Nucleocytoplasmic transport of hsp70-related proteins in Xenopus oocytes


Material Information

Nucleocytoplasmic transport of hsp70-related proteins in Xenopus oocytes
Physical Description:
xv, 144 leaves : ill. ; 29 cm.
Mandell, Robert Barry, 1963-
Publication Date:


Subjects / Keywords:
Research   ( mesh )
Oocytes   ( mesh )
Heat-Shock Proteins 70   ( mesh )
Xenopus   ( mesh )
Biological Transport   ( mesh )
Cytoplasm   ( mesh )
Cell Nucleus   ( mesh )
Department of Anatomy and Cell Biology thesis Ph.D   ( mesh )
Dissertations, Academic -- College of Medicine -- Department of Anatomy and Cell Biology -- UF   ( mesh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 1991.
Includes bibliographical references (leaves 124-143).
Statement of Responsibility:
by Robert Barry Mandell.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001665316
oclc - 25668249
notis - AHX7098
System ID:

This item is only available as the following downloads:

Full Text









I would like to thank Dr. Carl M. Feldherr for allowing

me to work in his laboratory throughout the past three

years. The investigations described in this dissertation

could not have been accomplished without his support and

guidance. He has performed and continues to perform

scientific research of the highest quality, and I am

grateful to have had the opportunity to learn from him.

I would like to extend special thanks to Dr. Robert

Cohen for his time and effort, including numerous

discussions, for providing critical reviews of the

manuscripts submitted for publication, and for serving as a

member of my supervisory committee. I would also like to

thank Dr. Thomas W. O'Brien for serving as a member of my

committee and for reviewing manuscripts submitted for

publication, and Dr. Robin Wallace for serving as a member

of my supervisory committee.

I would like to thank Debra Akin for her assistance

with the electron microscopy, for proofreading manuscripts

and foremost for making the time spent in lab more



I would like to thank Dr. Gillian Small for allowing me

to perform molecular biological experiments in her

laboratory, and for providing critical reviews of

manuscripts submitted for publication.

I would also like to thank Dr. William A. Dunn for

numerous and informative discussions regarding my data, and

his interest in my research.

I am grateful to Dr. Nancy Denslow and the Protein

Chemistry Core Facility for allowing me to perform the

densitometric analyses with their Image Analysis System.

I would like to acknowledge and thank Kari Eissinger

for her help in photographing and printing the figures

contained herein.

Much of the work presented in Chapters II and III is

reproduced from the Journal of Cell Biology, 1990, volume

111, number 5, part 1, pages 1775-1783 by copyright

permission of The Rockefeller University Press.

I am grateful to Dr. Gregory Flynn from Dr. J.

Rothman's lab at Princeton University for providing the

amino terminus-specific anti-hsc70 antibodies and the

synthetic peptides used to demonstrate that a-chymotrypsin

cleaves hsc70 from its carboxyl-terminus.

Finally, and most importantly, I thank my wife, Tamara,

for her love, patience, moral support and encouragement

through the good and bad times, and for the many sacrifices

she made while I completed my graduate education.







ABSTRACT. ... .. . .. xiii



Introduction. . . 1
Nuclear Permeability. . 3
Nuclear Efflux. . 27
Recycling Proteins. . 34
Statement of Research Topic 39


Introduction. . . 41
Materials and Methods . 42
Results . . 48
Discussion. . . 57


Introduction. . .. 61
Materials and Methods . 63
Results . . 68
Discussion. . .. 86

OF RAT HSC70. . .. 90

Introduction. . . 90
Materials and Methods . 92
Results . . 97
Discussion. . . 107


Summary of Results. . 116
Future Studies. . 120





1. Purification of B3 and B4. . .

2. Purification of Rat Hsc70. . .

3. Cyanogen Bromide Cleavage Maps of
Rat Hsc70 and B3 and B4. . .

4. Relative ATP Hydrolysis. . .

5. Nuclear Uptake Kinetics of 125I-labeled
B3/B4 (A) and 125I-labeled Rat Hsc70. .

6. Autoradiographs of One-Dimensional Gels
of Nuclear and Cytoplasmic Fractions
Following Cytoplasmic Injection of
125I-B3/B4 (A) and 125I-hsc70 (B) .


. 51

. 53

. 56

. 59

7. Nuclear Efflux Kinetics of 12sI-labeled
BSA (A), B3/B4 (B) and Hsc70 (C) 76

8. Autoradiographs of One-dimensional Gels of
Nuclear and Cytoplasmic Fractions 2 Hours
After Nuclear Injections of 125I-B3/B4 from
two Separate Experiments (A and B) 80

9. The Intracellular Distribution of
B3/B4-coated Gold, 30 Minutes After
a Cytoplasmic Injection. . ... 83

10. Characterization of a-Chymotrypsin Digestion
Products of Rat Brain Hsc70. ... 100

11. Amino Acid Sequence of Rat Hsc70 101

12. Autoradiographs of a One-Dimensional Gel
of Nuclear and Cytoplasmic Fractions
Following Cytoplasmic Injection of
125I-labeled, Chymotrypsin-digested Hsc70 104

13. Nuclear Uptake Kinetics of 125I-labeled,
Chymotrypsin-generated Polypeptides
and HSC70. . . .. 106


14. Autoradiograph of a One-dimensional Gel of
Nuclear and Cytoplasmic Fractions 2 Hours
After Nuclear Injections of 125I-labeled,
Chymotrypsin-digested Hsc70. . 109

15. Relative Nuclear Efflux Kinetics of
125I-labeled, Chymotrypsin-generated
Polypeptides and Hsc70 2 Hours After
Nuclear Injection. . 111



Table Page





Ala Alanine

a Alpha

A Angstrom

Asn Asparagine

Asp Aspartic Acid

ATP Adenosine-5'-triphosphate

ATPase Adenosine-5'-triphosphatase

Arg Arginine

B Beta

BSA Bovine Serum Albumin

cAMP Cyclic Adenosine Monophosphate

CAPS 3-[cyclohexamino]-l-propane-
sulfonic acid

CNBr Cyanogen Bromide

cpm Counts per minute

cDNA Complementary Deoxyribonucleic

DTT Dithiothreitol

EM Electron Microscopic

ER Endoplasmic Reticulum

Glu Glutamic Acid

GTP Guanosine Triphosphate

Y Gamma

3H Tritiated

HEPES N-2-Hydroxyethylpiperazine-N'
2-ethane sulfonic acid
1251 Iodinated

IgG Immunoglobulin

kD Kilodaltons

Lys Lysine

M Molar Concentration

mCi Millicurie(s)

mg Milligram(s)

MDCK Madine Darby Canine Kidney

ml Milliliter(s)

mM Millimolar

p Micron(s)

pCi Microcurie(s)

Pg Microgram(s)

Pm Micrometer(s)

pM Micromolar

NEM N-ethylmaleimide

NH2 Amino

nl Nanoliter(s)

NLS Nuclear Localization Signal

nm Nanometer(s)

nmol Nanomole(s)

NTPase Nucleoside Triphosphatase

P Probability

32P Radioactive Phosphorus


Phe Phenylalanine

Pro Proline

PAGE Polyacrylamide Gel

pI Isoelectric Point

Poly(A) Polyadenylic Acid

Poly(I) Polyinosinic Acid

PVP Polyvinylpyrrolidone

RNA Ribonucleic Acid

hNRNA Heteronuclear RNA

mRNA Messenger RNA

rRNA Ribosomal RNA

snRNA Small Nuclear RNA

snRNP Small Nuclear

tRNA Transfer RNA

RNP Ribonucleoprotein

35S Radioactive Sulfur

SDS Sodium Dodecyl Sulfate

Sm-Antigen Smith's Antigen

SV40 Simian Virus 40

T-Antigen Tumor Antigen

TCA Trichloroacetic Acid

UV Ultraviolet Radiation

Val Valine

WGA Wheat Germ Agglutinin


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Robert Barry Mandell

August 1991

Chairman: Dr. Carl M. Feldherr
Major Department: Anatomy and Cell Biology

Two 70kD polypeptides, B3 and B4, are present in

equivalent concentrations in the nucleus and cytoplasm of

Xenopus oocytes. The initial objectives of this study were

to determine if they 1) are members of the 70kD family of

heat shock proteins, and 2) recycle between the nuclear and

cytoplasmic compartments. Evidence based on high affinity

binding to ATP, cross reactivity of B3/B4-specific

antibodies with rat hsc70, and the similarity of cyanogen

bromide peptide maps with hsc70, verified that B3 and B4 are

hsp70-related. Nuclear uptake and efflux studies were

performed by microinjecting 125I-labeled B3/B4, rat hsc70 and

BSA into oocyte cytoplasms and nuclei, respectively, and

examining their subsequent intracellular distributions.

These polypeptides both accumulated in and left the nucleus


at rates faster than could be accounted for by diffusion;

additionally, B3/B4-coated gold particles as large as 120A

in diameter were able to enter the nucleus by passing

through the pores. Cell fusion experiments, in which

labeled, anucleate oocyte vegetal hemispheres were fused,

under oil, with nucleate unlabeled animal hemispheres,

demonstrated that cytoplasmic B3 and B4 could enter the

nucleus after equilibration was reached, arguing against the

existence of separate nuclear and cytoplasmic populations.

Collectively, these results show that B3, B4 and rat hsc70

are transported across the nuclear envelope and recycle

between the nucleus and cytoplasm.

Having demonstrated that hsc70 was bidirectionally

transported, our next objective was to identify domains

involved in this process. Limited proteolytic digestion

with a-chymotrypsin generated three major truncated proteins

of -67.5, 59.5 and 56.5kD. Reactivity with amino terminal-

specific antibodies show that carboxyl-terminal fragments

were removed. Nuclear uptake and efflux studies were

performed by microinjecting 125I-labeled proteins into oocyte

cytoplasms or nuclei, respectively, and determining their

subsequent nucleocytoplasmic distribution. The uptake

rates, while faster than BSA controls, were inversely

related to the size of the truncated proteins and reduced

compared to undigested hsc70. Digestion similarly reduced

the relative efflux rates of the truncated proteins. These


results indicate that the carboxyl-terminal domain of hsc70

is involved in its bidirectional exchange.



The cell nucleus likely evolved in response to greater

organizational and regulatory problems associated with the

increasing genome size of more complex organisms.

Compartmentalization of the chromatin presumably makes it

easier for cells to handle these problems. The nucleus is

separated from the cytoplasm by the nuclear envelope, which

consists of inner and outer nuclear membranes, pore

complexes and a proteinaceous lamina. The double nuclear

membrane, separated by a 100-600A perinuclear space, serves

as a physical barrier between the nucleus and cytoplasm.

The perinuclear space is continuous with the lumen of the

endoplasmic reticulum. The nucleoplasmic side of the inner

membrane is associated with a proteinaceous nuclear lamina

that consists primarily of 50-80kD intermediate filament-

related proteins, the lamins (Fisher et al., 1986; McKeon et

al., 1986; Franke, 1987). The lamina serves as a structural

scaffold for the envelope, and might be involved in its

disassembly and reformation during cell division (for

review, see Gerace and Burke, 1988). The inner and outer

membranes fuse at various intervals, forming the boundaries

of the nuclear pore complexes. The nuclear pore complex is

the site of macromolecular exchanges between the nucleus and

cytoplasm (Anderson and Beams, 1956; Feldherr, 1962, 1965;

Stevens and Swift, 1966; Feldherr et al., 1984; Dworetzky

and Feldherr, 1988). The nuclear pore complex could

therefore have a direct role in regulating nucleocytoplasmic

exchanges. Nuclear pore structure appears similar in most

eukaryotes (Franke, 1974; Maul, 1977), but their density

varies from 2-60/Pm2 of nuclear envelope (Franke and Sheer,

1974). Extensive electron microscopic analysis revealed

that each pore complex, ~120nm in diameter, has three

prominent features: 1) two coaxial rings located at the

inner and outer surfaces of the nuclear envelope, 2) a

radially aligned central spoke assembly and 3), a central

granule or plug (Unwin and Milligan, 1982). It has been

postulated that the central plug represents endogenous

substrates fixed in transit during preparation. Recent

investigations of pore structure using cryoelectron

microscopy and quantitative image processing have revealed

considerably more structural detail. The most significant

finding is a central ~390A diameter transporter (Akey,

1989). Probes specific for nuclear pore proteins as well as

a tagged nuclear protein labeled this assembly, which is

consistent with a functional role for the transporter in

facilitated protein uptake (Akey and Goldfarb, 1989). Akey

(1990) proposed a double-iris model for the transporter to


explain several configurations observed during

macromolecular translocation. Two irises, one each at the

nucleoplasmic and cytoplasmic faces of the transporter, open

and close sequentially to allow uptake of transported

karyophilic molecules, or efflux of RNP. When both are

closed, the transit pore diameter is ~90A, consistent with

the 90A channel available for diffusion calculated by Paine

et al. (1975).

