signifi-

cantly reduces its interaction with the GAG. The

dissociation constant of heparin to immobilised GST-

Tat was observed to be around 0.3 AM [74].

Work in CHO cells demonstrated that cell uptake

and association of Tat constructs containing the basic

peptide were effectively blocked by heparin [24,76],

pre-treatment of HeLa cells with heparinase III [20] or

pre-treatment of CHOs with glycosaminoglycan

lyases that specifically degrade HS chains ([76] and

Melikov et al., submitted for publication). Given

the homology of heparin to the surface sulfated

glycosaminoglycans, the observed Tat–heparin inter-

action could reflect the initial cell surface interactions

of Tat with exposed surface HS proteoglycans,

perhaps serving as the initial point of contact or even

a route of entry for Tat, as observed for some other

heparin-binding proteins. Sulfated glycosaminogly-

cans (GAGs), such as HS, increasingly implicated in

cell adhesion, are distributed ubiquitously on cell

surfaces as the carbohydrate component of proteogly-

cans [79].

Many microbial and viral particles enter cells using

HS receptors via a two-step process, adhering to the

cell surface by binding initially to GAGs followed by

internalisation. The foot-and-mouth disease virus

(FMDV) infects cells in such a way, its primary

contact being with a low-affinity HS proteoglycan

receptor, followed by transfer to the high-affinity

integrin receptor for endocytosis [80]. HS facilitates

entry of the FMDV and it was found that alteration of

the HS affinity had profound consequences for the

infectivity of the virus [81]. HSPG serve as cell

surface receptors for a number of natural ligands,

some of which include matrix proteins, such as

laminin, cell adhesion molecules (N-CAM) and

growth factors, such as fibroblast growth factor

(FGF) [82], insulin-like growth factor-binding pro-

H. Brooks et al. / Advanced Drug Delivery Reviews 57 (2005) 559–577570

tein-2 (IGFBP-2) [83] and vascular endothelial cell

growth factor (VEGF) [84].

The basic domain of Tat shown to be responsible

for the Tat–heparin interaction [74] has homology

with heparin-binding growth factors [85]. Structurally,

there appears to be no conserved conformation for this

domain, clusters of basic residues and a heparin-

binding capacity using heparin-sepharose chromatog-

raphy serve as common criteria when labelling a

peptide domain as heparin binding [86]. Studies are

yet to be done demonstrating direct competition of Tat

with natural ligands, such as the heparin-binding

proteins growth factors mentioned above, for the

binding sites of these cell surface receptors. This

would potentially provide useful information about

the initial binding and entry of the Tat peptide.

Binding to HSPGs is often followed by rapid

internalisation via endocytosis. There are multiple

proposals for the mechanism of internalisation via

such proteoglycans, ranging from simple endocytosis

via classical clathrin pathways [87] to alternative

routes, such as those mediated by the syndecan

HSPGs, those independent of coated pits or those

utilizing a much slower pathway of internalisation, as

was recently described for the perlecans [88]. In the

case of syndecan HSPGs, efficient internalisation is

triggered by a clustering of transmembrane and

cytoplasmic domains and then proceeds via a non-

coated pit pathway, possibly caveolae [88,89].

When CHO cells were pre-treated with chondroitin

ABC lyase to eliminate CS/DSPGs or heparitinase/

heparinase to cleave HS chains, all proteolytic treat-

ment resulted in a significant reduction in the uptake

of Tat peptide [29]. Likewise, treatment of CHO cells

with chlorate (which inhibits GAG sulfation) had a

similar inhibitory effect [29]. Mutant cells defective

for GAG synthesis show dramatically reduced TAT-

mediated transmembrane transport [24,31,76]. The

cell line CHO pgs D-677, which does not produce

HSPGs (due to a 10-fold reduction in N-acetylgluco-

saminyl-transferase and glucuronosyltransferase), pro-

duces chondroitin sulfates in excess of about a 3-fold

[90]. According to the work of Tyagi et al., these cells

show a 50% reduction in transactivation by recombi-

nant GST-Tat [76], equal to that observed in E-606

cells, which produce HSPGs that are undersulfated,

thereby indicating the importance of the sulfation step.

By contrast, the complete proteoglycan null mutant

pgs A-745 (deficient in xylosyl transferase, which

catalyses the first step in PG assembly/formation [91])

lacks all surface PGs and show a much severer

reduction in Tat uptake of around 80% [76]. CHO 745

cells also show no cell membrane adhesion of the

basic peptide or vesicle inclusion using confocal

microscopy [24], and show a reduced uptake of Tat

peptide/HS complexes [29], except where the ratios of

peptide to anion are particularly high (i.e., high excess

of peptide). Thus, the internalisation was if anything

more important in the HSPG-deficient cells.

In a study examining only EDTAwashed cells, the

milder D-677 mutant was not observed to have any

difference in its binding of Tat when compared to wild

type [31]. The severe A-745 mutant, however, showed

a reduction of 80–90%. Identical work in our

laboratory comparing GST-Tat-GFP and fluorescein-

labelled Tat peptide in only PBS-washed cells showed

that binding of the fusion protein was completely

inhibited in mutants compared to wild type, whereas

the peptide adhered to all three cell lines regardless of

their GAG expression [92]. The initial attachment of

Tat peptide to GAGs or any other molecule at the cell

surface would likely be influenced by an attached or

previously bound cargo (see above). The size of the

cargo, the overall charge involved, the way in which it

was coupled (N- versus C-terminal binding, covalent

binding, fusion protein or chemical coupling as well

as pure electrostatic interactions), and the degree of

exposure of the basic residues would all play a role in

influencing these initial cell-surface associations.