Nuclear Permeability

Both morphological and biochemical analyses have

revealed that the composition of the cell nucleus is

significantly different from the cytoplasm. However,

macromolecules, including proteins and RNAs, are

continuously moving between these compartments. Because the

nuclear envelope separates the nucleus and cytoplasm, it

represents a potential regulatory site. The accumulation of

nuclear constituents and regulation of exchanges could be

determined by the permeability properties of the nuclear

envelope, intranuclear binding or a combination of both


Ions and Small Molecules

The permeability properties of the nuclear envelope to

ions and small molecules has been studied using quantitative

autoradiographic techniques (Abelson and Duryee, 1949;


Horowitz, 1972; Horowitz and Moore, 1974), by direct

quantitation of labeled ions and small molecules in fast-

frozen and dissected nuclear and cytoplasmic fractions

(Naora et al. 1962; Century et al., 1970; Horowitz and

Fenichel, 1970), with electrophysiological techniques (Kanno

and Loewenstein, 1963), and by comparing the equilibrium

concentration of various ions and sucrose between a

cytoplasmic gelatin reference phase, the nucleus, and the

cytoplasm of oocytes (Horowitz and Paine, 1976). The uptake

rates of 2Na, "K, 3PO, 35SO, 1C-leucine, 4C-alanine, 3H-

sucrose and 3H-inulin were consistent with the idea that

they diffused through a 90A (diameter) aqueous channel.

Several findings appear to contradict this

interpretation. First, Loewenstein and Kanno (1963)

reported a significant electrical resistance across

Drosophila melanogaster salivary gland cell nuclear

envelopes, which suggested these membranes were not freely

permeable to ions. Second, the resistance measured across

nuclear envelopes of mouse embryo cells by patch-clamping

suggested that the pores might not be freely permeable to K+

and indicated the presence of a potassium channel (Mazzanti

et al., 1990). An explanation that might account for these

results was forwarded by Paine (1975), who suggested that a

significant increase in resistance across the nuclear

envelope, and thus an apparent barrier to ions, could occur

if half of the nuclear pores were blocked by


translocating macromolecules such as RNP. These findings

might therefore simply reflect obstructed diffusion


This possibility does not explain the results of Neylon

et al. (1990) who found that a thrombin-induced wave of free

calcium progressing rapidly through cultured smooth muscle

cells did not affect the free calcium concentration of the

nucleus. This result, while suggesting that the nuclear

envelope was not freely permeable to calcium ions, remains


Exogenous Macromolecules

Characterization of the physical properties of the

nuclear envelope with respect to the diffusion of

macromolecules has been facilitated by studies that measured

the intracellular distributions of cytoplasmically

microinjected tracers including colloidal gold particles,

ferritin, and fluorescent and radiolabeled dextrans and


Feldherr (1962, 1965) studied the intracellular

distribution of PVP-coated colloidal gold particles

microinjected into the cytoplasm of the amoebas Amoeba

proteus and Chaos chaos by electron microscopy. Gold

particles were shown to enter the nucleus through the

centers of the nuclear pores. Particles up to 85A in

diameter readily entered the nuclei, the uptake of those


from 89-106A was reduced, and particles of 110-170A were

almost entirely excluded, indicating 1) the envelope had an

exclusion limit of approximately 140A (diameter) and 2) the

rate of translocation was inversely proportional to the size

of the tracer.

Using quantitative fluorescence microscopy, Paine and

Feldherr (1972) measured the nuclear translocation of well-

characterized exogenous proteins from 17.5kD to 495kD

microinjected into the cytoplasm of Periplaneta americana

oocytes, and found that nuclear migration was limited by

their molecular size. Similar findings were obtained for

exogenous proteins and dextrans using amphibian oocytes

(Bonner, 1975a) and mammalian cultured cells (Stacey and

Allfrey, 1984). However, Stacey and Allfrey found that

hemoglobin (68kD) and myosin (468kD) both entered culture

cell nuclei immediately after microinjection. While the

rapid translocation of hemoglobin might be explained by its

dissociation into subunits, the results with myosin are

still unresolved.

The relationship of permeation rate to particle size

was quantitated by Paine et al. (1975), who performed

careful kinetic studies of the passage rates of

microinjected 3H-dextrans of known hydrodynamic radii across

the nuclear envelope of Rana pipiens oocytes. Intracellular

dextran distributions were determined by autoradiography of

sections from fast-frozen cells. Dextrans diffused through


the nuclear pores at rates inversely proportional to their

size, consistent with previous findings for exogenous

proteins. Using a mathematical model which considered

uptake rates, diffusional coefficients and the length of the

diffusion channel, they calculated the diameter of the

channel available for diffusion through the nuclear pores to

be ~90A. In a related study, Peters et al. (1984) measured

the nucleocytoplasmic exchange of dextrans in rat

hepatocytes by fluorescence microphotolysis and calculated

the diameter of the diffusion channel to be ~110A.

Collectively, experiments using exogenous tracers

indicated that 1) the nucleocytoplasmic exchange of

macromolecules occurs by diffusion through the nuclear

pores, 2) permeation rate is inversely proportional to

molecular size and 3) the size of the channels available for

diffusion are ~90, 110 and 140A for amphibian oocytes,

hepatocytes and amoebae, respectively.

Endogenous Macromolecules

The nucleocytoplasmic exchange of exogenous tracers

that were not specifically targeted to the nucleus allowed

for the characterization of the diffusion permeability

properties of the nuclear envelope. The use of endogenous

proteins allows for the elucidation of interactions and

mechanisms involved in karyophilic protein targeting.


Gurdon (1970) and Bonner (1975a) found that the Xenopus

oocyte nucleus rapidly accumulated histones to at least 100

times the concentration found in the cytoplasm, which

indicated that nuclei were capable of accumulating selected

proteins. Bonner (1975b) found that metabolically labeled

endogenous cytoplasmic and nuclear proteins microinjected

into Xenopus oocyte cytoplasms reconcentrated in the

cytoplasm and nucleus, respectively. The nuclear proteins

thus retained the information necessary for nuclear

localization. A significant amount of some labelled

cytoplasmic proteins, however, entered the nucleus. Bonner

therefore proposed the existence of three protein classes:

1) nuclear-localized N-proteins, 2) cytoplasmically-

localized C-proteins and 3) B proteins, found in both

compartments. A more detailed examination of this

distribution was performed by DeRobertis et al. (1978) who

used high resolution 2D-gel electrophoresis to define the

intracellular distribution of Xenopus laevis oocyte

proteins. 35S-labeled nuclear proteins microinjected into

oocyte cytoplasms rapidly re-entered the nucleus and were

greatly concentrated there. Two proteins in particular, N1

and N2, accumulated to ~120 fold higher concentration in the

nucleus, much faster than would be expected by diffusion

(Paine et al., 1975), which indicated that factors other

than size governed the accumulation of large nuclear

proteins. Consistent with these results, Feldherr (1975)


found that newly synthesized Xenopus oocyte nuclear proteins

from 94-150kD were three times more concentrated in the

nucleus than the cytoplasm after 6 hours, and their

accumulation rates were greater than would have been

predicted for exogenous tracers of similar size.

Thus, specific endogenous proteins accumulated in the

nucleus at significantly greater rates and extents than

exogenous molecules of comparable size. One mechanism

proposed to account for this was diffusion across the

nuclear envelope followed by selective binding and retention

in the nucleoplasm (Bonner, 1978). To investigate the role

of the nuclear envelope in regulating nucleocytoplasmic

exchanges of endogenous proteins and directly address the

question of selective intranuclear binding, Feldherr and

Pomerantz (1978) altered the permeability properties of

Xenopus oocyte nuclear envelopes by puncturing with fine

glass needles. This resulted in a substantial increase in

the nuclear uptake of 125I-labeled BSA, indicating that the

envelopes ability to act as a barrier to the diffusion of

large macromolecules was compromised. However, puncturing

had only a minimal affect on the nuclear uptake of

endogenous proteins. High-resolution 2D-gel analysis

(Feldherr and Ogburn, 1980), fluorography and direct

quantitation showed that of ~300 nuclear proteins

identified, the distribution of only 10-15 varied between

punctured and control nuclei, suggesting that 1) the


accumulation of nuclear proteins was dependent in part upon

intranuclear binding, and 2) because the uptake rates of

endogenous proteins were not effected, passage across the

nuclear envelope was not a rate-limiting step in

accumulation. Additionally, because a 90A diffusion channel

would allow significant exchange of small macromolecules,

Paine (1982) reasoned that small nuclear proteins were

absent from the cytoplasm because of selective nuclear

binding or incorporation into nuclear structural elements.

Facilitated Uptake

Endogenous oocyte nuclear proteins accumulated in the

nucleus at rates significantly faster that exogenous

macromolecules of similar size (Feldherr, 1975; DeRobertis

et al., 1978). Feldherr et al. (1983) provided evidence

suggesting the nuclear uptake and accumulation of endogenous

proteins larger than the exclusion limit for diffusion

involved a facilitated transport mechanism. They compared

the actual nuclear uptake rate of a 148kD Rana oocyte

protein, RN1, with the rate predicted by a) simple diffusion

and b) diffusion and selective binding in the nucleus. The

diffusion rates through 90A diameter pores were calculated

for molecular configurations of RN1 with various axial

ratios. Even in its most diffusive conformation and

assuming a high nuclear binding coefficient, simple

diffusion could only account for -1/20 the observed uptake


of RN1, which strongly suggested that the accumulation of

RN1 occurred by a facilitated transport process.

Clues to the mechanism by which endogenous proteins are

selected for facilitated nuclear uptake were first obtained

by Dingwall et al. (1982) who studied the nuclear

accumulation of nucleoplasmin, a major nuclear protein of

Xenopus oocytes (Mills et al., 1980). Nucleoplasmin is a

llOkD pentameric protein. Purified, radiolabeled

nucleoplasmin, microinjected into the oocyte cytoplasm,

accumulated to a N/C concentration ratio of ~475 after 24

hours. Proteolytic digestion with pepsin generated two

fragments, a core consisting of an amino-terminal trypsin-

resistant pentamer with an apparent monomer molecular weight

of 23kD, and tails consisting of carboxy-terminal domains of

~10kD. Limited trypsin digestion produced cores with 0-5

tails. Cores with no tails were excluded from the cell

nucleus, while cores with tails accumulated in the nucleus

at rates proportional to the number of tails (5 being

fastest, 1 slowest). Cores microinjected directly into

oocyte nuclei were retained there. Thus the nucleoplasmin

molecule could be broken down into two domains, 1) the core

and 2) the tail, which contains information necessary for

passage across the nuclear envelope. These results

suggested that a signal for the facilitated nuclear uptake

of nucleoplasmin is encoded somewhere in the molecular

structure of its carboxyl-terminal tail domain, and argued


against free diffusion and selective retention as a

mechanism for the accumulation of nucleoplasmin.

To determine the site of exchange of endogenous nuclear

proteins, Feldherr et al. (1984) microinjected gold

particles coated with the endogenous nuclear protein

nucleoplasmin into the cytoplasm of Xenopus oocytes, and

examined their translocation across the nuclear envelope

using electron microscopy. Gold particles acquire the

properties of the coating agent, so this method permitted

direct visualization of the pathway taken by endogenous

nuclear proteins. Particles from 50-200A in diameter were

clearly seen in transit passing through the central channels

of the nuclear pores, indicating that these were the sites

of nucleocytoplasmic exchange. The size of the channel

available for transport was therefore significantly larger

than the 90A channel available for diffusion. Additionally,

these findings ruled out the possibility that shape changes,

which could reduced the radius of large nuclear proteins to

the point where migration rates were accountable by simple

diffusion, occurred before they passed through the nuclear


Using an in vitro nuclear import assay composed of

isolated rat liver nuclei in a Xenopus egg extract, Newmeyer

et al. (1986a,b) showed that the uptake of nucleoplasmin was

both temperature- and ATP-dependent. Newmeyer and Forbes

(1988) and Richardson et al. (1988) further demonstrated


that import could be separated into two distinct steps,

binding at the nuclear envelope, followed by translocation

through the nuclear pores. Metabolic inhibitors,

temperature reduction and a lectin that bound to the nuclear

pores (WGA) inhibited the translocation, but not binding of

nucleoplasmin (identified by direct immunofluorescence) and

nucleoplasmin-coated gold (identified by EM) to the surface

of nuclear envelopes in cultured cells and Xenopus oocytes,


Nuclear Localization Signals (NLS)

Molecular biolological approaches have proven useful in

defining the specific sequences contained in the carboxyl-

tail domains of nucleoplasmin and regions of other nuclear-

localized polypeptides required for facilitated nuclear

transport. The first such experiments were reported by Hall

et al. (1984), who examined the intracellular distribution

of B-galactosidase fused by recombinant DNA methodology to

selected sequences of the yeast nuclear protein MATa2. A 13

amino acid sequence from MATa2 was sufficient to target

normally cytoplasmic 3-galactosidase to the yeast nucleus,

but was unable to target carrier proteins into the nucleus

of mammalian cells (Chelsky et al., 1989; Lanford et al.,

1990). This finding suggested that receptors for this

signal could be yeast or generally species specific.


Lanford and Butel (1984) identified a SV40-adenovirus 7

hybrid virus that encoded a cytoplasmically-localized mutant

form of the normally nuclear-localized large tumor antigen

(~94kD). This mutation resulted from the replacement of a

lysine with asparagine at amino acid 128, which indicated

that the alteration of a single amino acid was sufficient to

eliminate nuclear import. Deletion analyses of this region

performed by Kalderon et al. (1984) revealed a minimal

sequence of seven amino acids, Pro-Lys- Lys-Lys-Arg-Lys-

Val, was sufficient to facilitate nuclear accumulation of B-

galactosidase and pyruvate kinase. Further evidence that

the seven amino acid SV40 T-antigen NLS was a transport

signal was provided by Lanford et al. (1986) who

demonstrated that a peptide fragment containing the SV40 NLS

biochemically conjugated to proteins too large to passively

diffuse into the nucleus (e.g., BSA, IgG and ferritin)

facilitated their nuclear uptake. These SV40 NLS-conjugated

proteins did not move into the cytoplasm when injected

directly into the nucleus, which indicated that the NLS

functioned vectorially. However, these results did not

eliminate the possibility that the NLS was bound in the

nucleus and thus rendered unable to facilitate efflux.

More than 32 different NLSs from many different species

have been identified including Xenopus N1/N2 and

nucleoplasmin (Dingwall et al., 1988 and Kleinschmidt and

Seiter, 1988, respectively), human lamin A (Loewinger and


McKeon, 1988) and viral proteins including the SV40 large T-

antigen and adenovirus Ela (Lyons et al., 1987) (for review,

see Garcia-Bustos et al., 1990). While there is no

consensus sequence, NLS are characterized by short, basic

amino acid sequences (usually 8-10 residues) containing a

high proportion of lysine and arginine, and are generally

not restricted to specific peptide domains (Garcia-Bustos et

al., 1990). The SV40 NLS is the prototype to which most

others are compared, probably because it was the first shown

to function autonomously in mammalian (Kalderon et al.,

1984; Lanford et al., 1986) and amphibian cells (Goldfarb et

al., 1986; Dworetzky et al., 1988).

Multiple NLS increase the efficiency of nuclear protein

uptake. Dingwall et al. (1982) showed that the rate of

nucleoplasmin accumulation was directly proportional to the

number of NLS-containing tail domains. Lanford et al.

(1988, 1990) observed partial nuclear uptake with the mutant

SV40 NLS if a large number of these were conjugated to a

carrier protein, suggesting that NLS efficiency is additive.

This was consistent with results of Roberts et al. (1987)

who found that the extent of nuclear localization of

pyruvate kinase constructs containing partially defective

SV40 NLS was proportional to the number of signals. Many

proteins, including the polyoma large T-antigen (Richardson

et al., 1986), adenovirus DNA-binding protein (Morin et al.,

1989) and yeast ribosomal protein L29 (Underwood and Fried,


1990) have been shown to contain more than one NLS.

Multiple signals should increase the avidity of a

polypeptide for NLS binding components compared to a similar

protein with only one signal; thus, signal multiplicity

could represent a means of regulating nuclear uptake rates.

To directly investigate the effects of NLSs on the

properties of the nuclear pore transport channels, Dworetzky

et al. (1988) examined the effects of variations in the

number and sequence of targeting signals on nuclear

transport. The intracellular distribution of colloidal gold

particles coated with BSA preparations conjugated with

synthetic peptides containing the SV40 NLS microinjected

into the cytoplasm of Xenopus oocytes were determined by

electron microscopy. The functional size of the transport

channels and relative nuclear uptake were found to be

directly proportional to the number of active NLS.

Consistent with Feldherr et al. (1984), gold particles up to

260A in diameter with active signals were able to enter the

nucleus. Thus, the dimensions of the transport channel

increased in proportion to the number of NLSs to allow the

import of transported macromolecules. Additionally, gold

particles coated with nucleoplasmin accumulated faster than

those coated with any BSA-peptide conjugates, suggesting

that endogenous NLSs function more effectively.

Although minimal nuclear localization sequences

function autonomously, studies have suggested that domains


flanking targeting sequences might affect nuclear import.

Rihs and Peters (1989) showed that B-galactosidase hybrid

proteins containing residues 111-135 of SV40 large T-antigen

accumulated in the nucleus significantly faster than hybrids

containing only the minimal SV40 signal (residues 127-132;

minutes vs. hours, respectively). The additional sequence

does not resemble any NLS, suggesting it does not function

as one. Region 111-126 contains four potential

phosphorylation sites, but whether these are phosphorylated,

and the functional significance of phosphorylation with

respect to nuclear transport in general, remains unknown.

However, recent investigations have shown that a potential

casein kinase II site at serine 111/112 is the determining

factor that enhances transport (Rihs et al., 1991).

Recent findings have provided evidence for the

existence of additional classes of signals. Robbins et al.

(1991) defined a class of bipartite NLSs composed of two

interdependent basic domains separated by ~10 amino acids in

the Xenopus oocyte protein nucleoplasmin that is conserved

in a number of nuclear proteins including N1/N2, the

nucleolar protein N038 and the human glucocorticoid

receptor. Additionally, evidence for the existence of

different functional classes of NLSs was provided by Hall et

al. (1990). They showed that the yeast MATa2 nuclear

protein contains two distinct NLSs with no sequence

homology; deletion of either differentially affected nuclear


localization, suggesting that each signal mediates different

steps in the nuclear import process. However, the fact that

the SV40 NLS can function autonomously argues against this

idea. While multiple NLS can increase the rate of nuclear

uptake, the precise role of non-homologous NLSs remains


Studies of the nuclear uptake of snRNAs supports the

existence of a more complex nuclear import signal. SnRNAs

are a family nuclear RNA species characterized by a

trimethylguanosine cap and their association with specific

RNA binding proteins (Steitz et al., 1988b; Zieve et al.,

1988). SnRNAs have been shown to accumulate in the nucleus

after transient passage through the cytoplasm (Zieve et al.

1988). Nuclear import requires binding of particular snRNP

proteins called Sm antigens. Sm proteins are excluded from

the nucleus when not bound to SnRNA (DeRobertis et al.,

1982; Zeller et al., 1983; Mattaj and DeRobertis, 1985), and

therefore either lack functional NLSs, are bound in the

cytoplasm, or have NLSs that are exposed by interaction with

snRNA. SnRNA is excluded from the nucleus when not bound to

the Sm antigen (Mattaj and DeRobertis, 1985), indicating

that the NLS is likely contained in the protein. However,

Fischer and LUhrmann (1990) demonstrated that nuclear

transport of Ul snRNP was dependent upon the presence of the

trimethylguanosine cap, and Hamm et al. (1990) demonstrated

that both Sm binding and the cap are required for uptake of


UlsnRNA in Xenopus oocytes. Therefore, the snRNP uptake

signal could be more complex than the NLS sufficient for

karyophilic protein import.

Although the SV40 prototype NLS functions autonomously

in most systems, the existence of different NLS sequences,

differences in the effectiveness of particular NLSs and the

presence of additional sequences (not homologous to the SV40

NLS) that will initiate nuclear transport suggests there are

different functional classes of NLS.

Facilitated Nuclear Transport is Receptor-mediated

Several lines of independent evidence support the

hypothesis that nuclear protein import is a receptor-

mediated process. The ability of short amino acid sequences

to direct the nuclear localization of non-nuclear proteins

suggests that these sequences are recognized by cellular

proteins involved in nuclear import, analogous to receptor-

ligand interactions (Lanford et al., 1988). Feldherr et al.

(1984) observed that nucleoplasmin-coated gold particles

appeared to accumulate at the cytoplasmic surface of the

nuclear pores during translocation. Additionally, both

Newmeyer and Forbes (1988) and Richardson et al. (1988) also

found that karyophilic tracers were bound at the cytoplasmic

surface of the nuclear pores before translocation occurred.

These findings are consistent with the presence of receptors

at or near the cytoplasmic face of the nuclear pores.


Two important characteristics of a receptor-mediated

process are specificity and saturability. Goldfarb et al.

(1986) demonstrated that excess free NLS peptide reduced the

nuclear uptake rate of SV40 NLS-BSA conjugates microinjected

into Xenopus oocytes, which suggested that free NLS competed

with NLS-conjugated proteins for an import receptor, and

protein uptake was indeed saturable. An inactive form of

the SV40 NLS (121lys to asn) did not affect the uptake of the

conjugates, which demonstrated the specificity of an

unidentified NLS receptor.

NLS Binding Proteins

Available evidence suggests that the NLS initiates the

transport of nuclear proteins through the pores by

interacting with specific pore-localized receptors. A

common strategy in the search for nuclear import receptors

has been to generate NLS-peptides that contain a

photoactivatable cross-linking group, and use these to

identify NLS-binding proteins in cell extracts or on Western

blots. While a number of different NLS-binding proteins

have been identified in a number of species, there is no

conclusive evidence that they have a functional role in

facilitated transport.

NLS-binding proteins have been identified in the rat

liver cell nucleus (Adam et al., 1989; Yamasaki et al.,

1989), cytoplasm (Yamasaki et al., 1989) and specifically


associated with the nuclear envelope (Adam et al., 1989;

Benditt et al., 1989; Imamoto-Sonobe et al., 1990) that

range in size from 55-140kD. Nuclear-localized NLS binding

proteins have also been found in yeast (Silver et al., 1989;

Lee and Melese, 1989). Additionally, Li and Thomas (1989)

identified a 66kD NLS-binding protein in HeLa cell nuclear

extracts using an import assay based on lysolecithin

permeabilization of living cells and a 125I-labelled, UV-

activated transferrable cross-linker conjugated to SV40 NLS

peptides. Because their conditions more closely resemble

those in vivo, these results support the hypothesis that

NLS-binding proteins are actually involved in the import

process. In general, when tested, competition experiments

have shown that NLS-binding proteins bound heterologous

signals, although not always with the same affinity. This

implies that diverse signals might use a common transport

route. Binding was a) signal sequence-dependent, b)

competed by excess NLS (free peptide or conjugated), c) ATP-

independent and d) not affected by WGA, consistent with the

findings of Newmeyer and Forbes (1988) that WGA inhibits

transport but not the binding of karyophilic proteins.

Whether the NLS-binding proteins identified by

different groups are related remains to be determined. It

is clear, however, that NLS-binding proteins are not only

localized to the nuclear pores. One possible explanation

for these results is that NLS-mediated uptake involves a


pore-localized receptor specific for a nuclear protein/NLS-

binding protein complex. Karyophilic proteins might be

bound by NLS-binding proteins in the cytoplasm, and this

complex would then be recognized by specific receptors

localized at the pore transporter assembly. Once bound at

the pore, the complex could dissociate, allowing for the

translocation of the karyophilic protein. Alternatively,

the NLS-binding protein could function as a recycling

carrier and the entire complex could be translocated. Once

in the nucleus, the NLS-binding protein would dissociate and

be transported back to the cytoplasm. The latter is

consistent with the localization of the NLS-binding protein

to both cellular compartments. However, one cannot rule out

the possibility that the apparent distribution of these

proteins is an artifact of cell fractionation.