Aside from the known affinity for HSPGs, other

cell surface receptors have also been implicated in Tat

binding. Yeast 2-hybrid screens controlled with

subsequent GST pull down assays confirmed the

binding of the full-length Tat protein to both HSPGs

and LRP (low-density lipoprotein receptor family)

[93]. They also confirmed that the domain responsible

(amino acids 34–48) was just before the cluster of

basic residues (49–57), meaning that any elongation

of this minimal sequence to include the a-helical-like

structure located just prior to the translocating domain

might result in differences in affinity and the pathway

of internalisation.

Tat and its basic domain have been proposed to

bind many cell surface receptors. One study in

particular showed the Tat peptide as causing the

release of acetylcholine from human and rat chol-

H. Brooks et al. / Advanced Drug Delivery Reviews 57 (2005) 559–577 571

inergic terminals [94]. The release was dependent on

calcium, effected through voltage sensitive calcium

channels, and inhibited strongly by cadmium, as well

as the mGluR and IP3R antagonists (heparin and

xestosponginC). Further studies showed immunopre-

cipitation of the Tat peptide with various anti-

integrin antibodies suggesting that the vitronectin-

binding integrin (alphaVbeta5) is the cell surface

protein responsible for binding to the basic domain

of Tat [95]. A natural ligand of this receptor,

vitronectin, also contains a related basic peptide

sequence (KKQRFRHRNRKG) in its heparin-bind-

ing domain, which served to competitively inhibit

binding.

Another group showed in the same year that

antibodies to the h-4 integrin subunit were able to

inhibit cell attachment to Tat specifically [96], yet

were unable to demonstrated co-precipitation. They

found instead a strong relation to a 90-kDa surface

membrane protein in both Molt3 and PC12cells.

Although HSPGs can resolve at around this size, they

vary considerably depending on the saccharide chain

length from 12 kDa for the HS chain, to 61 kDa for

the core protein, and 90–190 kDa for the intact PG

[97]. Rusnati et al. found that a positive correlation

existed between the size of heparin oligosaccharides

and their capacity to inhibit the internalisation of Tat

[98]. Given the size heterogeneity of HSPGs, it is

most likely that the observed 90-kDa protein is an

additional separate factor in Tat cell membrane

adhesion.

Taken together, studies on the binding of Tat would

implicate more than one component involved in initial

cell membrane attachment. A strong argument for the

role of GAGs, in particular HS, has been assembled

from the data of many independent studies; however,

the lack of complete inhibition by mutant, enzymatic

digestion or competition studies would tend to

preclude their exclusivity.

6. Possible mechanisms of internalisation

Membrane association or binding occurs at any

temperature, including the metabolically inhibiting

48C. In traversing the extracellular membrane, how-

ever, the Tat peptide behaves in an energy-dependent

manner requiring temperatures above 4 8C and ATP.

Only one group [18] continues to observe uptake at 4

8C with a modified Tat peptide [Tat-(48–60:P59W)].

The initial association is followed by a rapid

translocation to the cytosolic side most probably

within vesicle-like structures, of which some at least

are acidified according to colocalisation studies with

pH markers [29] or inhibition of vesicle acidification

[Melikov et al., submitted for publication]. The

fluorescence of labelled Tat peptide when observed

in live cells is often described as being punctate or

vesicular [24,29,75] and more rarely as diffuse

cytosolic [18,34]. It has been observed to be close

to the membrane at early time points, progressing to

larger aggregations with a more perinuclear type

pattern at longer time points and can be observed to

continue on into the nucleus [2,29,99].

Various drugs have been shown to affect the entry

or distribution of the Tat peptide. For instance, Golgi

destabilisation of HeLa cells (brefeldinA) converts the

punctate vesicular staining to a more cytoplasmic,

even distribution, while having no effect on control

dextran [100]. Ammonium chloride halted the stain-

ing of the nucleus, but appeared to have no effect on

the vesicular pattern, leading the authors to conclude

that the basic peptide is normally released from

vesicles after endosomal uptake by means of a

mechanism requiring endosome acidification [99].

Chloroquine, another inhibitor of endosomal acid-

ification appeared likewise to inhibit the release of Tat

from vesicles [100], enhancing the vesicular staining

for both Arg9 and Tat peptides and reducing/

eliminating any diffuse cytoplasmic staining. No

effect on Tat was seen with a similar drug, bafilomy-

cinA, nor of wortmannin, a potent PI-kinase inhibitor.

In our hands, the drug monensin, which causes de-

acidification of cytoplasmic compartments, resulted in

an increase in the fluorescent signal of FITC-labelled

Tat, indicating that the internalised label had been

sequestered in acidic compartments, which masked

the level of fluorescence [Melikov et al., submitted for

publication]. In a nice experiment using the similar

Arg9 peptide, the authors co-cultured different cells

and then showed that distribution of the peptide was

different according to cell type (vesicular and cyto-

solic in MC57 cells, while only vesicular in HeLa

cells) [100].

CytochalasinD is known to depolymerise the actin

cytoskeleton causing clustering of caveolae at the cell

H. Brooks et al. / Advanced Drug Delivery Reviews 57 (2005) 559–577572

surface. A construct of Tat, namely, GST-Tat-EGFP,

was found to be restricted to the plasma membrane

area following exposure to cytochalasinD, whereas

nocodazole treatment, which preferentially disrupts

the actin microtubules, resulted in perinuclear fluo-

rescence similar to untreated controls [24]. Trans-

ferrin, often used as a marker for early endosomes

following the clathrin-coated vesicular pathway, has

been observed to partially colocalise with the Tat

peptide [5,31,99]. The same has been observed for the

non-clathrin markers, such as cholera toxin [24,71] or

the SV40 virus [21], which are internalised via

caveolin-cholesterol rich domains.

Inhibition of the caveolin pathway by the drug

nystatin reduced the Tat peptide reporter h-galactivity by 50% in CHO and HepG2 (a rather

surprising finding when it has been reported that

HepG2 cells lack Cav-1 [101]), but showed some cell

specificity, having no effect on buffalo green monkey

(BGM) cells [34]. Nystatin also inhibited Tat-phage-

mediated gene transfer up to 50% of control values,

whereas DEAE-dextran-mediated gene transfer

remained unaffected [21].

The kinetics of internalisation of a GST-Tat-GFP

conjugate and the cholera toxin were reported to be far

slower than the comparatively rapid internalisation of

transferrin [71]. The peptide conjugate showed no

evidence of co-localisation with either transferrin, the

marker EE1 (early endosome antigen-1) or lysotracker

dye. The authors concluded that for their conjugate at

least, Tat was internalised via a non-clathrin-depend-

ent route, possibly caveolae. A large number of, but

not all, vesicles containing Tat were positive for

Caveolin-1. Tat uptake was also shown to be inhibited

by the sequestration of cholesterol by methyl-h-cyclodextrin [71]. It should be noted, however, that

uptake of the GST-Tat-GFP conjugates were studied

in the presence of 100 AM chloroquine (Tat protein (1

Ag/ml)) [76,77], which, although used as a lysosomal

trophic agent to reduce degradation of the parent Tat

protein [102], is nonetheless going to interfere with

vesicular recycling and affect the subcellular local-

isation of both control markers and the Tat peptide.

The pH neutral environment of caveolae would, in

any case, conflict with data obtained regarding the

acidic nature of at least some of the Tat-containing

vesicles and the partial colocalisation observed with

transferrin for non-conjugated peptide.

The data on Tat and Tat-cargoes would tend to

preclude the dominance of one exclusive pathway of

entry into the cell. The lack of complete inhibition by

selective drugs or complete colocalisation with known

markers strongly suggests a multiplicity of entry

pathways for this sticky basic peptide. Aside from

clathrin and caveolae, other mechanisms of crossing

the plasma membrane include the non-clathrin/non-

caveolin-type pathway(s) (lipid rafts or microdo-

mains, e.g., IL-2 receptor), macropinocytosis (platelet

derived growth factor), potocytosis (folate receptor)

and phagocytosis (specialised cells only).

Perhaps we must also entertain the idea that Tat is

simply an opportunistic peptide, adhering strongly to

the cell surface on the basis of its charge to any

negative offerings, such as lipids or proteins, and then

being internalised through natural cell membrane

recycling on regions or microdomains, presumably

captured by any type of endocytic vesicle.

Cell plasma membrane turnover continues con-

stitutively at an estimated rate of ~2%/min [103], in

other estimates as fast as 5%/min [104], meaning

100% of the cells surface is internalised nonspecifi-

cally in less than an hour, notwithstanding the faster

receptor mediated or stimulated routes of uptake. To

this end, most drugs inhibiting cell membrane

recycling will show an effect on Tat uptake, yet not

prevent it entirely as long as other possible routes

exist. Competitors for cell surface binding, however,

would presumably be more effective at reducing the

level of peptide internalised as is seen for heparin and

the similar glycosaminoglycans.

7. Conclusions

The Tat peptide delivery strategy is now widely

used to improve the cellular delivery of a very large

panel of cargo molecules. The increase of the

biological response of peptides, proteins or oligonu-

cleotides upon their coupling to the Tat peptide has

been assessed in several recent studies, although the

precise mechanism of entry is far from being firmly

identified. Major controversies still exist in the

literature, probably as the direct result of previously

reported misleading results, due to the presence of

fixation artifacts, even in mild conditions, affecting

molecules containing a strong cationic cluster of

H. Brooks et al. / Advanced Drug Delivery Reviews 57 (2005) 559–577 573

amino acids, such as the Tat peptide. Nowadays, it

seems that most of the studies are performed

following uptake on live (unfixed) cells.

As discussed in this review, comparisons between

the different reports relative to the transducing

properties of the Tat peptide are very often difficult,

because of the very wide diversity of the Tat

sequence used as a cell-penetrating peptide, although

it seems very likely that the Tat peptide correspond-

ing to its basic domain showed the best ubiquitous

positive effect. More importantly, the very large

differences within the cargoes, in terms of size (from

some hundreds of daltons to massive particles up to

200 nm of diameter), in terms of composition and,

therefore, of physicochemical properties, make it

very difficult to extract clear information allowing

the definition of a universal mechanism of entry.

Moreover, some reports revealed that additional

parameters involving the cargo could be very

important. These include the type of linkage

between the peptide and the cargo, its length, the

orientation of the peptide relative to the cargo, the

quantity and peptide exposure at the overall surface

of the cargo.

In addition, most of the experiments aimed at

defining the mechanism of entry of this Tat-derived

cell-penetrating peptide have been performed with a

large heterogeneity of cell types. Altogether, we

counted about 50 different cell types used for

studying the Tat delivery process, from primary cells

to established cell lines, derived from various species

or tissues, and studied either in vitro or in vivo. The

experiments were often performed under variable

conditions in terms of kinetics (from minutes to

hours), concentration of the chimera (from nano-

molar to hundreds of micromolars) and protocols

applied to estimate the uptake efficiency or the

subcellular localisation of the chimeras (from bio-

logical activity to fluorescence detection). When

drugs were used to block different entry pathways

proposed as candidates for Tat-mediated cellular

entry of cargo, an absolute blockade of the entry

or the full biological inhibition has not been always

fully achieved, excluding the possibility that a

discrete entry process could also be involved.

Moreover, the detection of such a phenomenon is

always limited in each application by detection limits

of the tools available.

This Tat peptide (and probably some of the

arginine polymers shown to be closely related to

Tat) was found to induce the cellular uptake of very

different molecules. It appears however, that this

process will suffer from a lack of cellular specificity

since it seems that it is effective in a very large

number of different cell types. Anionic structures at

the cell surface are probably nonspecific agents

interacting with the Tat peptide to increase the local

concentration of the Tat-bound cargo at the cell

surface before allowing its cellular entry through

general endocytosis pathways. Conditions inducing

the entry through any of the possible routes (caveolae,

clathrin-dependant endocytosis, macropinocytosis,

fluid-phase endocytosis. . .) are far from being fully

understood and will certainly require a complete study

with regard to the possible influences of all the

various parameters discussed here.