NLS-binding proteins could represent a means for cells

to regulate nuclear uptake. These proteins might serve as

"adapters" which recognize diverse NLS for presentation to

pore receptors. Consistent with the existence of multiple

NLS-binding proteins and the findings that NLSs are not all

equally effective, pore receptors might have different

affinities for karyophilic protein/adaptor complexes which

could determine uptake rates.


Cytoplasmic Components are Required for Nuclear Uptake

Cytoplasmic components have been shown to be necessary

for nuclear transport. Newmeyer and Forbes (1990)

identified an NEM-sensitive cytosolic factor, NIF-1, which

was necessary for both nuclear protein binding and transport

in an in vitro import assay. Binding and transport in NEM-

treated cytosol were rescued by the addition of NIF-1. It

has not been determined whether NIF-1 is a NLS-binding

protein. Adam et al. (1990) also found that cytosolic

components were required for nuclear import in an in vitro

nuclear transport system that employed digitonin-

permeabilized HeLa and normal rat kidney cells. Transport

required the addition of exogenous cytosol, suggesting that

digitonin treatment resulted in the loss of soluble

cytosolic components) necessary for nuclear import.

Consistent with Newmeyer and Forbes (1990), NEM treatment

inhibited transport. However, neither of these

investigations demonstrated that the NEM-sensitive factors

are directly involved in nuclear transport. One alternative

is that they represent components, possibly of the nuclear

envelope, necessary for, but not directly involved in

facilitated transport.

Components of the Pore Complex

It is clear that the translocation of transported

macromolecules across the nuclear envelope occurs through


the nuclear pores. Thus, the identification and

characterization of pore constituents could yield

information about the exchange mechanisms. The pore complex

has a mass of -125 X 106 daltons (Maul, 1977; Reichelt et

al., 1990). Because of its relative insolubility,

biochemical characterization of the pore complex has proven

to be difficult, and only a few protein components have been

identified. A 190kD, concanavalin A-binding, integral

membrane glycoprotein (gpl90) from vertebrate cells (Gerace

et al., 1982) was one of the first pore-associated

polypeptides identified. The primary structure deduced from

a cDNA sequence revealed a polypeptide of ~205kD with two

transmembrane segments (Wozniak et al. 1989). Including the

oligosaccharides, the total estimated mass is ~210kD, hence

the polypeptide was renamed gp210. It is present in -16-25

copies per pore complex, and is thought to function as a

membrane anchor for pore complex components and/or as a

nucleating site for pore complex assembly.

Berrios et al. (1983, 1986) identified an ATPase

closely related to non-muscle myosin heavy chain in

Drosophila and other eukaryotic cell nuclear envelope

fractions. Recent immunoelectron microscopic analysis

confirmed its localization to the nuclear pore complex

(Berrios et al., 1991). A model was proposed wherein the

annular subunits of the nuclear pore complex are formed by

the heads, the cylindrical walls of the pore lumen by the


tails, and the hydrolysis of ATP required for facilitated

transport through the pores is catalyzed by the head regions

of these molecules. However, this model has not been

substantiated by any direct evidence.

A family of at least eight o-linked n-

acetylglucosamine-modified nuclear pore glycoproteins (45-

210kD, called nucleoporins) have been localized to the

periphery of the nuclear pores (Davis and Blobel, 1986;

Finlay et al., 1987; Hanover et al., 1987; Holt et al.,

1987; Park et al., 1987; Snow et al., 1987;). Finlay et al.

(1987) demonstrated that WGA bound to n-acetylglucosamine

residues at or near the nuclear pores and completely

inhibited the facilitated nuclear transport of

fluorescently-labeled nucleoplasmin. Newmeyer and Forbes

(1988) found that WGA inhibited import but not the binding

of nucleoplasmin and human serum albumin-SV40 NLS-peptide

conjugates to the cytoplasmic face of the nuclear pores,

which suggested that a specific receptor site remained

accessible. These findings were confirmed by Yoneda et al.

(1987) and Dabauvalle et al. (1988); passive diffusion was

unaffected by WGA (DaBauvalle et al., 1988; Jiang and

Schindler, 1986; Yoneda et al., 1988), which suggested that

these glycoproteins functioned in the translocation step.

Antibodies raised against these proteins also inhibited

nuclear protein uptake and RNA efflux in Xenopus oocytes

(Featherstone et al., 1988). Additionally, pores


reconstituted from Xenopus egg extracts depleted of n-

acetylglucosamine-modified proteins were transport-deficient

(Finlay and Forbes, 1990), which supports the involvement of

these proteins in nuclear transport. Consistent with this

interpretation, Akey and Goldfarb (1989) localized these

glycoproteins to the transporter assembly of the nuclear


The most prominent member of this family, p62 (Hanover

et al., 1987), was cloned, sequenced, and found to contain

regions of homology with both collagen and filamentous

proteins including myosin and intermediate filaments (Starr

et al., 1990). Rat p62 overexpressed in monkey cells

associated with the nuclear envelope without interfering

with nuclear transport, suggesting that the nucleoporins are

functionally conserved. Consistent with this

interpretation, Finlay and Forbes (1990) demonstrated that

nuclear transport could be restored to transport-defective

reconstituted Xenopus nuclear pores by the addition of rat

liver pore WGA binding proteins. Additionally, two

essential Saccharomyces cerevisiae proteins have been

cloned. NSP1 (86kD; Nehrbass et al., 1990) and NUP1 (113kD;

Davis and Fink 1990) are recognized by monoclonal antibodies

against the nucleoporins. NSP1 was localized to the nuclear

pores by immunoelectron microscopy (Nehrbass et al., 1990),

and NUP1 was localized to the nuclear periphery by

immunofluorescence (Davis and Fink, 1990). Affinity-


purified anti-NSPl antibodies also stained the nucleus of

MDCK cells, and recognized an ~80kD protein on immunoblots.

Collectively, these findings suggest that the nucleoporins

are a conserved family of pore-associated glycoproteins that

might have similar functional roles in nucleocytoplasmic

transport and/or pore complex/envelope structure.

Nuclear Efflux

Paine (1975) compared the movement of different sized

exogenous tracers across the nuclear envelope of salivary

gland cell nuclei and found similar diffusion rates both

into and out of the nucleus. Stacey and Allfrey (1984) and

Lanford et al. (1986) further demonstrated that proteins

effectively too large to passively diffuse into the nucleus,

such as BSA and IgG, do not diffuse out. These findings

suggest that, analogous to protein import, the nuclear

efflux of transported proteins is a signal mediated,

facilitated process. However, nothing is known about the

mechanism, nor have any targeting sequences been described

for the facilitated efflux of any proteins. Nuclear uptake

signals do not appear sufficient to initiate the transport

of polypeptides from the nucleus to the cytoplasm (Feldherr

et al., 1984; suggested by Newmeyer et al., 1986a,b), which

suggests that efflux signals are separate and distinct from

uptake signals. However, the possibility that NLSs are

bound in the nucleus can not be excluded.

RNA Efflux

Because several features of RNA export appear analogous

to nuclear protein import, and most RNA leaves the nucleus

completed with protein in the form of RNP, the study of RNA

efflux and the proteins involved in this process might give

clues to the mechanisms involved in protein transport.

RNA Efflux Occurs through the Nuclear Pores

Analogous to protein uptake, RNA efflux occurs through

the nuclear pores. EM views of RNP particle export were

obtained for Balbani ring RNPs in Chironomus tentans

salivary gland cells (Stevens and Swift, 1966; Skoglund et

al., 1983; Mehlin et al., 1991) and RNP particles in Xenopus

oocytes (Franke, 1974) that clearly demonstrate the nuclear

pores are the sites of exchange. Consistent with these

observations, Dworetzky and Feldherr (1988) showed that gold

particles coated with tRNA, 5sRNA, Poly(A) and Poly(I) leave

the nucleus through the central regions of the pores.

Featherstone et al. (1988) found that a monoclonal antibody

against the nucleoporins inhibited the export of mature tRNA

and 5sRNA, while Bataill6 et al. (1990) showed that the

export of mature ribosomal subunits was inhibited by WGA.

These findings suggest that, analogous to protein uptake,

the nucleoporins might have a role in facilitated RNA


RNA Efflux is Facilitated

Transfer RNA

Transfer RNA originates from pre-tRNA transcripts which

undergo intranuclear processing to generate mature tRNAs.

Only mature tRNAs exit the nucleus. Peters (1986)

determined the stokes radius of tRNAPhQ to be 34A, which

suggests it is too large to rapidly diffuse from the

nucleus. Zasloff (1983) showed that efflux was temperature

dependent, saturable by free tRNA and could be competed by

different tRNA species, which strongly suggested that efflux

is a facilitated receptor mediated process.

Specific sequences and proper conformation are required

for efflux (Zasloff, 1982; Tobian, 1985); tRNA efflux rates

were reduced by single base mutations, and mutations which

alter 20 and 30 tRNA structure always perturbed both

processing and transport.

Ribosomal RNA

Eukaryotic ribosomes are composed of RNA and protein,

sediment at 80s, and are composed of two subunits, large and

small. Ribosomal proteins enter the nucleus and assemble

with nascent rRNA to form a pre-ribosomal particle. This is

cleaved, forming 40s and 60s subunits containing nearly

mature 18s and 28s:5.8s rRNAs, respectively. rRNPs,

including the ~1500-3000kD ribosomal subunits (Watson et

al., 1984), are too large to passively diffuse through the


pores. Additionally, Giese and Wunderlich (1983)

demonstrated that rRNP efflux from isolated Tetrahymena

nuclei was ATP- and temperature dependent, which are

characteristics of a facilitated process. Consistent with

this interpretation, Bataill6 et al. (1990) examined the

export of intact radiolabeled ribosomal subunits

microinjected into Xenopus oocyte nuclei, and showed that

transport from the nucleus was ATP- and temperature-

dependent and saturable, which also suggests efflux is

receptor mediated.


In oocytes, 5s rRNA associates with ribosomes by the

addition of separate 5s-containing ribonucleoproteins (RNPs)

in the cytoplasm (Steitz et al., 1988a). Newly transcribed

nuclear 5s rRNA transiently interacts with the La antigen, a

polypeptide probably involved in the termination of

transcription (Gottleib and Steitz, 1989). La is then

replaced by either ribosomal protein L5 or the 5s-specific

TFIIIA, and each of these two RNPs migrates out of the

nucleus and accumulates in the cytoplasm. RNA molecules

rendered unable to interact with L5 and TFIIIA are retained

in the nucleus (Guddat et al., 1990), which suggests that L5

and TFIIIA are involved in the efflux of 5sRNA. It is not

yet known whether these findings are conserved in somatic


Messenger RNA

Messenger RNA matures from nucleus-restricted hnRNA.

Only mature mRNAs are transported out of the nucleus (Jacobs

and Birnie, 1982; Kindas-MUgge and Sauermann, 1985), which

suggests that posttranscriptional maturation of mRNA from

hnRNA is necessary for mRNA transport. Posttranscriptional

modifications performed before efflux include splicing,

methylation, capping and polyadenylation (reviewed in

Agutter, 1988).

There is some evidence suggesting that mRNA is not

freely diffusible in the nucleus and cytoplasm, but instead

is associated with the nuclear matrix and cytoskeleton

(reviewed in Schrdder et al., 1988a). A current model

suggests that mRNA transport is a solid-state process

characterized by a) transport of RNA to the nuclear envelope

along nuclear matrix elements, b) translocation through the

nuclear pores, and c) transport along the cytoskeletal

matrix (Agutter, 1985a). Messenger RNA is translocated as

mRNP which is too large to diffuse through the nuclear

pores, and efflux has been shown to be both ATP- and

temperature-dependent (reviewed in Schrbder et al., 1988a).

These findings suggest that mRNA efflux occurs by

facilitated transport.

Two important questions regarding RNA efflux are 1) is

the RNA efflux signal contained in the associated proteins

or the RNA itself, and 2) what envelope-associated and


cellular components are involved in RNA transport? While

generalities cannot be made for all classes of RNA, most of

the data pertinent to these questions has come from the

study of mRNA efflux. Evidence from several groups suggests

that the poly(A) tail of mRNA might serve as a signal for

mRNA translocation, but this function cannot be general

since many mRNA species are not polyadenylated (reviewed by

Agutter, 1988). However, the poly(A) segment of mRNA is

associated with different poly(A)-binding proteins in the

nucleus and cytoplasm (Schwartz and Darnell,1976; van

Eekelen et al., 1981; Baer and Kornberg, 1983; Sachs and

Kornberg, 1985). Although there is no direct evidence that

these proteins have a role in transport, mRNA translocation

might be dependent upon dissociation of the nuclear poly(A)

associated proteins, while association with cytoplasmic

poly(A) binding proteins might prevent the message from

being transported back into the nucleus (Schrbder et al.