References

[1] S. Ruben, A. Perkins, R. Purcell, K. Joung, R. Sia, R.

Burghoff, W.A. Haseltine, C.A. Rosen, Structural and func-

tional characterization of human immunodeficiency virus tat

protein, J. Virol. 63 (1989) 1–8.

[2] S. Fawell, J. Seery, Y. Daikh, C. Moore, L.L. Chen, B.

Pepinsky, J. Barsoum, Tat-mediated delivery of heterologous

proteins into cells, Proc. Natl. Acad. Sci. U. S. A. 91 (1994)

664–668.

[3] E. Vives, P. Brodin, B. Lebleu, A truncated HIV-1 Tat protein

basic domain rapidly translocates through the plasma

membrane and accumulates in the cell nucleus, J. Biol.

Chem. 272 (1997) 16010–16017.

[4] S. Futaki, T. Suzuki, W. Ohashi, T. Yagami, S. Tanaka, K.

Ueda, Y. Sugiura, Arginine-rich peptides. An abundant

source of membrane-permeable peptides having potential as

carriers for intracellular protein delivery, J. Biol. Chem. 276

(2001) 5836–5840.

[5] J.P. Richard, K. Melikov, E. Vives, C. Ramos, B. Verbeure,

M.J. Gait, L.V. Chernomordik, B. Lebleu, Cell-penetrating

peptides. A reevaluation of the mechanism of cellular uptake,

J. Biol. Chem. 278 (2003) 585–590.

[6] M.A. Lindsay, Peptide-mediated cell delivery: application in

protein target validation, Curr. Opin. Pharmacol. 2 (2002)

587–594.

[7] M.J. Gait, Peptide-mediated cellular delivery of antisense

oligonucleotides and their analogues, Cell. Mol. Life Sci. 60

(2003) 844–853.

[8] M. Zhao, R. Weissleder, Intracellular cargo delivery using tat

peptide and derivatives, Med. Res. Rev. 24 (2004) 1–12.

[9] A. Astriab-Fisher, D. Sergueev, M. Fisher, B.R. Shaw, R.L.

Juliano, Conjugates of antisense oligonucleotides with the

Tat and antennapedia cell-penetrating peptides: effects on

H. Brooks et al. / Advanced Drug Delivery Reviews 57 (2005) 559–577574

cellular uptake, binding to target sequences, and biologic

actions, Pharm. Res. 19 (2002) 744–754.

[10] C. Bonny, A. Oberson, S. Negri, C. Sauser, D.F. Schorderet,

Cell-permeable peptide inhibitors of JNK: novel blockers of

beta-cell death, Diabetes 50 (2001) 77–82.

[11] J.S. Wadia, R.V. Stan, S.F. Dowdy, Transducible TAT-HA

fusogenic peptide enhances escape of TAT-fusion proteins

after lipid raft macropinocytosis, Nat. Med. 10 (2004)

310–315.

[12] S. Futaki, I. Nakase, T. Suzuki, Z. Youjun, Y. Sugiura,

Translocation of branched-chain arginine peptides through

cell membranes: flexibility in the spatial disposition of

positive charges in membrane-permeable peptides, Biochem-

istry 41 (2002) 7925–7930.

[13] E.P. Loret, E. Vives, P.S. Ho, H. Rochat, J. Van Rietschoten,

W.C. Johnson Jr., Activating region of HIV-1 Tat protein:

vacuum UV circular dichroism and energy minimization,

Biochemistry 30 (1991) 6013–6023.

[14] D. Derossi, A.H. Joliot, G. Chassaing, A. Prochiantz, The

third helix of the Antennapedia homeodomain translocates

through biological membranes, J. Biol. Chem. 269 (1994)

10444–10450.

[15] S.D. Kramer, H. Wunderli-Allenspach, No entry for TAT(44–

57) into liposomes and intact MDCK cells: novel approach to

study membrane permeation of cell-penetrating peptides,

Biochim. Biophys. Acta 1609 (2003) 161–169.

[16] J.M. Sabatier, E. Vives, K. Mabrouk, A. Benjouad, H.

Rochat, A. Duval, B. Hue, E. Bahraoui, Evidence for

neurotoxic activity of tat from human immunodeficiency

virus type 1, J. Virol. 65 (1991) 961–967.

[17] D. Derossi, S. Calvet, A. Trembleau, A. Brunissen, G.

Chassaing, A. Prochiantz, Cell internalization of the third

helix of the Antennapedia homeodomain is receptor-inde-

pendent, J. Biol. Chem. 271 (1996) 18188–18193.

[18] P.E. Thoren, D. Persson, P. Isakson, M. Goksor, A. Onfelt, B.

Norden, Uptake of analogs of penetratin, Tat(48–60) and

oligoarginine in live cells, Biochem. Biophys. Res. Commun.

307 (2003) 100–107.

[19] J.P. Gratton, J. Yu, J.W. Griffith, R.W. Babbitt, R.S. Scotland,

R. Hickey, F.J. Giordano, W.C. Sessa, Cell-permeable

peptides improve cellular uptake and therapeutic gene

delivery of replication-deficient viruses in cells and in vivo,

Nat. Med. 9 (2003) 357–363.

[20] T. Suzuki, S. Futaki, M. Niwa, S. Tanaka, K. Ueda, Y.

Sugiura, Possible existence of common internalization

mechanisms among arginine-rich peptides, J. Biol. Chem.

277 (2002) 2437–2443.

[21] A. Eguchi, T. Akuta, H. Okuyama, T. Senda, H. Yokoi, H.

Inokuchi, S. Fujita, T. Hayakawa, K. Takeda, M. Hasegawa,

M. Nakanishi, Protein transduction domain of HIV-1 Tat

protein promotes efficient delivery of DNA into mammalian

cells, J. Biol. Chem. 276 (2001) 26204–26210.

[22] Y.L. Tseng, J.J. Liu, R.L. Hong, Translocation of liposomes

into cancer cells by cell-penetrating peptides penetratin and

tat: a kinetic and efficacy study, Mol. Pharmacol. 62 (2002)

864–872.

[23] U. Koppelhus, S.K. Awasthi, V. Zachar, H.U. Holst, P.

Ebbesen, P.E. Nielsen, Cell-dependent differential cellular

uptake of PNA, peptides, and PNA-peptide conjugates,

Antisense Nucleic Acid Drug Dev. 12 (2002) 51–63.

[24] A. Ferrari, V. Pellegrini, C. Arcangeli, A. Fittipaldi, M.

Giacca, F. Beltram, Caveolae-mediated internalization of

extracellular HIV-1 tat fusion proteins visualized in real time,

Molec. Ther. 8 (2003) 284–294.

[25] C. Rudolph, C. Plank, J. Lausier, U. Schillinger, R.H. Muller,

J. Rosenecker, Oligomers of the arginine-rich motif of the

HIV-1 TAT protein are capable of transferring plasmid DNA

into cells, J. Biol. Chem. 8 (2003) 8.

[26] J.C. Mai, H. Shen, S.C. Watkins, T. Cheng, P.D. Robbins,

Efficiency of protein transduction is cell type-dependent and

is enhanced by dextran sulfate, J. Biol. Chem. 277 (2002)

30208–30218.

[27] P.O. Falnes, J. Wesche, S. Olsnes, Ability of the Tat basic

domain and VP22 to mediate cell binding, but not membrane

translocation of the diphtheria toxin A-fragment, Biochem-

istry 40 (2001) 4349–4358.

[28] A. Nori, K.D. Jensen, M. Tijerina, P. Kopeckova, J. Kopecek,

Tat-conjugated synthetic macromolecules facilitate cytoplas-

mic drug delivery to human ovarian carcinoma cells,

Bioconjug. Chem. 14 (2003) 44–50.

[29] S. Sandgren, F. Cheng, M. Belting, Nuclear targeting of

macromolecular polyanions by an HIV-Tat derived peptide.

Role for cell-surface proteoglycans, J. Biol. Chem. 277

(2002) 38877–38883.

[30] R. Bhorade, R. Weissleder, T. Nakakoshi, A. Moore, C.H.

Tung, Macrocyclic chelators with paramagnetic cations are

internalized into mammalian cells via a HIV-tat derived

membrane translocation peptide, Bioconjug. Chem. 11

(2000) 301–305.

[31] S. Console, C. Marty, C. Garcia-Echeverria, R. Schwendener,

K. Ballmer-Hofer, Antennapedia and HIV transactivator of

transcription (TAT) bprotein transduction domainsQ promote

endocytosis of high molecular weight cargo upon binding to

cell surface glycosaminoglycans, J. Biol. Chem. 278 (2003)

35109–35114.

[32] E.L. Snyder, B.R. Meade, S.F. Dowdy, Anti-cancer protein

transduction strategies: reconstitution of p27 tumor suppres-

sor function, J. Control. Release 91 (2003) 45–51.

[33] V.P. Torchilin, T.S. Levchenko, TAT-Liposomes: a novel

intracellular drug carrier, Curr. Protein Pept. Sci. 4 (2003)

133–140.

[34] I.A. Ignatovich, E.B. Dizhe, A.V. Pavlotskaya, B.N.

Akifiev, S.V. Burov, S.V. Orlov, A.P. Perevozchikov,

Complexes of plasmid DNA with basic domain 47–57 of

the HIV-1 Tat protein are transferred to mammalian cells by

endocytosis-mediated pathways, J. Biol. Chem. 278 (2003)

42625–42636.

[35] S. Violini, V. Sharma, J.L. Prior, M. Dyszlewski, D. Piwnica-

Worms, Evidence for a plasma membrane-mediated perme-

ability barrier to Tat basic domain in well-differentiated

epithelial cells: lack of correlation with heparan sulfate,

Biochemistry 41 (2002) 12652–12661.

H. Brooks et al. / Advanced Drug Delivery Reviews 57 (2005) 559–577 575

[36] V.P. Torchilin, T.S. Levchenko, R. Rammohan, N. Volodina,

B. Papahadjopoulos-Sternberg, G.G. D’Souza, Cell trans-

fection in vitro and in vivo with nontoxic TAT peptide-

liposome-DNA complexes, Proc. Natl. Acad. Sci. U. S. A.

100 (2003) 1972–1977.

[37] V.P. Torchilin, R. Rammohan, V. Weissig, T.S. Levchenko,

TAT peptide on the surface of liposomes affords their

efficient intracellular delivery even at low temperature and

in the presence of metabolic inhibitors, Proc. Natl. Acad. Sci.

U. S. A. 98 (2001) 8786–8791.

[38] P.E. Thoren, D. Persson, M. Karlsson, B. Norden, The

antennapedia peptide penetratin translocates across lipid

bilayers—the first direct observation, FEBS Lett. 482

(2000) 265–268.

[39] E. Vives, C. Granier, P. Prevot, B. Lebleu, Structure activity

relationship study of the plasma membrane translocating

potential of a short peptide from HIV-1 Tat protein, Lett.

Pept. Sci. 4 (1997) 429–436.

[40] P.A. Wender, D.J. Mitchell, K. Pattabiraman, E.T. Pelkey, L.

Steinman, J.B. Rothbard, The design, synthesis, and evalua-

tion of molecules that enable or enhance cellular uptake:

peptoid molecular transporters, Proc. Natl. Acad. Sci. U. S. A.