Messenger RNA efflux requires the activity of a nuclear

envelope-associated NTPase (for recent review see Schr6der

et al., 1987). Evidence based on similarities between the

properties of the NTPase and of nucleocytoplasmic RNA

transport with respect to substrate specificity, kinetics,

activation energy and sensitivity to inhibitors support its

role in mRNA efflux. Most importantly, however, is that the

activity of the enzyme in intact nuclear envelopes (Bernd et


al., 1982; Reidel et al., 1987) and purified enzyme

(Schr6der et. al., 1986) is stimulated by poly(A) or

poly(A)+ mRNA.

The regulation of mRNA efflux likely involves

intranuclear components including a nuclear envelope-

associated protein kinase (Lam et al., 1979) and phosphatase

(Steer et al., 1979). Schr6der et al. (1988b) has evidence

suggesting that the rate of ATP-dependent NTPase-mediated

nucleocytoplasmic translocation of poly(A)+ mRNA is

determined by the poly(A)-sensitive phosphorylation and

dephosphorylation of a Triton-insoluble, 106kD poly(A)

binding protein associated with the nuclear envelope. This

protein was shown to bind poly(A) in an ATP-dependent manner

and phosphorylation increased its affinity for poly(A)

(McDonald and Agutter, 1980; Bachmann et al., 1984).

Additionally, a monoclonal antibody against nuclear

envelope-localized poly(A)-binding proteins inhibited mRNA

efflux from isolated oviduct cell nuclei and mouse L-cell

nuclei (Clawson et al., 1984), consistent with a role for

envelope-associated poly(A) binding proteins in mRNA efflux.

A 34kD protein from liver cell cytoplasm that

specifically stimulated mRNA efflux from isolated nuclei has

been purified (Moffett and Webb, 1983; Agutter, 1985b). The

role of this protein, its mode of action, and whether it

also functions in protein efflux are still not known.


In summary, RNA efflux occurs by facilitated transport

through the nuclear pores. Some of the components involved

in this process have been identified, but their precise

roles and whether they are also required for facilitated

protein efflux remains to be determined.

Recycling Proteins

Evidence for the existence of proteins that shuttled

between the nucleus and cytoplasm was first presented by

Goldstein (1958) who used nuclear transplantation to study

the movement of radiolabeled nuclear proteins in Amoeba

proteus. Results suggested that nuclear proteins migrated

from the labeled nucleus, through the cytoplasm, and

accumulated in the originally unlabeled nucleus (Goldstein

and Prescott, 1967; Legname and Goldstein, 1972). These

proteins were partially characterized and found to be

heterogeneous with respect to size (Jelinek and Goldstein,

1973). Approximately 1/3 of the shuttling proteins were

represented by a single 2.3kD species that comprised ~17% of

the soluble nuclear protein and 3-4% of the total cellular

protein. The role of this or other amoeba shuttling

proteins has not been determined, nor are there any clues

about the mechanisms involved in their bidirectional


Additional shuttling proteins have been identified in

different systems. One of the first was the non-histone


chromosomal protein HMG1. Rechsteiner and Kuehl (1979)

showed that when unlabeled HeLa cells were fused with I-

labeled-HMGl-injected cells, the nuclear-localized labeled

protein equilibrated between both nuclei within 12 hours,

indicating that HMG1 shuttled between the nucleus and

cytoplasm. Because of its relatively small size (26kD), the

nucleocytoplasmic exchange of HMG1 could likely be the

result of diffusion and binding, rather than facilitated

transport. The relevance of recycling to its functional

role remains unclear.

Conclusive evidence that nucleolar proteins shuttled,

first suggested by Ege et al. (1971) and Ringertz et al.

(1971), was presented by Borer et al. (1989). They

monitored the migration of chicken nucleolar proteins in

polyethylene glycol-induced chick/mouse heterokaryons by

indirect immunofluorescence microscopy and found that the

chick nucleolar proteins nucleolin (92kD) and No38 (38kD)

accumulated in the mouse nuclei, independent of protein

synthesis. Antibodies against these proteins microinjected

into the cytoplasm of cultured chicken cells were carried

into the nucleus (also independent of protein synthesis),

which suggested that the antigens appeared transiently in

the cytoplasm. Non-specific IgG, which is too large to

diffuse across the nuclear envelope, was excluded from the

nucleus. Since the nucleolus is a site of ribosome

biogenesis, and recycling is a property that might be


expected of a carrier protein, Borer et al. (1989) suggested

that these proteins might be involved in the

nucleocytoplasmic exchange of ribosomal components. Because

nucleolin is probably too large to enter and leave the

nucleus by passive diffusion, its exchange likely requires

bidirectional facilitated transport.

Some proteins, including the cAMP-dependent protein

kinase, appear to enter and leave the nucleus in response to

specific signals. Most, if not all physiological effects

produced by elevation in cAMP levels in vertebrate cells are

mediated by the activation of cAMP-dependent protein kinases

(for review, see Cohen, 1982). At least two major forms

have been distinguished in most species, referred to as type

I and type II. Both are composed of two regulatory and two

catalytic subunits (for review, see Flockhart and Corbin,

1982). Nigg et al. (1985) used indirect immunofluorescence

microscopy to show that the catalytic subunit of the type II

cAMP-dependent protein kinase, normally associated with the

Golgi complex, dissociated from the Golgi-associated

regulatory subunits and accumulated in the nucleus in

response to adenylate cyclase stimulation. The effect was

rapidly reversible in that the catalytic subunits

reassociated with the Golgi minutes after cessation of

adenylate cyclase stimulation. Functionally, these findings

suggest that nuclear translocation of activated protein

kinase subunits could represent an important link between


hormonal stimulation and physiological responses. It cannot

be inferred whether translocation of the kinase requires

facilitated transport because of its relatively small size


A second example of a signal-mediated recycling protein

is the glucocorticoid steroid receptor (GR). Steroid

hormone receptors function in the nucleus by binding to DNA

in response to hormonal stimulation and activate specific

genes. The GR is located primarily in the cytoplasm, and is

translocated to the nucleus only after steroid is bound

(reviewed in Pratt, 1990). Each cytoplasmic GR monomer is

associated with two molecules of the 90kD heat shock protein

hsp90 (Catelli et al., 1985). Hormone activation results in

the dissociation of hsp90 and nuclear translocation of GR.

Picard and Yamamoto (1987) defined two distinct NLSs present

in the GR, NLS1 and NLS2. GR without NLS2 was always

excluded from the nucleus. Because nuclear localization was

hormone-dependent, they suggested that hormone binding

dissociated hsp90 or caused a conformational change, which

exposed NLS2.

The mechanisms involved in the nuclear efflux of the GR

are not known. Cellular levels remain constant and

unliganded cytoplasmic receptors are regenerated from

nuclear localized receptors in the presence of protein

synthesis inhibitors (Munck et al., 1972; Bell and Munck,

1973; Raaka and Samuels, 1983; Munck et al., 1984),


indicating that it indeed recycles. Studies have suggested

that a change in the phosphorylation state of the GR might

be required for efflux (Mendel et al., 1990), but this

interpretation is only speculative.

Another protein that recycles in response to signals is

the 70kD heat shock protein hsp70. Indirect

immunofluorescence with hsp-specific monoclonal antibodies

on Drosophila tissue culture and salivary gland cells

demonstrated that after its rapid, stress-induced

nuclear/nucleolar accumulation, hsp70 was redistributed (in

4-5h) back to the cytoplasm (Velaquez and Lindquist, 1984).

Because no significant breakdown of hsp70 was reported

during the time frame of these experiments (Mirault et al.,

1977), the cytoplasmic signal likely reflected the efflux of

intact proteins. The role of hsp70 is still not clear,

though it was suggested that its localization to the

nucleus/nucleoli is required for the recovery of nucleolar

function (Pelham, 1984; Lewis and Pelham, 1985). As with

nucleolin, hsp70 and related proteins that shuttle between

the nucleus and cytoplasm can serve as models for the study

of recycling proteins and facilitated nuclear efflux.

Hsp70-related proteins play a role in protein translocation

into the ER and mitochondria (Murikami et al., 1988;

Zimmerman et al., 1988; Chirico et al., 1988; Deshaies et

al., 1988; Kang et al., 1990), and a role in


nucleocytoplasmic exchanges might be consistent with this

general function.

Statement of Research Topic

Although there is evidence suggesting that protein

efflux is both facilitated and signal-mediated, there is

little functional data regarding the transport of proteins

from the nucleus. The signal for protein efflux is still

undefined, and it is not clear whether uptake and efflux

signals are necessarily distinct. Most of our knowledge of

efflux comes from the study of RNA, but it has not been

determined whether protein efflux is governed by the same

components and mechanisms.

A polypeptide capable of bidirectional exchange

(recycling) could serve as an ideal model to further

investigate protein transport. One objective of this study

is to identify an endogenous recycling protein in the

Xenopus oocyte, because this system is readily amenable to

studies of nuclear transport. We have identified two 70kD

Xenopus oocyte proteins, B3 and B4, as potential recycling

proteins because 1) they are present in both the nucleus and

cytoplasm and 2) unlike other targeted nuclear proteins such

as nucleoplasmin (Dingwall et al., 1988) that accumulate in

the nucleus, B3 and B4 equilibrate between the nucleus and

cytoplasm in ~3 hours (Feldherr and Ogburn, 1980). Thus,

our first objective is to determine if these specific


polypeptides recycle across the nuclear envelope.

Based on similarities in molecular weight and

isoelectric points, it was suggested that these polypeptides

might be constitutively expressed heat shock proteins

similar to the heat shock cognates (hsc70s) characterized in

Drosophila (Craig et al., 1983; Horrell et al., 1987; King

and Davis, 1987;). This is especially interesting because

heat shock proteins have been shown to facilitate protein

transport into mitochondria and the ER (Chirico et al.,

1988; Deshaies et al., 1988; Murikami et al., 1988;

Zimmerman et al., 1988; Kang et al., 1990). Therefore,

another objective is to determine whether B3 and B4 are

members of the 70kD family of heat shock proteins.

To learn more about domains necessary for nuclear

uptake and efflux, experiments were performed to define

regions of an hsp70-related polypeptide involved in

bidirectional transport. Limited proteolytic digestion was

successfully employed to define domains important for the

uptake of nucleoplasmin (Dingwall et al., 1982), and a

similar strategy was employed in the following study. The

results of these experiments might indicate whether

additional signals are involved in uptake and if single

domains are involved in bidirectional transport.



Two prominent 70kD Xenopus laevis oocyte proteins,

designated B3 and B4 (pI 5.58 and 5.75, respectively), are

present in equivalent concentrations in both the nucleus and

cytoplasm (DeRobertis et al., 1978; Feldherr and Ogburn,

1980). These polypeptides are also present in adult Xenopus

tissues, including liver, kidney, pancreas, brain, and heart

(Dreyer and Hausen, 1983). Since B3 and B4 are not

restricted to oocytes, it is likely that they have a

universal role in cell function.

King and Davis (1987) and Horrell et al. (1987)

suggested that B3 and B4 might be constitutively expressed

heat shock proteins, similar to the heat shock cognates

(hsc70s) characterized in Drosophila (Craig et al., 1983).

These conclusions were based on similarities in the

molecular weights and isoelectric points of these proteins.

Most hsp70-related proteins have an endogenous

substrate-dependent ATPase activity which is essential for

their function (reviewed by Lindquist and Craig, 1988).

Determining whether B3/B4 have endogenous ATPase activity,

and whether this activity can be modulated by cellular


components involved in nuclear transport might also provide

clues to whether they play a role in nuclear transport.

The objectives in this study were to determine whether

B3 and B4 are constitutively expressed heat shock proteins,

and whether they possess the endogenous ATPase activity

common to most members of the 70kD family of heat shock

proteins (Lindquist and Craig, 1988). We demonstrate, based

on several lines of evidence, that B3 and B4 are 70kD heat

shock proteins. We were unable to detect an endogenous

ATPase activity in B3/B4, but interestingly, B3/B4 reduced

the activity of a nuclear envelope-associated ATPase(s).