97 (2000) 13003–13008.

[41] D.J. Mitchell, D.T. Kim, L. Steinman, C.G. Fathman, J.B.

Rothbard, Polyarginine enters cells more efficiently than

other polycationic homopolymers, J. Pept. Res. 56 (2000)

318–325.

[42] L.R. Wright, J.B. Rothbard, P.A. Wender, Guanidinium rich

peptide transporters and drug delivery, Curr. Protein Pept.

Sci. 4 (2003) 105–124.

[43] O. Zelphati, F.J. Szoka, Mechanism of oligonucleotide

release from cationic liposomes, Proc. Natl. Acad. Sci.

U. S. A. 93 (1996) 11493–11498.

[44] S.R. Schwarze, A. Ho, A. Vocero-Akbani, S.F. Dowdy, In

vivo protein transduction: delivery of a biologically active

protein into the mouse, Science 285 (1999) 1569–1572.

[45] S.R. Schwarze, S.F. Dowdy, In vivo protein transduction:

intracellular delivery of biologically active proteins,

compounds and DNA, Trends Pharmacol. Sci. 21 (2000)

45–48.

[46] J.S. Wadia, S.F. Dowdy, Protein transduction technology,

Curr. Opin. Biotechnol. 13 (2002) 52–56.

[47] R. Montesano, J. Roth, A. Robert, L. Orci, Non-coated

membrane invaginations are involved in binding and internal-

ization of cholera and tetanus toxins, Nature 296 (1982)

651–653.

[48] L. Pelkmans, J. Kartenbeck, A. Helenius, Caveolar endocy-

tosis of simian virus 40 reveals a new two-step vesicular-

transport pathway to the ER, Nat. Cell Biol. 3 (2001)

473–483.

[49] J. Alblas, L. Ulfman, P. Hordijk, L. Koenderman, Activation

of Rhoa and ROCK are essential for detachment of migrating

leukocytes, Mol. Biol. Cell 12 (2001) 2137–2145.

[50] H.M. Moulton, M.H. Nelson, S.A. Hatlevig, M.T. Reddy,

P.L. Iversen, Cellular uptake of antisense morpholino

oligomers conjugated to arginine-rich peptides, Bioconjug.

Chem. 15 (2004) 290–299.

[51] L.L. Chen, A.D. Frankel, J.L. Harder, S. Fawell, J. Barsoum,

B. Pepinsky, Increased cellular uptake of the human

immunodeficiency virus-1 Tat protein after modification with

biotin, Anal. Biochem. 227 (1995) 168–175.

[52] S. Futaki, W. Ohashi, T. Suzuki, M. Niwa, S. Tanaka, K.

Ueda, H. Harashima, Y. Sugiura, Stearylated arginine-rich

peptides: a new class of transfection systems, Bioconjug.

Chem. 12 (2001) 1005–1011.

[53] P. Sazani, S.H. Kang, M.A. Maier, C. Wei, J. Dillman, J.

Summerton, M. Manoharan, R. Kole, Nuclear antisense

effects of neutral, anionic and cationic oligonucleotide

analogs, Nucleic Acids Res. 29 (2001) 3965–3974.

[54] S. Futaki, Arginine-rich peptides: potential for intra-

cellular delivery of macromolecules and the mystery of

the translocation mechanisms, Int. J. Pharm. 245 (2002)

1–7.

[55] J. Kyte, R.F. Doolittle, A simple method for displaying the

hydropathic character of a protein, J. Mol. Biol. 157 (1982)

105–132.

[56] M. Hallbrink, A. Floren, A. Elmquist, M. Pooga, T. Bartfai,

U. Langel, Cargo delivery kinetics of cell-penetrating

peptides, Biochim. Biophys. Acta 1515 (2001) 101–109.

[57] F. Albericio, D. Andreu, E. Giralt, C. Navalpotro, E. Pedroso,

B. Ponsati, M. Ruiz-Gayo, Use of the Npys thiol protection

in solid phase peptide synthesis. Application to direct

peptide–protein conjugation through cysteine residues, Int.

J. Pept. Protein Res. 34 (1989) 124–128.

[58] M.S. Bernatowicz, R. Matsueda, G.R. Matsueda, Preparation

of Boc-[S-(3-nitro-2-pyridinesulfenyl)]-cysteine and its use

for unsymmetrical disulfide bond formation, Int. J. Pept.

Protein Res. 28 (1986) 107–112.

[59] E. Vives, B. Lebleu, Selective coupling of a highly basic

peptide to an oligonucleotide, Tetrahedron Lett. 38 (1997)

1183–1186.

[60] B. Lebleu, I. Robbins, L. Bastide, E. Vives, J.E. Gee,

Pharmacokinetics of oligonucleotides in cell culture, Ciba

Found. Symp. 209 (1997) 47–54. discussion 54-49.

[61] T. Zhu, Z. Wei, C.H. Tung, W.A. Dickerhof, K.J. Breslauer,

D.E. Georgopoulos, M.J. Leibowitz, S. Stein, Oligonucleo-

tide-poly-l-ornithine conjugates: binding to complementary

DNA and RNA, Antisense Res. Dev. 3 (1993) 265–275.

[62] Z. Wei, C.H. Tung, T. Zhu, W.A. Dickerhof, K.J. Breslauer,

D.E. Georgopoulos, M.J. Leibowitz, S. Stein, Hybridization

properties of oligodeoxynucleotide pairs bridged by poly-

arginine peptides, Nucleic Acids Res. 24 (1996) 655–661.

[63] E.P. Feener, W.C. Shen, H.J. Ryser, Cleavage of disulfide

bonds in endocytosed macromolecules. A processing not

associated with lysosomes or endosomes, J. Biol. Chem. 265

(1990) 18780–18785.