Materials and Methods

Xenopus laevis were purchased from Xenopus I, Ann Arbor,

MI., maintained in artificial pond water, and fed beef heart

every other day. Ovaries were surgically removed from frogs

anesthetized on ice. O-R2 medium (Wallace et al., 1973) was

used as an extracellular medium in all experiments.

Isolation of B3, B4 and hsc70

B3 and B4 were isolated using a modification of the

procedure described by Welch and Feramisco (1985).

Approximately 25ml of isolated ovary was homogenized in a

Dounce homogenizer (Wheaton) on ice. An equal volume of an

ice-cold solution containing 10mM Tris-acetate pH 7.5, 10mM

NaCl, and 0.1mM EDTA was added, and the homogenate was


centrifuged at 100,000g for 60 minutes at 4C. The

supernatant was clarified twice with 1,1,2-

trichlorotrifluoroethane (Laskey et al., 1978). The

remainder of the procedure followed that described by Welch

and Feramisco except the ATP-agarose column was washed with

3mM GTP before being eluted with 3mM ATP. The ATP eluent

was extensively dialyzed against 15mM ammonium bicarbonate,

and concentrated to 4ml using an Amicon ultrafiltration cell

(model 8010, Amicon Corp., Danvers, MA). It was then

aliquoted into 0.5ml fractions, frozen in dry-ice/acetone,

and lyophilized to dryness. One to 2mg of protein was

obtained from 25ml of ovary.

Rat hsc70 was purified from brains collected from three

freshly guillotined rats. The procedure of Welch and

Feramisco was followed except the DEAE step was omitted.

Purity of both B3 and B4 and rat hsc70 were determined

by two-dimensional isoelectric focusing polyacrylamide gel

electrophoresis (O'Farrell and O'Farrell, 1977). The

proteins were stored dry at -20C.

Since B3 and B4 co-purify, the isolate will be referred

to in the subsequent discussion as B3/B4.


To generate polyclonal antibodies, two female New

Zealand White rabbits were immunized subcutaneously every 14

days with B3 or B4 protein spots cut from unfixed two-


dimensional gels of whole oocyte protein. Prior to

injection, the gel spots were homogenized in RIBI adjuvant

system (Immunochem Research Inc., Hamilton, MT).

Approximately 20pg of B3 or B4 was used for each

immunization. After the third immunization, Western blots

of two-dimensional gels of oocyte protein (Towbin et al.,

1979) were probed with serum from each rabbit. Both

polyclonal antisera reacted with B3 and B4. Polyclonal

antibodies were also prepared against isolated B3/B4 using

the same injection regimen described above. Approximately

20pg of isolated B3/B4 were used for each immunization. The

antibodies were purified by passing sera over an affinity

column made by conjugating purified B3 and B4 with CNBr-

Sepharose 4B (Sigma). The column was washed with 20mM Tris-

HCl (pH 7.2), 300mM NaCl, and bound antibodies were eluted

by a low pH wash (0.1M glycine-HCl pH 2.6, 500mM NaCl). The

eluent was dialyzed into 20mM Tris-HCl pH 7.2, 100mM NaCI,

and the IgGs were purified using Protein A-Sepharose


Cyanogen Bromide-Cleavage Peptide Mapping

Cyanogen bromide-cleavage peptide mapping was performed

using a modification of the procedure described by Sokolov

et al. (1989). Approximately 50pg of purified B3 and B4,

and 25pg of purified rat brain hsc70 were resolved on


two-dimensional gels. The proteins were visualized by

staining unfixed gels for 15 minutes with 0.0625% Coomassie

Blue in 25mM KC1, 3mM NaCl, 1.8mM K2HPO4, 1.2mM KH2PO4, pH

7.0, and destaining in deionized water for 5 minutes. B3,

B4 and hsc70 spots were cut from the gels, fragmented and

lyophilized to dryness. 454l (per gel spot) of CNBr

(200mg/ml in 70% formic acid) was added to tubes containing

B3, B4 or hsc70. The gel fragments were incubated for 12

hours at 37C, then lyophilized to dryness. They were then

rehydrated with 100O1 deionized water, and relyophilized.

This step was repeated twice. 40pl (per spot) of complete

sample buffer at pH 7.2 (Laemmli, 1970) containing

bromophenol blue was added to each tube. After a 10 minute

incubation period (at 21C), the samples were neutralized

with pl aliquots of 3M NaOH until the indicator turned from

yellow/green to blue. The fragments were then run at the

equivalent of 2 protein spots per lane on a 12% SDS-PAGE

(Laemmli, 1970). They were then electrophoretically

transferred to a polyvinylidine difluoride (PVDF) membrane

(Immobilon, Millipore) in 20mM CAPS (Sigma; pH 9.9,

dissolved in 10% methanol). The blots were stained with

0.1% Coomassie Blue (Serva) in 80% methanol and air dried.

Protein Sequencing

CNBr-cleaved peptide fragments were fractionated and

resolved as described above. Fragments to be sequenced were


cut from the dried Immobilon membrane and applied directly

to an ABI 470A gas-phase amino acid sequencer with an on-

line PTH analyzer (Applied Biosystems Inc., Foster City,

CA.). Sequencing was performed by the Protein Chemistry

Core Facility at the University of Florida.

Phosphate Labeling

Two different procedures were employed. First, cells

were metabolically labeled with 500pCi/ml of [y32P]-

orthophosphate as described by Khrone and Franke (1980),

except the incubation medium was O-R2 (Wallace et al.,

1973). Second, oocytes were injected with 50nl of imCi/ml

[y P]-orthophosphate as described by Karsenti et al.

(1987), except the injection medium was 12mM K2HPO4, 102mM

KC1 and ll.lmM NaCl.

Sixteen hours after metabolic labeling or 45 minutes

post injection, the oocytes were washed and fixed in 10% ice

cold TCA. The cells were extracted with 3 changes of 50mls

10% TCA for 18 hours at 40C, then with 50ml acetone for 6

hours. The distribution of 32P was determined by

autoradiographic analysis of proteins from four cells from

each group homogenized in 80pl of sample buffer and

fractionated by two-dimensional isoelectric focusing PAGE

(O'Farrell and O'Farrell, 1977).

ATPase Assay

We essentially followed the procedure described by

Green and Liem (1989) except for the following

modifications: The assay was performed in 20mM HEPES pH

7.4, 25mM KC1, 2.0mM MgCl2, 0.3mM MnCl2, 0.5mM CaCl2 and

0.8mM DTT. Each experiment contained 20pM ATP and lpCi

[y"P]-ATP in a final volume of 40pl. The concentration of

the remaining components, when included, were as follows:

B3/B4, 0.375pM; BSA, 3.67pM; nucleoplasmin, 3.63pM; 10pg

nuclear envelopes/40pg assay. Poly(A) was used at a

concentration of 0.66pM based on the results obtained by

Bernd et al. (1982) for stimulation of the rat liver

poly(A)-dependent nuclear envelope NTPase. Quantitation of

ATP hydrolysis was performed using a modification of the

procedure described by Richter and Klink (1978). The

reactions were terminated by the addition of 10 volumes

(400pl) of an ice-cold 40mg/ml slurry of activated charcoal

(Sigma) in 2% TCA. The mixture was vortexed and incubated

on ice for 10 minutes. The charcoal was pelleted by

centrifugation for 15 minutes (Eppendorph centrifuge model

5415), and 300pl aliquots were assayed for free 32p in a

Beckman LS5000TD scintillation counter.

Isolation of Nuclear Envelopes

Rat liver cell nuclear envelopes were prepared

according to the procedure of Kaufmann et al. (1983).


Additionally, the envelopes were washed once by

centrifugation and resuspension with 20mM HEPES pH 7.4, 25mM

KC1, 2.0mM MgCl2, 0.3mM MnCI2, 0.5mM CaCI2 and 0.8mM DTT.

Isolation of Nucleoplasmin

Nucleoplasmin was purified from mature Xenopus oocytes

by affinity chromatography using affinity-purified anti-

nucleoplasmin antibodies coupled to sepharose beads

(Dworetzky and Feldherr, 1988). Purity was confirmed by

one-dimensional gel analysis.


Isolation and Characterization of B3, B4 and rat hsc70

A typical Coomassie Blue stained two-dimensional gel

obtained for Xenopus oocyte nuclei is shown in Figure 1A.

B3 and B4 have apparent molecular weights of 70kD, and pI's

of 5.58 and 5.75, respectively.

A Coomassie Blue stained two-dimensional gel of

proteins purified from whole oocyte 100,000 g supernatant

using ATP affinity chromatography is shown in Figure 1B; two

major polypeptides are apparent. Based on their apparent

molecular weights (70kD), pI's (5.6 and 5.8), and reactivity

with B3- and B4-specific polyclonal antibodies (Figure IC),

we concluded that these isolated polypeptides are B3 and B4.

B3 and B4 might be products of the same gene that have

different isoelectric points because of differential post-


translational modifications. One modification that can

result in large changes in pI is phosphorylation.

Phosphorylation experiments that assayed 1) steady-state

phosphorylation in oocytes metabolically labeled with ["P]-

phosphate for 16 hours and 2) rapid phosphorylation events

in oocytes 45 minutes post microinjection with [32P]-

phosphate indicated that B3 and B4 are not phosphoproteins

(data not shown).

A two-dimensional gel of the purified rat hsc70 is

shown in Figure 2. One major species is present, with an

apparent molecular weight of 70kD and a pI of 5.65. These

values are consistent with those reported for hsc70 (Pelham,

1984; Pelham, 1986; Lindquist and Craig, 1988). Its

identity was confirmed by comparing the amino-terminal

sequence of a 31kD CNBr-cleavage peptide fragment (see

below) to the predicted sequence of hsc70 derived from a

full length cDNA clone (O'Malley et al., 1985). The two

sequences are compared in Table 1.


Obtained from 31kD Fragment V N F I A F K K K K
Predicted from cDNA V NHF I A E F K RK HK K

* The amount of material was not sufficient to resolve
these residues

Figure 1. Purification of B3 and B4.

Panel A shows a two-dimensional gel of proteins
from 40 oocyte nuclei stained with Coomassie Blue.
The positions of B3 and B4 are indicated.
Isoelectric points and molecular weights are
indicated along the abscissa and ordinates,

Panel B shows a Coomassie Blue stained two-
dimensional gel of B3 and B4 purified from whole
oocyte 100,000g supernatant by ATP affinity

Panel C shows a Western blot of a two-dimensional
gel of 40 oocyte nuclei probed with anti-B3/B4
rabbit serum (1:100 dilution). Antibodies were
visualized using alkaline phosphatase-conjugated
goat-anti-rabbit secondary antibodies (Sigma).
The small labeled spots adjacent to the major
spots are most likely isoforms of B3 and B4.
Identical results were obtained using either anti-
B3 or anti-B4.

6.3 5.8

5.3 4.8

- *

7.4 6.8








0 m

4 +) 0

U 4J

4-1 U 0
4 -4
-4 0
r.0 v
C 0 0
0I to4

$4 U

CO $4 W 5
0 0 *i

C: 0 00
04 0

0 0 9

U # 0 )


0 4
CP 0

-r4 : $4
SS '4*
C I H-

3) -H to
~4J 00
to oc 00

4i.4 4.)
0 *4M0
*H C *^ H

S.0 04
S+1 U

40 4. 0

(1O too
3 *-40

r4 52.-Iv
4.)0 0
#- *C

*. 0 4

04: >H
< HN
0^ 3 CO
(S U-I -
4JC4-4 0

*2 $4 $4

H (02
t0 -> f
C 0 o<
H N^>



a w




(0 CO
0) (0



w i


Affinity-purified anti B3/B4-specific polyclonal antibodies

were found to be strongly cross-reactive with isolated hsc70

(Figure 2, Inset). The minor spots present in the gel shown

in Figure 2 are probably impurities since they do not react

with anti-B3/B4 antibodies. In addition, these antibodies

were found to be specific for hsc70 when 2D Western blots of

rat brain 100,000g supernatant were probed (data not shown).

CNBr-cleavage mapping studies were performed with

isolated B3, B4 and hsc70. The results are shown in Figure

3. The patterns obtained for the three polypeptides are


ATPase Assay

We were unable to detect any ATPase activity for B3/B4

alone, or in the presence of a 10-fold molar excess of BSA

or Xenopus nucleoplasmin (data not shown), but we can not

rule out the possibility that some activity might be

detected with a more sensitive assay or under different

conditions of, for example, temperature or ion

concentration. Figure 4A shows that ATPase activity was

detected in the nuclear envelope preparation, as expected,

because the envelopes have been shown to contain at least 3

different NTPases (matrix, envelope and pore-associated; for

review, see Schroder, 1988a). Interestingly, B3/B4 reduced

the envelope-associated ATPase activity. The activity was

not affected by BSA or nucleoplasmin, indicating the affect

Figure 3. Cyanogen Bromide Cleavage Maps of Rat Hsc70 and
Oocyte B3 and B4.

Digested peptides were resolved on a 12% SDS-
polyacrylamide gel, then electrophoretically
transferred to a PVDF membrane and stained with
Coomassie Blue. The patterns obtained for rat
hsc70 (lane 1), oocyte B3 (lane 2) and B4 (lane 3)
are not distinguishably different.





ia\ -




was not simply due to the addition of protein or nuclear

targeted protein (nucleoplasmin). The inhibitory effect of

B3/B4 was not competed by the addition of 10-fold molar

excesses of BSA or nucleoplasmin. To address whether B3/B4

was affecting the poly(A)-dependent NTPase, the effect of

poly(A) and B3/B4 on the nuclear envelope were also examined

(Figure 4B). The addition of poly(A) should have increased

the envelope-associated ATPase activity (Bernd et al.,

1982), which was not observed. This suggests that either 1)

the NTPase required for mRNA transport was inactive in this

experimental system, or 2) the poly(A) was insufficient for

its stimulation. Therefore, while nuclear envelope-

associated NTPase activity was inhibited by B3/B4, we could

not determine which NTPase(s) was inhibited.


In this study, we obtained data that show the

following: 1) B3 and B4 have molecular weights and

isoelectric points consistent with those established for

hsp70-related polypeptides (Pelham, 1986; Lindquist and

Craig, 1988); 2) B3 and B4 bind with high affinity to ATP, a

property shared by hsp70-related proteins (Welch and

Feramisco, 1985; Chappell et al., 1986; Lindquist and Craig,

1988); 3) B3 and B4-specific polyclonal antibodies cross-

react with a well characterized, constitutively expressed

rat heat shock protein, hsc70, and 4) a comparison of

Figure 4. Relative ATP Hydrolysis.

Time (minutes) and cpm are indicated along the
abscissa and ordinate, respectively. NE, nuclear
envelopes; NE+BSA, nuclear envelopes with BSA;
NE+NP, nuclear envelopes with nucleoplasmin;
NE+B3/B4, nuclear envelopes with B3/B4;
NE+B3/B4+BSA, nuclear envelopes with B3/B4 and
BSA; NE+B3/B4+NP, nuclear envelopes with B3/B4 and

5 10 15 20

TIME (minutes)


CNBr-cleavage maps of B3, B4 and hsc70 reveals no

distinguishable differences. These results provide

conclusive evidence that B3 and B4 are members of the 70kD

family of heat shock proteins.

The ATPase assay was designed to closely resemble the

conditions for the assay described by Green and Liem (1989),

as well as those for in vitro nuclear transport studies

performed by Agutter (1983), Markland et al. (1987) and

Reidel and Fasold (1987). While no endogenous ATPase

activity was detected for B3/B4 by our assay, they did

specifically reduce nuclear envelope-associated activity.

Further studies employing broader assay conditions and

determine which specific NTPases are affected might provide

additional clues to the function of B3/B4.

Interestingly, Bernd et al. (1982) demonstrated that

microtubules inhibited the activity of the poly(A)-dependent

NTPase activity of isolated nuclear envelopes. Hsp70-

related proteins, including hsc70 (B-internexin), copurify

with isolated microtubules (Weatherbee et al., 1978, 1980;

Green and Liem, 1989). Thus an attractive possibility is

that the NTPase inhibition was actually caused by the heat

shock proteins. Investigations using hsp70-depleted

microtubule preparations are needed to resolve this




Metabolically-labeled B3 and B4 equilibrate between the

nucleus and cytoplasm in approximately three hours (Feldherr

and Ogburn, 1980). Exogenous proteins of equivalent size

that are known to enter the nucleus by passive diffusion

require a much longer time to equilibrate (e.g., BSA;

Bonner, 1975; Paine, 1975; Paine et al., 1975), indicating

that the uptake of B3 and B4 is transport-mediated. There

are two likely explanations why these polypeptides do not

accumulate in the nucleoplasm like other targeted oocyte

nuclear proteins such as nucleoplasmin (Dingwall et al.,

1988) and N1/N2 (Kleinschmidt and Seiter, 1988). First,

only a portion of the cytoplasmic pool of B3 and B4 might be

available to the nucleus. This could be the result of

either binding or compartmentalization within the cytoplasm.

Second, B3 and B4 might shuttle between the nucleus and


The objective of this study is to determine if B3/B4

are capable of recycling between the nucleus and cytoplasm.

Both hsp70 and hsc70 are concentrated in the nucleus in

response to heat or chemical stress (Welch and Feramisco,



1984; Pelham, 1984; Pelham, 1986) and apparently are

redistributed back to the cytoplasm during recovery

(Velaquez and Lindquist, 1984; Welch and Mizzen, 1988). It

was not determined, however, whether these proteins are able

to continuously recycle under non-stress conditions.

Demonstrating that B3 and B4 are recycling heat shock

cognate proteins could give useful clues as to the functions

of this family of polypeptides, especially since they are so

highly conserved (Lindquist and Craig, 1988). For example,

recycling would be consistent with the hypothesis that these

proteins act as carriers in signal-mediated nuclear

transport. By analogy, hsp70 has been shown to be involved

in the translocation of proteins across mitochondrial and

endoplasmic reticulum membranes (Chirico et al., 1988;

Deshaies et al., 1988).

Microinjection and cell fusion studies provide evidence

that B3 and B4 are transported across the nuclear envelope

and are able to recycle between the nucleus and cytoplasm.

Comparable experiments performed with rat hsc70 demonstrate

that it behaves similarly to B3 and B4 when microinjected

into oocytes. This suggests recycling could be a conserved

property of the 70kD heat shock cognates.


Materials and Methods

Protein lodination

Approximately 40yg of purified B3/B4, rat hsc70 or BSA

(Sigma) were reacted with Bolton and Hunter reagent

(Amersham; Bolton and Hunter, 1973). The labeled proteins

were separated from uncoupled label using a Sephadex G-25

column equilibrated with intracellular medium, which

contains 102mM KC1, 11.1mM NaCI, 7.2mM K2HPO,, 4.8mM KH2PO4

(Feldherr and Pomerantz, 1978). The average specific

activities for B3/B4, hsc70 and BSA were approximately

5.0x107, 6.3xl07, and 1.0x109 cpm/nmole, respectively. For

B3/B4, hsc70 and BSA, respectively, approximately 4500, 4000

and 80,000 cpm were injected into the cytoplasm, and 450,

400 and 5400 cpm were injected into the nucleus.


Microinjection experiments were performed using

manually defolliculated late stage 5 or stage 6 oocytes

(Dumont, 1972). For cytoplasmic injections, the pipettes

were calibrated to deliver approximately 50nl of solution.

The injectates, approximately 4ng of labeled protein in

intracellular medium, were directed into the vegetal

hemisphere of the oocytes to avoid accidental contact with

the nucleus. It is estimated that the amount of protein

injected represented between 0.8 and 2.0% of the total

cellular B3/B4. For nuclear injections, cells were


centrifuged for 8-10 minutes at approximately 650g (Kressman

and Birnsteil, 1980; Feldherr et al., 1984). As a result of

centrifugation, the nucleus migrates to the inner surface of

the animal pole and, due to the displacement of pigment

granules, can be readily identified. Nuclei were injected

with 5nl of the protein solutions. All microinjection

experiments were carried out at 210C in O-R2 medium.

Quantitation of Nuclear Uptake of Injected Proteins

To measure the nuclear uptake rates of B3/B4 and rat

hsc70, iodinated proteins were microinjected into oocyte

cytoplasms, and the cells were manually enucleated 1, 2, 3,

6 and 15 hours post-injection (20 cells per time point).

125I-labeled BSA controls were injected and enucleated after

6 and 15 hours. Isolated nuclei and cytoplasms were

transferred into 95% ethanol immediately after fractionation

(Feldherr and Richmond, 1977). To determine the amount of

label in nuclear and cytoplasmic fractions at each time

point, 125I was measured in pooled fractions using a Packard

Auto-Gamma Scintillation Spectrometer (model 5110). To

determine if there was breakdown of the injected proteins

and also confirm the quantitation obtained by gamma

counting, aliquots of nuclear and cytoplasmic fractions were

analyzed by one-dimensional gel autoradiography (Towbin et

al., 1979), and quantitated by densitometry (Soft Laser

Scanning Densitometer, Model SLR-2D/1D, Biomed Instruments,


Fullerton, CA). Each lane was analyzed separately to

determine the amount of breakdown in the individual cellular

compartments. The linear range for densitometric analysis

was established using standard curves of 3-4 different time

exposures of the same gel.

Quantitation of Nuclear Efflux of Injected Proteins

To study nuclear efflux, we injected iodinated B3/B4,

rat hsc70 or BSA directly into oocyte nuclei. Two hours

post-injection, cells were enucleated and both nuclear and

cytoplasmic fractions were immediately transferred to 95%

ethanol. 125I was measured in individual nuclei and their

corresponding cytoplasms as described above. Analysis of

individual cells made it possible to distinguish between

successful injections, in which the injectate was introduced

entirely into the nucleus, and unsuccessful injections in

which there was leakage into the cytoplasm. Efflux

experiments were performed for two hours to minimize the

long-term effects of puncturing the nuclear envelope.

Aliquots of nuclei and cytoplasms were also analyzed by

one-dimensional gel autoradiography, as described above, to

determine if there was breakdown of the injected protein and

to confirm the results obtained by direct gamma analysis.


Measurement of Endogenous Nuclear and Cytoplasmic

Concentrations of B3 and B4

To determine the intracellular distribution of B3 and

B4, double-label analysis was employed as described

previously (Feldherr and Ogburn, 1980) with some

modifications. Experimental cells were metabolically

labeled with 3H-leucine and cell fractions were

simultaneously electrophoresed with nuclei isolated from

35S-methionine-labeled cells. The 35S-labeled nuclei served

as internal standards. To avoid labeling the surrounding

follicle cells that are known to contain hsp70 (Horell et

al., 1987), 40nl of intracellular medium containing lpCi of

3H-leucine (143Ci/mmol) was injected into the cytoplasm of

oocytes, and the cells were immediately transferred to O-R2

medium containing 3mM cold leucine in order to dilute any

labeled amino acids that may have leaked from the oocytes

following injection. Cells were transferred to O-R2 medium

without leucine after 2 hours. The oocytes were enucleated

nine hours post-injection, and the nuclei and cytoplasms

were fixed in 95% ethanol. Both fractions were resolved on

two-dimensional gels, and B3 and B4 protein spots were

excised. Radioactivity was measured using a Beckman

LS5000TD liquid scintillation counter, and the H:35S ratios


Stabilization of Gold Colloid

Colloidal gold particles 20-120A in diameter were

prepared by reducing chloroauric acid with a saturated

solution of white phosphorous in ether (Feldherr, 1965).

The gold sols were stabilized with B3/B4, concentrated,

dialyzed, and microinjected as described by Dworetzky and

Feldherr (1988). Fixation and electron microscopic analysis

were also performed according to Dworetzky and Feldherr


Effect of Temperature on Uptake of B3 and B4

Oocytes were injected cytoplasmicly with 125I-labeled

B3/B4 or 125I-labeled myoglobin and were either incubated at

210C or immediately transferred to O-R2 at 40C. Fourteen to

17 oocytes were injected in each experimental group. Three

hours after injection the oocytes were enucleated, and the

radioactivity in isolated nuclear and cytoplasmic fractions

was determined as described above.

Hemi-Cell Fusion Experiments
125I-labeled B3/B4 or 125I-labeled BSA was microinjected

into the cytoplasm of defolliculated oocytes. The cells

were allowed to equilibrate for 18 hours at room temperature

in O-R2 medium. Following the procedure of Feldherr et al.

(1988), injected cells were then bisected under oil along

their animal-vegetal boundaries. Labeled vegetal


hemispheres were then fused with unlabeled nucleatedd)

animal hemispheres. After 4 hours, nuclei were isolated and

both nuclear and cytoplasmic fractions were fixed in 95%

ethanol and analyzed for 125I. One-dimensional gel

autoradiography was also performed to control for breakdown

of the labeled protein and also confirm the quantitation

obtained from the gamma counter. The values for B3/B4 and

BSA were obtained from 7 and 8 fusions, respectively.