[64] H.M. Moulton, J.D. Moulton, Peptide-assisted delivery of

steric-blocking antisense oligomers, Curr. Opin. Mol. Ther. 5

(2003) 123–132.

[65] C. Tung, S. Mueller, R. Weissleder, Novel branching

membrane translocational peptide as gene delivery vector,

Bioorg. Med. Chem. 10 (2002) 3609.

[66] M. Lewin, N. Carlesso, C.H. Tung, X.W. Tang, D. Cory, D.T.

Scadden, R. Weissleder, Tat peptide-derivatized magnetic

H. Brooks et al. / Advanced Drug Delivery Reviews 57 (2005) 559–577576

nanoparticles allow in vivo tracking and recovery of

progenitor cells, Nat. Biotechnol. 18 (2000) 410–414.

[67] M. Lundberg, M. Johansson, Positively charged DNA-

binding proteins cause apparent cell membrane trans-

location, Biochem. Biophys. Res. Commun. 291 (2002)

367–371.

[68] M. Lundberg, S. Wikstrom, M. Johansson, Cell surface

adherence and endocytosis of protein transduction domains,

Molec. Ther. 8 (2003) 143–150.

[69] D.C. Anderson, E. Nichols, R. Manger, D. Woodle, M. Barry,

A.R. Fritzberg, Tumor cell retention of antibody Fab

fragments is enhanced by an attached HIV TAT protein-

derived peptide, Biochem. Biophys. Res. Commun. 194

(1993) 876–884.

[70] A. Astriab-Fisher, D.S. Sergueev, M. Fisher, B.R. Shaw, R.L.

Juliano, Antisense inhibition of P-glycoprotein expression

using peptide-oligonucleotide conjugates, Biochem. Pharma-

col. 60 (2000) 83–90.

[71] A. Fittipaldi, A. Ferrari, M. Zoppe, C. Arcangeli, V.

Pellegrini, F. Beltram, M. Giacca, Cell membrane lipid rafts

mediate caveolar endocytosis of HIV-1 tat fusion proteins, J.

Biol. Chem. 278 (2003) 34141–34149.

[72] M. Rusnati, C. Urbinati, A. Caputo, L. Possati, H. Lortat-

Jacob, M. Giacca, D. Ribatti, M. Presta, Pentosan polysulfate

as an inhibitor of extracellular HIV-1 Tat, J. Biol. Chem. 276

(2001) 22420–22425.

[73] M. Rusnati, G. Taraboletti, C. Urbinati, G. Tulipano, R.

Giuliani, M.P. Molinari-Tosatti, B. Sennino, M. Giacca, M.

Tyagi, A. Albini, D. Noonan, R. Giavazzi, M. Presta,

Thrombospondin-1/HIV-1 tat protein interaction: modulation

of the biological activity of extracellular Tat, FASEB J. 14

(2000) 1917–1930.

[74] M. Rusnati, G. Tulipano, C. Urbinati, E. Tanghetti, R.

Giuliani, M. Giacca, M. Ciomei, A. Corallini, M. Presta, The

basic domain in HIV-1 Tat protein as a target for polysulfo-

nated heparin-mimicking extracellular Tat antagonists, J.

Biol. Chem. 273 (1998) 16027–16037.

[75] D.A. Mann, A.D. Frankel, Endocytosis and targeting of

exogenous HIV-1 Tat protein, EMBO J. 10 (1991)

1733–1739.

[76] M. Tyagi, M. Rusnati, M. Presta, M. Giacca, Internalization

of HIV-1 tat requires cell surface heparan sulfate proteogly-

cans, J. Biol. Chem. 276 (2001) 3254–3261.

[77] M. Rusnati, D. Coltrini, P. Oreste, G. Zoppetti, A. Albini, D.

Noonan, F. d’Adda di Fagagna, M. Giacca, M. Presta,

Interaction of HIV-1 Tat protein with heparin. Role of the

backbone structure, sulfation, and size, J. Biol. Chem. 272

(1997) 11313–11320.

[78] H.C. Chang, F. Samaniego, B.C. Nair, L. Buonaguro, B.

Ensoli, HIV-1 Tat protein exits from cells via a leaderless

secretory pathway and binds to extracellular matrix-associ-

ated heparan sulfate proteoglycans through its basic region,

Aids 11 (1997) 1421–1431.

[79] M. Salmivirta, K. Lidholt, U. Lindahl, Heparan sulfate: a

piece of information, FASEB J. 10 (1996) 1270–1279.

[80] T. Jackson, F.M. Ellard, R.A. Ghazaleh, S.M. Brookes, W.E.

Blakemore, A.H. Corteyn, D.I. Stuart, J.W. Newman, A.M.

King, Efficient infection of cells in culture by type O foot-

and-mouth disease virus requires binding to cell surface

heparan sulfate, J. Virol. 70 (1996) 5282–5287.

[81] E.E. Fry, S.M. Lea, T. Jackson, J.W. Newman, F.M. Ellard,

W.E. Blakemore, R. Abu-Ghazaleh, A. Samuel, A.M. King,

D.I. Stuart, The structure and function of a foot-and-mouth

disease virus-oligosaccharide receptor complex, EMBO J. 18

(1999) 543–554.

[82] A. Yayon, M. Klagsbrun, J.D. Esko, P. Leder, D.M. Ornitz,

Cell surface, heparin-like molecules are required for binding

of basic fibroblast growth factor to its high affinity receptor,

Cell 64 (1991) 841–848.

[83] T. Arai, W. Busby Jr., D.R. Clemmons, Binding of insulin-

like growth factor (IGF) I or II to IGF-binding protein-2

enables it to bind to heparin and extracellular matrix,

Endocrinology 137 (1996) 4571–4575.

[84] D. Mohanraj, T. Olson, S. Ramakrishnan, A novel method to

purify recombinant vascular endothelial growth factor

(VEGF121) expressed in yeast, Biochem. Biophys. Res.

Commun. 215 (1995) 750–756.

[85] A. Albini, R. Benelli, M. Presta, M. Rusnati, M. Ziche, A.

Rubartelli, G. Paglialunga, F. Bussolino, D. Noonan, HIV-tat

protein is a heparin-binding angiogenic growth factor,

Oncogene 12 (1996) 289–297.

[86] Y.G. Brickman, M.D. Ford, D.H. Small, P.F. Bartlett, V.

Nurcombe, Heparan sulfates mediate the binding of basic

fibroblast growth factor to a specific receptor on neural

precursor cells, J. Biol. Chem. 270 (1995) 24941–24948.

[87] J.S. Bartlett, R. Wilcher, R.J. Samulski, Infectious entry

pathway of adeno-associated virus and adeno-associated

virus vectors, J. Virol. 74 (2000) 2777–2785.

[88] I.I. Fuki, R.V. Iozzo, K.J. Williams, Perlecan heparan sulfate

proteoglycan. A novel receptor that mediates a distinct

pathway for ligand catabolism, J. Biol. Chem. 275 (2000)

31554.

[89] K.J. Williams, I.V. Fuki, Cell-surface heparan sulfate

proteoglycans: dynamic molecules mediating ligand catabo-

lism, Curr. Opin. Lipidol. 8 (1997) 253–262.

[90] K.S. Rost, J.D. Esko, Microbial adherence to and invasion

through proteoglycans, Infect. Immun. 65 (1997) 1–8.

[91] K. Lidholt, J.L. Weinke, C.S. Kiser, F.N. Lugemwa, K.J.

Bame, S. Cheifetz, J. Massague, U. Lindahl, J.D. Esko, A

single mutation affects both N-acetylglucosaminyltransferase

and glucuronosyltransferase activities in a Chinese hamster

ovary cell mutant defective in heparan sulfate biosynthesis,

Proc. Natl. Acad. Sci. U. S. A. 89 (1992) 2267–2271.

[92] M. Silhol, M. Tyagi, M. Giacca, B. Lebleu, E. Vives,

Different mechanisms for cellular internalization of the HIV-

1 Tat-derived cell penetrating peptide and recombinant

proteins fused to Tat, Eur. J. Biochem. 269 (2002) 494–501.

[93] Y. Liu, M. Jones, C.M. Hingtgen, G. Bu, N. Laribee, R.E.

Tanzi, R.D. Moir, A. Nath, J.J. He, Uptake of HIV-1 Tat

protein mediated by low-density lipoprotein receptor-related

protein disrupts the neuronal metabolic balance of the

receptor ligands, Nat. Med. 6 (2000) 1380–1387.

[94] M. Feligioni, L. Raiteri, R. Pattarini, M. Grilli, S. Bruzzone,

P. Cavazzani, M. Raiteri, A. Pittaluga, The human immuno-

H. Brooks et al. / Advanced Drug Delivery Reviews 57 (2005) 559–577 577

deficiency virus-1 protein Tat and its discrete fragments

evoke selective release of acetylcholine from human and rat

cerebrocortical terminals through species-specific mecha-

nisms, J. Neurosci. 23 (2003) 6810–6818.

[95] B.E. Vogel, S.J. Lee, A. Hildebrand, W. Craig, M.D.

Pierschbacher, F. Wong-Staal, E. Ruoslahti, A novel

integrin specificity exemplified by binding of the alpha

v beta 5 integrin to the basic domain of the HIV

Tat protein and vitronectin, J. Cell Biol. 121 (1993)

461–468.

[96] B.S. Weeks, K. Desai, P.M. Loewenstein, M.E. Klotman, P.E.

Klotman, M. Green, H.K. Kleinman, Identification of a novel

cell attachment domain in the HIV-1 Tat protein and its 90-

kDa cell surface binding protein, J. Biol. Chem. 268 (1993)

5279–5284.

[97] R. Katoh-Semba, A. Oohira, S. Kashiwamata, Nerve growth

factor-induced changes in the structure of sulfated proteo-

glycans in PC12 pheochromocytoma cells, J. Neurochem. 59

(1992) 282–289.

[98] M. Rusnati, G. Tulipano, D. Spillmann, E. Tanghetti, P.

Oreste, G. Zoppetti, M. Giacca, M. Presta, Multiple

interactions of HIV-I Tat protein with size-defined heparin

oligosaccharides, J. Biol. Chem. 274 (1999) 28198–28205.

[99] T.B. Potocky, A.K. Menon, S.H. Gellman, Cytoplasmic and

nuclear delivery of a TAT-derived peptide and a {beta}-

peptide after endocytic uptake into HeLa cells, J. Biol. Chem.

278 (2003) 50188–50194.

[100] R. Fischer, K. Kohler, M. Fotin-Mleczek, R. Brock, A

stepwise dissection of the intracellular fate of cationic

cell-penetrating peptides, J. Biol. Chem. 279 (2004)

12625–12635.

[101] T. Fujimoto, H. Kogo, K. Ishiguro, K. Tauchi, R. Nomura,

Caveolin-2 is targeted to lipid droplets, a new bmembrane

domainQ in the cell, J. Cell Biol. 152 (2001) 1079–1085.

[102] A.D. Frankel, C.O. Pabo, Cellular uptake of the tat protein

from human immunodeficiency virus, Cell 55 (1988)

1189–1193.

[103] G. Kilic, R.B. Doctor, J.G. Fitz, Insulin stimulates membrane

conductance in a liver cell line: evidence for insertion of ion

channels through a phosphoinositide 3-kinase-dependent

mechanism, J. Biol. Chem. 276 (2001) 26762–26768.

[104] E.M. Neuhaus, T. Soldati, A myosin I is involved in

membrane recycling from early endosomes, J. Cell Biol.

150 (2000) 1013–1026.