Nucleocytoplasmic Distribution of Endogenous B3 and B4

Double-labeling experiments were performed to establish

the intracellular distribution of endogenous B3 and B4. The

results of two experiments, performed on groups of 15 and 18

cells, are shown in Table 2.


% of total protein in the nucleus*

Protein Exp. 1 Exp. 2

B3 6.4 10.7
B4 10.1 11.0

These results were obtained 9 hours after radiolabeling.
Based on the data shown in Figure 5A, equilibrium between
the nucleus and cytoplasm should have been achieved by this


We found that an average of 9.5% of the endogenous B3 and B4

are localized in the nucleus at equilibrium. Correcting for

yolk content, the nucleus occupies approximately 12% of the

total cell volume (Bonner, 1975).

Nuclear Uptake of 12"I-labeled BSA, B3 and B4

Figure 5A shows the data from three separate 125I-

labeled B3/B4 microinjection experiments. Equilibration of

labeled B3 and B4 was nearly complete by 6 hours post-

injection, and at 15 hours approximately 9.0% of the total

injected 125I-labeled B3/B4 was present in the nucleus. The

uptake kinetics are consistent with those reported for

endogenous B3 and B4 (Feldherr and Ogburn, 1980). The

intracellular distribution is also similar to that obtained

for the endogenous species (Table 2). Thus, the labeled

tracers reflect the migration patterns of endogenous B3/B4.

Nuclear uptake of 125I-labeled B3/B4 at equilibrium was

approximately 25-fold greater than that observed for

microinjected 12I-BSA (9.1% vs. 0.36%; see Figure 5A).

Figure 6A is an autoradiograph of a one-dimensional gel

of nuclear and cytoplasmic aliquots from injected cells.

There was no significant breakdown of the labeled proteins.

Densitometric analysis demonstrated that greater than 95% of

the labeled proteins in the cells had an apparent molecular

weight of 70kD. A minor 68kD species is present in the

cytoplasmic, but not nuclear aliquots. Whether this

Figure 5. Nuclear Uptake Kinetics of 125I-labeled B3/B4 (A)
and 12.I-labeled Rat Hsc70 (B).

Open circles, B3/B4; closed circles, hsc70; open
squares, BSA controls. Each curve represents an
independent experiment.

Hours Post-Injection

Figure 6. Autoradiographs of One-dimensional Gels of Nuclear
and Cytoplasmic Fractions Following Cytoplasmic
Injection of 12sI-B3/B4 (A) and 125I-hsc70 (B).

In both A and B, equivalent counts were applied to
each lane of a one-dimensional gel from fractions
taken 1 (lanes 1 ans 2), 2 (lanes 3 and 4), 3
(lanes 5 and 6), 6 (lanes 7 and 8) and 15 (lanes 9
and 10) hours after injection. Nuclear fractions
are shown in lanes 1, 3, 5, 7 and 9, and
cytoplasmic fractions are shown in lanes 2, 4, 6,
8 and 10. It has not been established whether the
minor low molecular weight bands seen in A are
breakdown products or impurities in the

Inset, Figure 5A. Autoradiograph of a two-
dimensional gel showing labeled B3 and B4 from a
nuclear fraction 6 hours after a cytoplasmic
injection. The migration pattern is
indistinguishable from that observed for the
isolated protein shown in Figure 1B.

1 2 3 4 5 6 7 8 9 10



1 2 34 56 7 89 10

"W --'wo lm 00



represents a breakdown product of B3 and B4 (Mitchell et

al., 1985) or a minor contaminant is unclear. The exclusion

of this 68kD polypeptide from the nucleus while B3 and B4

are rapidly taken up is consistent with the view that

translocation of these proteins is a transport-mediated

process. The inset in Figure 6A is an autoradiograph of a

two-dimensional gel which shows that equivalent amounts of

microinjected 1251-labeled B3 and B4 enter the nucleus. This

is consistent with the data obtained for endogenous B3 and

B4 (see Table 2).

Nuclear Uptake of 125I-labeled hsc70

Uptake experiments identical to those described for

B3/B4 were performed using hsc70 (Figure 5B). The nuclear

uptake rates of hsc70 were similar to those obtained for

B3/B4, and were approximately 24-fold greater than observed

for BSA. Figure 6B shows no significant breakdown of the

injected hsc70; greater than 95% of the labeled protein

remained intact throughout the course of the experiment as

determined by densitometric analysis.

Nuclear Efflux of 1"I-labeled B3 and B4

Nuclear efflux was measured by injecting 125I-labeled

protein directly into oocyte nuclei and measuring the amount

that subsequently entered the cytoplasm. Figure 7A and 7B

are distribution graphs, showing the percent of

Figure 7. Nuclear Efflux Kinetics of 125I-labeled BSA (A),
B3/B4 (B) and Hsc70 (C).

Single-cell analysis was performed, with the
number of cells indicated on the ordinates, and
percent of injected protein located in the
cytoplasmic fraction two hours after nuclear
injection indicated on the abscissa.

40 r--





0 ,- ,- -. -

Percent of Total in the Cytoplasm

I -

1 -


uTn. FA -F

VA. VA. -,wI. VA.


nuclear-injected BSA and B3/B4 found in the cytoplasm of

individual cells two hours post-injection.
125I-labeled BSA controls (Figure 7A) show that in

approximately 70% of the cells, less than ten percent of the

injected BSA entered the cytoplasm. Since nuclear uptake

and efflux of BSA occurs by passive diffusion (Bonner, 1975;

Paine, 1975; Paine et al., 1975; Lanford et al., 1986), it

is expected that only a few percent of the injected BSA

would passively diffuse from the nucleus in two hours.

Therefore, we considered the 0 to 10% distribution peak to

represent successful nuclear injections, and values greater

than 10% representative of partial or missed injections.

Results of nuclear injections of B3/B4 (Figure 7B) show

that in approximately 70% of the cells, 30 to 70% of the

labeled protein had entered the cytoplasm after two hours.

Values greater than 70% were considered partial or missed

nuclear injections. The means and standard deviations for

the amount of injected BSA and B3/B4 found in the cytoplasm

of the groups containing 70% of the cells were 6.21.1 and

465.7 percent, respectively. If the 50 to 60% values for

B3/B4 efflux are excluded from the calculations (i.e.,

considered to be unsuccessful injections), the lowest efflux

rates obtained for B3/B4 are still significantly greater

than the values obtained for BSA, indicating that B3/B4

leave the nucleus faster than predicted by passive



Figure 8 (A and B) shows autoradiographs of one-

dimensional gels from two separate nuclear injection

experiments. Figure 8A shows an aliquot of B3/B4 injected

nuclei from cells found within the 30-60% distribution peak

(lane 1), and an aliquot from ten pooled cytoplasms from the

same cells (lane 2). Figure 8B shows aliquots from

fractions identical to those in 8A (lanes 1 and 2), and, in

addition, an aliquot of ten pooled cytoplasms from cells in

which 70 to 90% of the label was cytoplasmic (lane 3). In

addition to the 70kD B3/B4 band, a lower molecular weight

polypeptide (approximately 68kD), analogous to that seen in

the nuclear uptake experiments, was present in the injected

nuclei (this is especially clear in Figure 8A) and

cytoplasmic aliquots from cells considered to represent

partial or missed nuclear injections. However, it was not

found in cytoplasmic fractions from successful nuclear

injections, supporting our earlier conclusion that it does

not cross the nuclear envelope. Since this protein is non-

exchangeable, it can serve as a marker for the site of

injection, and its distribution is consistent with our

interpretation of successful vs. partial or missed nuclear

injections. The proportion of the 68kD species, and its

migration pattern, was found to vary in different

experiments, probably the result of using different

preparations of B3/B4.

H $4

N 0

0H m

4 $4

0 4

C g
0 )

41 V
a 0
04 g

oM O








4 0
O i
$4 .

0 0
A a


Su 0

0 (a

0 >4 to
k u -,

O (0

00l 0
Hj *

3 44 4J M.
C 0 U
0) 0

0 V. *1

O 0

0 0 03

440 00
**o 0 3%,t


c >

o P-4
00 41

I4 U
4J 0 C ('
w 0

0 4 0
4) 02-


X 0 0'H

C + 0
0 U.- 00

*) C- 00 -


A ,'" : r,
,,"' ..-i -". "" ": '"'" "
, C O |.r.. .., ..:. ,.. .,
(0 .. -.. ., .

Nuclear Efflux of hsc70

Nuclear efflux of hsc70 was measured as described above

for B3 and B4. Results of nuclear injections of hsc70

(Figure 7C) show that in approximately 70% of the cells, 50

to 70% of the labeled protein had entered the cytoplasm

after two hours. Values greater than 70 percent were

considered bad injections. The mean and standard deviation

for the cytoplasmic concentration of hsc70 in the successful

experiments was 604.4 percent. The lowest efflux rates

obtained for hsc70 (50 to 60%) are still significantly

greater than the values obtained for BSA, demonstrating that

hsc70 also leaves the nucleus faster than expected for

passive diffusion.

The means of the efflux values from successful hsc70

and B3/B4 nuclear injections are significantly different.

Although the precise interpretation of this result is not

clear, it could reflect differences in signal efficiency, or

the interaction of B3/B4 with oocyte nuclear components.

Colloidal Gold Uptake and Temperature Effects

The intracellular distribution of B3/B4-coated gold

particles 30 minutes after cytoplasmic injection is

illustrated in Figure 9. Gold is seen associated with the

cytoplasmic surface of the nuclear pores and extending

through the centers of the pores. Particles over 120A in

diameter (including the coat material) can readily penetrate

Figure 9. The Intracellular Distribution of B3/B4-coated
Gold, 30 Minutes After a Cytoplasmic Injection.

Particles are present in the cytoplasm (C) and the
nucleus (N), and also adjacent to and within the
nuclear pores. Bar, 0.5pm.

0* i




'V 4

* .2

* 0.


the nuclear envelope. The size distribution of the

particles located in the nucleus and cytoplasm from 3 cells

are shown in Table 3.


Percentage of particles in each
Total no. size (diameter) class
of particles
measured 20-40 40-60 60-80 80-100 100-120A

Nucleus 792 11.4 40.0 33.8 11.9 2.0

Cytoplasm 797 10.8 42.0 34.5 9.3 2.7

Particle dimensions do not include the thickness of the
protein coat. The coat material adds approximately 30A to
the overall diameter (Dworetzky and Feldherr, 1988).

Chi-square analysis confirmed that the two populations are

not significantly different (p=0.9). The N/C particle ratio

is 0.18, approximately 30- and 20-fold greater than that

obtained for particles of similar size coated with PVP

(Feldherr et al., 1984), and BSA (Dworetzky et al., 1988),


Signal-mediated transport into the nucleus has been

shown to be a temperature-dependent process (eg., Newmeyer

et al., 1986a; Richardson et al., 1988). To determine the

effect of temperature on the nuclear uptake of B3/B4, the

intracellular distribution of microinjected 125I-labeled

B3/B4 was compared in cells maintained at 210C and 40C. The

percent incorporation of labeled B3/B4 into the nucleus


after 3 hours was 6.5% and 2.8%, respectively, which

represents a 2.3-fold reduction. For comparison, the

nuclear incorporation of a diffusible molecule, 125I-labeled

myoglobin (17.8 kD) (see Bonner, 1975b and Paine et al.,

1975), was also studied. 12.3% and 15% of the injected

myoglobin was present in the nucleoplasm after 3 hours at

21C and 40C, respectively. Although the effect of

temperature on B3/B4 transport is not as great as observed

for other nuclear proteins (i.e., nucleoplasmin, see

Newmeyer et al., 1986a), the results are significantly

different than those obtained for passive diffusion.

Hemi-Cell Fusion Experiments

The nuclear uptake of 12sI-B3/B4 following fusion of

labeled vegetal hemispheres from equilibrated cells with

unlabeled animal hemispheres was significantly greater than

that observed for 125I-BSA controls (5% vs. 1%; p<.001 by

Student's t-test). The higher than expected value for the

nuclear uptake of BSA 0.19% (expected) vs. 1.0% (observed),

reflects error inherent in the experimental procedure,

possibly the result of cytoplasmic contamination of the

isolated nuclei, or damage to the nuclear envelope during

the nuclear isolation procedure. These factors are more

difficult to control during oil isolation compared to

aqueous isolation procedures.

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd