nucleocytoplasmic protein traffic in single mammalian cells studied by fluorescence microphotolysis

Biochimica et Biophysica Acta 930 (1987) 419-431 419 Elsevier

BBA 12119

N u c l e o c y t o p l a s m i c p ro t e in t ra f f ic in s ingle m a m m a l i a n cells

s tudied by f l u o r e s c e n c e mic ropho to lys i s

B a r b a r a Schulz a n d R e i n e r Pe te r s

Max-Planck-lnstitut fftr Biophysik, Frankfurt (F.R.G.)

(Received 21 April 1987)

Key words: Protein transport; Fluorescence microphotolysis; (Nucleus)

Fluorescence microphotolysis was employed to measure in single living cells the kinetics of nucleocytoplasmic transport and the coefficients of intracellular diffusional mobility for the nuclear non-chromosomal protein nucleoplasmin. Nucleoplasmin was isolated from Xenopus ovary and labeled fluorescently. By injection into Xenopus oocytes it was ascertained that fluorescent labeling did not interfere with normal nuclear accumulation. Upon injection into the cytoplasm of various mammalian cell types nucleoplasmin was rapidly taken up by the nucleus. In rat hepatoma cells the half-time of nuclear uptake was approx. 5 rain at 37 o C; the nucleoeytoplasmic equilibrium concentration ratio had a maximum of 6.5 ± 1.4 and depended on the injected amount. Upon co-injection of ATPases or reduction of temperature to 10°C a nucleocytoplasmic equilization but no nuclear accumulation was observed. Equilization was fast (time constant 65 s at 23°C), similar to that of 10-kDa dextran permeating the nuclear envelope by simple diffusion through functional pores. Nucleoplasmin (160 kDa), however, is too large to permeate passively the nuclear envelope, which is apparent from the fact that its tryptic 'core' fragment (I00 kDa) could not permeate the nuclear envelope. On the other hand, a large fluorescent protein, phycoerythrin (240 kDa), was targeted to the nucleus by conjugation with nucleoplasmin. In the nucleus-to-cytoplasm direction the nuclear envelope was completely impermeable to nucleoplasmin, independently of temperature or ATP depletion. Nucleoplasmin, its core fragment, phycoerythrin and the phycoerythrin-nucleoplasmin conjugate were mobile in both cytoplasm and nucleus.

Introduction

Nucleocytoplasmic protein traffic has profound implications for both the structural organization

Abbreviations: HTC, hepatoma tissue culture; PMSF, phenyl- methylsulfonyl fluoride; SMPB, succinimidyl-4-(p-maleimidophenyl)butyrate; SPDP, succinimidyl-3-(pyridyldithio)pro- prionate; SDS, sodium dodecyl sulfate. FITC, fluorescein isothiocyanate; DMSO, dimethylsulfoxide; TRITC, tetramethylrhodamine isothiocyanate.

Correspondence: R. Peters, Max-Planck-lnstitut fiJr Biophysik, Kennedyallee 70, 6000 Frankfurt 70, F.R.G.

of the nucleus and the regulation of gene expres- sion (for a review see Refs 1-4). Although nucleocytoplasmic traffic, in general, is a complex process involving many parameters the central aspect is transport across the nuclear envelope. For exogenous compounds the nuclear envelope behaves like a molecular sieve with a functional pore radius of 45-60 A [5-9]. With regard to endogenous compounds the situation is more complex. In amphibian oocytes certain nuclear non-chromatin proteins exceeding in size that of functional pore radii rapidly permeate the nuclear envelope and accumulate in the nucleus after cytoplasmic injection [10-16]. Among those 'karyo-

0167-4889/87/$0350 © 1987 Elsevier Science Publishers B.V. (Biomedical Division)

420

philic' proteins, nucleoplasmin was studied in greatest detail. Nucleoplasmin is the most' abun- dant protein of the Xenopus oocyte nucleus and facilitates nucleosome assembly in vitro [12,13,15, 17-20]; for review, see Refs. 21 and 22.

It was suggested that nucleoplasmin and other karyophilic proteins carry a 'signal ' for nuclear location [11]. Meanwhile, the molecular nature of nuclear location signals has been disclosed, although in the first instance, not in nucleoplasmin but a rather different nuclear protein, the large tumor antigen (large T) of the Simian virus 40 [23,24]. Directed mutagenesis of the large T gene showed that the nuclear location signal consists of a sequence of seven or eight amino acid residues. The sequence is a permanent part of the mature protein not removed during nucleocytoplasmic transport. Furthermore, the sequence has a large functional autonomity [25]: It induces nuclear location when relocalized within large T or fused to prokaryotic or eukaryotic cytoplasmic proteins. Synthetic peptides containing the nuclear location sequence of large T can target proteins to the nucleus of Xenopus oocytes [26] and mammalian cells [27]. By now, nuclear location sequences have been described in a variety of nuclear proteins [28-33]. Homologous sequences were also found in nucleoplasmin [34] and in another karyophilic protein of amphibian oocytes [35]. However, the heterogeneity among these sequences is consider- able. The mechanism by which nuclear location sequences exert their function(s) is still unresolved.

In the past, quantitative studies of nucleocytoplasmic transport were almost completely confined to amphibian oocytes. This article employs fluorescence microphotolysis [4] to measure nucleocytoplasmic traffic of fluorescently labeled nucleoplasmin in single living somatic mammalian cells. The results confirm that the karyophilic nature of the amphibian protein is recognized by various mammalian cell types. Novel aspects are disclosed by the quantification of transport and molecular mobility. It appears that nuclear uptake in rat hepatoma cells consists of two separate processes: a fast permeation of the nuclear envelope and a rate-limiting accumulation in the nucleus. Some aspects of this study have been reported in preliminary form [36].

Materials and Methods

Purification, enzymatic digestion and fluorescent labeling of nucleoplasmin. Nucleoplasmin was isolated as described by Dingwall et al. [15]. In short, 40 ml of ovary from Xenopus laevis were homogenated. The homogenate was brought to 50 ml by addition of a solution containing 120 mM KC1, 2 mM MgC12 and 20 mM Tris-HCl (pH 7.5). This was centrifuged for 25 min in an SS34 rotor of a Sorval centrifuge at 3500 rpm. The cytosolic fraction was rescued and centrifuged for 30 min in a Ti50 rotor of a Beckman ultracentrifuge at 40 000 rpm. The cytosolic fraction was rescued and the high-speed centrifugation step was repeated. The cytosolic fraction was rescued and extracted twice with 1,1,2-trichlorofluoroethane. The clarified cy- tosol was applied to a DEAE-cellulose column (0.9 cm diameter, 10 cm length) equilibrated with buffer A (50 mM NaC1, 1 mM ethylendiamin- tetraacetate (EDTA), 0.1 mM phenylmethyl- sulfonyl fluoride (PMSF), 1 mM mercaptoethanol and 25 mM Tris-HC1, pH 7.5). The column was eluted with a linear NaC1 gradient (50-500 mM in buffer A). The eluted fractions were analyzed for nucleoplasmin by sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis according to Laskey et al. [12]. Fractions containing nucleoplasmin were pooled. Ammonium sulfate was added to a final concentration of 358 g/1. After 4 h at 4 ° C the sample was centrifuged. The super- natent was applied to a phenyl-Sepharose column (0.9 cm diameter, 10 cm length) equilibrated in buffer B (1.5 M ammonium sulfate and 20 mM Tris-HC1, pH 7.6). The column was eluted with a solution containing 20 mM Tris-HC1, pH 7.6. Fractions were analyzed for nucleoplasmin by SDS polyacrylamide gel electrophoresis. Fractions containing nucleoplasmin were pooled. Nucleoplas- rain was washed in an Amicon cell on a PM10 filter with 25 mM Tris-HC1 to give a final vol. of 4 ml and stored at - 7 0 ° C . Purification on a DEAE-cellulose column was repeated in the described manner.

In order to prepare the tryptic nucleoplasmin core [15], 250 /~g of purified nucleoplasmin was washed on Amicon PM10 filters with Tris-HC1 buffer (50 mM, pH 7.5), concentrated to 2 ml and incubated with 7.5 /~g trypsin (100 U / m g , Boeh-

ringer, Mannheim, F.R.G.) for 1 h at 33°C [44]. Digestion was terminated by addition of 96 ~g PMSF. The protein solution was washed with a 0.85% (w/v) solution of NaC1 in a centricon 10 cell and concentrated to 1 ml. On SDS polyacrylamide gel electrophoresis a single band of 80 kDa was observed.

Labeling with fluorescein isothiocyanate (FITC) or tetramethylrhodamine isothiocyanate (TRITC) was performed according to Goldmann [37]. 250 /~g of purified nucleoplasmin or core, contained in 1 ml 0.85% (w/v ) NaC1 solution, were mixed with 0.1 ml of a solution containing 4 m g / m l FITC (Sigma Chemie G m b H , Deisenhofen, F.R.G.) in 0.5 M carbonate buffer (pH 9.5). The reaction was terminated after 30 min (22 ° C) by titration to pH 7.0. The labeled protein was extensively washed on Amicon PM10 filters to remove free dye and then concentrated to 80 /~1 in a centricon 10 cell. The labeling ratio was estimated by quantitative gel electrophoresis. 2 /tl of a nucleoplasmin solution were subjected to a heat step (80 o C, 10 min) and then run on a 12% SDS-polyacrylamide gel according to Laemmli [38] together with FITC- labeled bovine serum albumin as standard. The gel was scanned for fluorescence, then stained with procion navy blue [39] and scanned for ab- sorption. From the scans the amounts of FITC and protein were estimated. The labeling ratio, calculated according to Goldman [37], was approx. 2 mol of FITC per mol of nucleoplasmin monomer. In the case of the nucleoplasmin core, the labeling ratio was substantially smaller. Un- bound FITC was less than 10% of total FITC.

Conjugation of nucleoplasmin with phycoerythrin Conjugation [40] involved three steps. (i) 250

/~g purified nucleoplasmin, dissolved in 100 ~1 of 0.1 M phosphate buffer (pH 7.5), were incubated with 3 ~l of a 5 mM solution of succinimidyl-4- (p-maleimidophenyl)butyrate (SMPB) in dimethylsulfoxide (DMSO). After 2 h at 22°C, the protein was washed in a centricon 10 cell with 0.1 M phosphate buffer (pH 7.5) and concentrated to 100 #l. (ii) 500/~g of R-phycoerythrin modified by reaction with succinimidyl 3-(pyridyldithio)pro- prionate (SPDP), obtained from Molecular Probes (Junction City, Oregon, U.S.A.), were dissolved in 0.5 ml and mixed with 17.5/~1 of a 0.5 M solution

421

of dithiothreitol. The reaction was allowed to pro- ceed for 15 min at 22°C. The modified phycoerythrin was then washed on Amicon XM 50 filters and concentrated to 100 /tl. (iii) The samples prepared in (i) and (ii) were mixed and incubated for 20 h at 4°C. The reaction was terminated by addition of 15 nmol of N-ethyl- maleimide. The sample was then applied to a Sephacryl S-300 column (1 cm diameter, 60 cm length) and eluted with a buffer containing 50 m M NaC1 and 50 mM phosphate buffer (pH 7.0) [41]. Three clearly separated fractions were obtained and shown by gel electrophoresis to contain nucleoplasmin, phycoerythrin and the nucleoplasmin-phycoerythrin conjugate. In the conjugate the molecular ratio of nucleoplasmin-to-phycoerythrin was about 1 : 1.

Injection of Xenopus oocytes and fluorescence measurements in oocyte sections

Pieces of Xenopus ovary were incubated for 4 h in a modified Barth solution to which collagenase (0.5 g / l ) had been added. Stage VI oocytes were injected with about 80 nl of a solution containing either FITC-labeled nucleoplasmin (0.35 mg/ml ) , the FITC-labeled tryptic core fragment of nucleoplasmin (0.50 mg/ml ) , or FITC-labeled bovine serum albumin (100 mg/ml) . Injected oocytes, together with non-injected control cells of the same ovary piece, were kept for 8±10 h at 19°C in a modified Barth solution'to which Ficoll (25 g / l ) had been added. Incubation was followed by freezing, sectioning and fixation. Extensive pre- cautions were taken to prevent redistribution of injected compounds during these procedures. Eggs were placed as a thin layer on a piece of cellulose filter paper which was shot into liquid propane. Quenched oocytes were kept in liquid nitrogen. In a kryomicrotom the frozen oocytes were sectioned at - 1 8 ° C. Sections (thickness 12 #m) were trans- ferred onto slides covered previously with a bovine serum albumin/glycerol solution and kept over night in the microtom at - 1 8 ° C under a para- formaldehyde atmosphere. Sections were trans- ferred into a dessicator and slowly warmed up to room temperature. In these sections the subcellu- lar distribution of injected compounds was studied by microfluorometry. The fluorescence microphotolysis apparatus was employed at low magnifica-

422

tion. The 488-nm line of the argon laser served for fluorescence excitation. A 3.2-fold objective lens was used to illuminate and measure circular areas of 0.2 mm diameter. Micrographs of the fluorescence distribution in oocyte sections were obtained with an image-intensifying video system (Hamamatsu, model 1966-12).

Cultivation of mammafian cells, microinjection and microscopic fluorescence measurements

Cultivation of cells, generation of polykaryons by treatment of monolayer cultures with poly(eth- ylene glycol), microinjection, and the measurement of intracellular mobility and nucleocytoplasmic transport by fluorescence microphotolysis were as described [8,9].

Results

Effects of fluorescent labeling on nuclear uptake of nucleoplasmin

It seemed important to study the effects of f luorescent labeling on nuclear uptake of nucleoplasmin. F ITC and T R I T C are amino-reac- tive chromophores which may couple, for instance, to lysine residues of putative nuclear location sequences [34] and, thus, may influence nuclear uptake. Therefore, we measured nuclear accumulation of FITC-labeled nucleoplasmin in

the homologous system, the Xenopus oocyte. In a typical experiment, about 25 ng of nucleoplasmin was injected into the cytoplasm of a stage VI Xenopus oocyte. After 8-10 h at 20 °C the oocyte was quenched in liquid propane and sectioned at - 2 0 ° C . In the sections, the fluorescence of nuclear and cytoplasmic areas was measured em- ploying the fluorescence microphotolysis apparatus with a low-power objective lens. The measurements were complicated by the fact that the ooplasm has a rather strong and quite variable autofluorescence. Therefore, with each batch of injected cells, non-injected control cells were prepared and treated exactly as the injected cells. In addition to nucleoplasmin the FITC-labeled tryptic core fragment of nucleoplasmin and FITC-labeled bovine serum albumin were studied in the same manner. Nucleoplasmin was found to be accumulated in the nucleus, whereas the core fragment and bovine serum albumin were re- stricted to the cytoplasm (Fig. 1). The nucleocytoplasmic fluorescence ratio, Fn/c, was about 16 for nucleoplasmin (range 3.9-31.8), 0.06 for the core fragment and 0.12 for bovine serum albumin (Table I).

The results suggest that fluorescent labeling does not noticeably interfere with nuclear uptake of nucleoplasmin. It may be important in this respect that the labeling ratio of our preparation

TABLE l

EQUILIBRIUM DISTRIBUTION OF FLUORESCENTLY LABELED NUCLEOPLASMIN AND BOVINE SERUM ALBUMIN AFTER INJECTION INTO THE CYTOPLASM OF XENOPUS OOCYTES

Values represent relative fluorescence intensity in counts per second as measured by the single-photon counting system of the fluorescence microphotolysis apparatus. The mean+_ S.D. is given with the number of measurements in parenthesis. Fn/c is the nucleocytoplasmic fluorescence ratio.

Nucleoplasmin Tryptic core fragment Bovine serum of nucleoplasmin albumin

Cytoplasm

Injected cells 6074_+ 2264 (39) 2706 +_ 1 689 (34) 34458 _+ 16415 (10 Controlcells 5833+3093 (9) 1023 + 476(13) 2153 + 305(19) Injected-controls 241_+3833 1683 +1755 32305 +16417

Nucleus

Injected cells 4456_+ 1 953 (39) 207 _+ 73 (34) 4 270 _+ 1490 (10) Control cells 605_+ 525 (9) 109 + 19 (13) 393 _+ 6 (19) Injected controls 3 851 _+ 2022 98 _+ 75 3 877 _+ 1 490

Fn/c 16+ 0.1 0.06_+ 0.96 0.12_+ 1.97

k

Fig. 1. Equilibrium distribution of fluorescently labeled nucleoplasmin and bovine serum albumin in Xenopus oocytes. Micrographs were obtained by means of an intensified video system. (a) Fluorescence micrography of a section from an oocyte which had been injected into the cytoplasm with FITC-

423

was rather low, namely about 2 mol of F I T C per mol of nucleoplasmin monomer (see Materials and Methods). Dingwall et al. [15] employed ra- diotracer and cell f ract ionat ion methods to s tudy nuclear uptake of nucleoplasmin in Xenopus oocytes. At incubat ion times and temperatures comparab le to those of the present s tudy nucleocytoplasmic concentra t ion ratios of nucleoplasmin were found to be relatively large, up to 100 in some cases. Below is shown that in rat hepa toma cells the degree of nuclear accumulat ion depended on the total injected amount , and seemed to be saturable. This might also hold for the Xenopus oocyte. Dingwall et al. [15] injected substantially smaller amounts of nucleoplasmin (1.3 ng per oocyte) than we did (25 ng per oocyte).

Nuclear accumulation of fluorescently labeled nucleoplasmin in various somatic mammalian cells

F1TC- or TRITC-labeled nucleoplasmin was in t roduced into cultured mammal ian cells by pres- sure microinjection. The following cell types were studied: hepa toma tissue culture (HTC) cells, p r imary rat hepatocytes, Vero cells, BSC cells, LLC-PK1 cells and T24 cells, a human lung carc inoma line. In all these cells, nucleoplasmin was observed to accumulate rapidly in the nucleus after cytoplasmic injection (microphotographs not shown). In H T C cells kept at 37 ° C accumulat ion was completed within 10 min after injection (for a sequence of micrographs showing the time course of accumulat ion see Fig. 2 in Ref. 36).

Concentration dependence of nuclear accumulation The kinetics of nuclear uptake was measured in

H T C polykaryons. After injection of FITC-labeled nucleoplasmin into the cytoplasm, a small circular photometr ic field was used to measure fluorescence alternately in a nuclear and a cytoplasmic, perinuclear area. The optical condit ions were such (40-fold objective lens with numerical aperture = 0.75) that the i l luminating beam traversed the cell approximately as cylinder and that fluorescence was collected f rom the total cellular cross section

labeled nucleoplasmin and incubated for 8-10 h at 19°C before cryofixation. The diameter of the nucleus is about 300 /~m. (b) Phase-contrast image of the same section. (c) Fluores- cence rnicrograph of an oocyte injected with FITC-labeled

bovine serum albumin and treated as the oocyte in (a).

424

[8]. In the following, the ratio of nuclear and perinuclear fluorescence signal is referred to as the nucleocytoplasmic fluorescence ratio, Fn/c. In order to judge nuclear uptake kinetics it is essen- tial to know the Fn/c ratio of both a small molecule which freely distributes in the whole cell and a large molecule which cannot permeate the nuclear envelope. FITC-labeled dextrans can be used for that purpose [9]. Fn/c of a small permeable dextran (10 kDa) was 1.0 _+ 0.1 (mean _+ S.D. of 11 measurements) for times of more than 10 min after cytoplasmic injection. This indicated that the height, h, of the cell, i.e., its vertical diameter, is the same in nuclear and perinuclear areas. Fn/c of a large impermeable dextran (157 kDa) was 0.4 _+ 0.1 (mean + S.D. of 22 measurements), independent of time after cytoplasmic injection. This value agrees with previous conclu- sions [9] that in the nuclear area the cytoplasmic layer accounts for 40% and the nucleus for 60% of cellular height. In order to relate fluorescence to concentration we assume that the efficiency of fluorescence excitation, emission and collection is about equal in cytoplasm and nucleus. Together with the preliminaries mentioned (constancy of cellular height, h, cylindrical beam) it follows that the nucleocytoplasmic concentration ratio (Cn/c) is given by:

C n / c = ( F n / c - 1 ) / H + 1 (1)

where H is the fraction of h taken by the nucleus. If H = 0.6, Eqn. 1 reduces to Cn/c = 1.7 Fn/c - 0.7.

Nuclear accumulation of nucleoplasmin depended on the injected amount. In a first series of experiments, a solution was injected which contained about 1 g / l of nucleoplasmin. From the injected volume (5-10% of the cell volume), we estimated roughly the amount of nucleoplasmin injected per nucleus of the polykaryons to be in the order of 50-100 fg. Under these conditions (Fig. 2, circles), Fn/c increased from an initial value of 0.4 to a maximum of 1.7 +_ 0.3 (mean _+ S.D. of 44 measurements); the half-time of accumulation was approx. 5 rain. In a second series of experiments, the nucleoplasmin solution was diluted 10-times to yield a concentration of 0.1 g / l and an estimated injected amount of 5-10 fg

Fn/c 60

50

."- 0-

30-

20-

10-

• A •

.e .. AA,,

2. L4 " "

I ' " " ":':" """ ":': ~'o 2'o ~ o ~o

I 5'0 6'0

tlm~nl

Fig. 2. Nuclear uptake kinetics of nucleoplasmin in HTC polykaryons. HTC polykaryons were injected into the cytoplasm with a solution of FITC-labeled nucleoplasmin. The nucleocytoplasmic fluorescence ratio, F n / c , is plotted versus time after injection. Concentration of nucleoplasmin in the

injectate was about 1 g / I (circles) or 0.1 g / l (triangles).

per nucleus. Under these condition (Fig. 2, triangles), the initial Fn/c and the half-time of accumulation were similar or identical as in the case of the higher concentration. The maximum Fn/c, however, was 4.2 _+ 0.9 (mean _+ S.D. of 14 measurements). According to Eqn. 1, the measured fluorescence ratios correspond to maximum nucleocytoplasmic concentration ratios of 2.2 +_ 0.5 (1 g / l ) and 6.5 _+ 1.4 (0.1 g/l) , respectively.

The fraction, X, of injected nucleoplasmin molecules accumulated in the nucleus at equilibrium (molecules in nucleus/molecules in cell) was roughly estimated from the following equa- tion:

X = C n / c / ( C n / c + 1 / V ) (2)

where V is the nucleocytoplasmic volume ratio. Assuming deliberately that V is about 1 / 4 in the case of HTC polykaryons, it follows that X is about 0.35 in case of the higher nucleoplasmin concentration (1 g / l ) and 0.62 in case of the lower nucleoplasmin concentration (0.1 g/ l) . Thus, at the higher concentration, about 17.5 35 fg of the injected 50-100 fg were accumulated in the nucleus. At the lower concentration, about 3.1-6.2 of the injected 5-10 fg were accumulated in the nucleus. This result suggests that nuclear accumulation of nucleoplasmin is saturable.

Dependence of nuclear uptake on temperature reduction and A TP depletion

The effects of temperature reduction and ATP depletion on nuclear uptake of FITC-labeled nucleoplasmin are shown in Fig. 3. In all cases, a concentrated nucleoplasmin solution (1 g / l ) was injected into the cytoplasm of HTC polykaryons. The experimental series comprised also control experiments with nucleoplasmin at 37 o C, with the

10 20 30 I '

Fn/c 2.0. ° o o o

o o o

% o o Oo~ o o

o goO %

1.0 ~o ° o

o

".'.. ""'o " 0

1.0 o o o o t3 o o

• . Oo •% O ~

I " ~ s A

• . ' : t • • • •

0 1

40 50 60 ,

0 0

o o o

o

o

0 0 0

A

Fn/c 1'0 t d , ,]

e t , B - , - J' ~ ; • , . *. * , ~ . % ° e

Oq 0 lo 2'0 io io 5'0 60

Time [mini Fig. 3. Effects of temperature, ATPases and tryptic dissection on the nuclear uptake kinetics of nucleoplasmin in HTC polykaryons. The nucleocytoplasmic fluorescence ratio, Fn/c, is plotted versus time after cytoplasmic injection. (a) At 37 o C, nucleoplasrnin (open circles) is rapidly taken up by the nucleus , whereas a large dextran of 157 kDa (full circles) is excluded from the nucleus. (b) At 10 o C, nucleoplasmm (open squares) rapidly equilizes between nucleus and cytoplasm but does not accumulate in the nucleus. For comparison, the data of the 157-kDa dextran from (a) are also plotted (full circles). (c) Upon co-injection of ATPases, nucleoplasmin (open triangles) rapidly equilizes between nucleus and cytoplasm but does not accumulate in the nucleus (37 o C). Full circles are data of the 157-kDa dextran for comparison. (d) The core fragment of nucleoplasmin (open triangles) is excluded from the nucleus (37°C). Full circles are data of the 157 kDa dextran for

comparison.

425

tryptic core fragment of nucleoplasmin, and with a large dextran (157 kDa) which cannot permeate the nuclear envelope. Fig. 3d shows that the tryptic nucleoplasmin core fragment cannot permeate the nuclear envelope. Fn/c of the core fragment was independent of time after cytoplasmic injection, it amounted to 0.43 _+ 0.10 (mean + S.D. of 30 measurements) and was not significantly different from that of the 157-kDa dextran. Injected into the nucleus the core fragment cannot permeate the nuclear envelope (data not shown).

At 10°C (Fig. 3b), Fn/c of nucleoplasmin increased within 10 min after cytoplasmic injection from 0.4 to an equilibrium value of 0.73 _+ 0.14 (mean + S.D. of 17 measurements). Similar rela- tions were observed at 37°C when the ATPase apyrase was added to the injectate at a very high concentration (0.3 U / # I ) and co-injected with nucleoplasmin (Fig. 3c). It has been shown recently by Newmeyer et al. [42] that apyrase effi- ciently depletes cells of ATP pools and prevents the accumulation of nucleoplasmin in nuclear envelope vesicles generated in a cell-free system. In HTC cells, Fn/c of nucleoplasmin (Fig. 3c) increased under these conditions within a few minutes from an initial value of 0.4 to an equilibrium value of 0.81 + 0.11 (mean + S.D. of 16 measurements).

It can be argued that the kinetics of ATP depletion by apyrase were not established and that it remained unresolved whether ATP depletion was much faster then nucleocytoplasmic transport. However, we have also measured the kinetics of equilization under equilibrium conditions (see Fig. 4 below), i.e., at times sufficiently long after injection of apyrase. In addition, we have also tried to employ other means for depleting cells from metabolic energy. HTC cells were kept for 48 h in a glucose-free medium, then fused to yield polykaryons and injected with FITC-labeled nucleoplasmin. In such cells nucleoplasmin could enter the nucleus but was not accumulated (Fn/c in all cases was less than 1.3).

The results illustrated in Fig. 3 opened the possibility that temperature reduction and ATPase co-injection prevented accumulation of nucleoplasmin in the nucleus, but not its permeation through the nuclear envelope and its equilization between cytoplasm and nucleus. In the present

4 2 6

context, equilization means that the equilibrium distribution of nucleoplasmin between cytoplasm and nucleus is only determined by accessible space and solubility in intracellular aqueous phases. In the case of equilization, the Fn/c value of nucleoplasmin is expected to be somewhat smaller than 1.0. Previous experiments [45| with isolated, leaky nuclei showed that the equilibrium distribution of dextrans between external medium and nuclear interior depended on molecular mass. Ap- parently, the nucleus interior is organized in form of a rather fine meshwork only partially accessible to large molecules such as nucleoplasmin (160 kDa).

Furthermore, Fig. 3b and c suggested that nuclear envelope permeat ion is a fast process not adequately resolved by Fn/c measurements. In order to investigate this possibility, the kinetics of equilization between nucleus and cytoplasm were measured at 2 2 ° C by fluorescence microphotolysis. Nucleoplasmin and apyrase were co-injected into the cytoplasm of H T C polykaryons. After an equilibration time of 10 rain, fluorescence was measured in the nuclear region using a large illuminated field which covered most of the nucleus. Then, the nucleus was depleted of fluorescence by an intensive laser flash and the influx from cytoplasm to the nucleus was followed by fluorescence measurement at the initial low beam power. A typical example of such measurements is shown in Fig. 4a. For comparison, a measurement relating to the nucleocytoplasmic flux of a small dextran which can freely permeate the nuclear envelope is displayed in Fig. 4b. An evaluation of the measurements in terms of the rate constant , k, of influx yielded 0.015 + 0 .06 / s for nucleoplasmin (mean_+ S.D. of 13 measurements) and 0.019_+ 0 .004 /s for the 10-kDa dextran (mean + S.D. of 18 measurements). This corresponds to relaxation times ( l / k ) of 66 s and 53 s, respectively.

Impermeability of the nuclear envelope to nucleoplasmin in the nucleus-to-cytoplasm direction

Nuclear injection was employed to study whether nucleoplasrnin can permeate the nuclear envelope in the direction from nucleus to cytoplasm. In this approach, FITC-labeled nucleoplasmin was injected into a nucleus of a H T C polykaryon and fluorescence measurements were

1.0

~o.8

~0.6

t.l-

0.2

1.0"

U

U ~0.8 0

0.6

Q

I I I I

I I 1 I

1 2 3 4 Time [min]

Fig. 4. Kinetics of nucleocytoplasmic equilization as measured by fluorescence mierophotolysis. (a) FITC-labeled nucleoplasmin was injected together with ATPases into the cytoplasm of HTC polykaryons kept at 22 ° C. After an equilibration time of more than 10 rain, a nucleus was illuminated with an attenuated laser beam (radius 8 /~m). Fluorescence was measured and denoted as 1.0. At t = 0.5 min, the beam power was increased by a factor 10000 for a short time (1/15-1/4 s) to deplete the nucleus of fluorescence. Fluorescence measurement was then continued at the initial low beam power and influx from cytoplasm to nucleus was monitored. The kinetics of fluorescence recovery yield the rate constant of influx. (b) The

analog experiment with a 10-kDa dextran.

performed on the injected nucleus, a different non-injected nucleus and a cytoplasmic region. Representat ive measurements are displayed in Fig. 5. Usually, quite large nucleocytoplasmic fluorescence ratios could be established in this manner. However, under none of the condit ions employed (37°C , 3 7 ° C with co-injection of apyrase, 2 2 ° C and 10 ° C) was a leakage of nucleoplasmin from nucleus into cytoplasm observed. Within the accu- racy and time scale of our experiments, no indica- tions for, e.g., the establishment of an nucleocytoplasmic equilibrium or the accumulat ion of nucleoplasmin in non-injected nuclei were obtained.

It is not self-evident that the nuclear envelope reseals after puncture by a micropipette. In oocytes, for instance, reflux from nucleus to cytoplasm has been described to occur under t he se condit ions [43]. However, the frequency of nuclear

0 5 1 e °°°°

0 i i ,

~o o o o o

(D

0 = Or , , L I0~ ...... sc

I -~ o5J n-

0~, e • l , , J

1.0Ion. d o o o o

0"5~ n a a *,

1'0 . . . . 0 0 20 30 &O 50 tlmin]

Fig. 5. Impermeability of the nuclear envelope to nucleoplasmin in the nucleus-to-cytoplasm direction. FITC-labeled nucleoplasmin was injected into nuclei of HTC polykaryons. Relative fluorescence of an injected nucleus (circles), a different nucleus (squares) of the same polykaryon, or the cytoplasm (triangle) is plotted versus time after injection. (a) 37 ° C, (b)

37 ° C with co-injection of ATPases, (c) 22 ° C, (d) 10 o C.

envelope resealing may depend on the cell type. In experiments with dextrans, we have previously observed that resealing is a frequent event in H T C cells and that a permanent leak in the nuclear envelope can be recognized by a fast nucleocytoplasmic equilization [9].

Nuclear targeting of a large chromoprotein, phycoerythrin, by conjugation with nucleoplasmin

Feldherr et al. [44] showed previously that col- loidal gold particles can be targeted to the nucleus of Xenopus oocytes by coating with nucleoplas- rain. In the case of mammal ian cells, Sugawa et al. [46] introduced nucleoplasmin and nucleoplasmin- protein conjugates into the cytoplasm by red-cell- mediated fusion and studied transport to the nucleus by cell fractionation methods. We chose a

427

more direct approach involving needle-injection and video-enhanced fluorescence microscopy. For that purpose, covalent conjugates of nucleoplasmin with the highly fluorescent chromoprotein, R-phycoerythrin, were prepared. Although the stoichiometry of the complex was not rigorously quantitated, the column elution profile and gel electrophoresis pattern suggested that only 1 or 2 molecules of nucleoplasmin were bound per molecule of phycoerythrin.

R-phycoerythrin (240 kDa), whether in its na- tive form or modified with pyridylsulfide, did not permeate the nuclear envelope. It remained in the cytoplasm (Fig. 6, e and f) or in the nucleus (not shown) after a corresponding injection. The nucleoplasmin-phycoerythrin conjugate, on the other hand, was able to permeate the nuclear envelope and to evenly distribute in the cell after cytoplasmic injection. In some, but not all, cases equilization was followed by nuclear accumulation in a manner similar to that of labeled nucleoplasmin (Fig. 6, a-d) . The kinetics of nuclear uptake were not determined. Video-enhanced fluorescence microscopy suggested, however, that nuclear uptake of the conjugate was considerably slower than that of FITC-labeled nucleoplasmin.

Translational mobility of nucleoplasmin, its core fragment, phycoerythrin and the nucleoplasmin- phycoerythrin conjugate

The intracellular mobility of all fluorescent species employed in this study was measured by fluorescence microphotolysis. Measurements were evaluated in terms of the mobile fraction, R, and the apparent translational diffusion coefficient, D. Results are collected in Table II. In both cytoplasm and nucleus, 70-90% of injected nucleoplasmin was mobile with an apparent diffusion coefficient of 2 -3 ~m2/s at 37°C. Intracellular mobility of nucleoplasmin was independent of the injected amount or the time after injection. The values given in Table II pertain to the situation of maximum nuclear accumulation but similar values hold for times immediately after injection. The tryptic core fragment of nucleoplasmin was also quite mobile in cytoplasm and nucleus although nuclear values are somewhat smaller than the corresponding values of nucleoplasmin. It may be noted that the diffusion coefficient of phycoeryth-

428

Fig. 6. Nuclear uptake of a nucleoplasmin-phycoerythrin conjugate in HTC polykaryons. (a-d) The covalent conjugate of nucleoplasmin and R-phycoerythrin was injected into the cytoplasm of a HTC polykaryon at 37 ° C. Fluorescence micrographs were taken at (a) 1 min, (b) 10 min, (c) 20 rain and (d) 30 min after injection. (e) R-phycoerythrin was injected into the cytoplasm of a HTC polykaryon at 37°C; the micrograph was taken 30 min after injection. (f) Pyridylsulfide-modified R-phycoerythrin, an intermediate in the conjugation of R-phycoerythrin and nucleoplasmin, was injected into the cytoplasm of a HTC polykaryon at

37 o C; the micrograph was taken 30 min after injection.

429

TABLE II

INTRACELLULAR MOBILITY OF FLUORESCENTLY LABELED N U C L E O P L A S M I N AND OTHER PROTEINS IN HTC CELLS

D, Apparent translational diffusion coefficient; R, mobile fraction. Values represent mean + S.D. of 13 20 measurements at 37 ° C.

Nucleoplasmin a Tryptic core Phycoerythrin Nucleoplasmin-

1 g/1 0.1 g/1 fragment phycoerythrin conjugate

Cytoplasm D ( # m2/s) 2.3 + 0.5 3.0 +__ 0.7 1.9 + 0.2 4.6 ± 1.1 5.0 ± 1.0 R(%) 90 ±5 78 _+8 79 ±6 93 ±7 94 ± 4 Nucleus D ( # m / / s ) 1.9_+0.8 2.1 +0.8 0.8±0.2 4.4± 1.3 4.8_+ 1.0 R(%) 89 ±7 69 ±8 61 ±9 93 ± 4 88 +10

a Concentration in the injected fluid.

rin (240 kDa) was about 2-fold larger than that of the smaller nucleoplasmin (160 kDa), an observation discussed below.

Discussion

In the present study, quantitative fluorescence microscopic methods were employed to measure the intracellular transport of the amphibian nuclear non-chromosomal protein nucleoplasmin in somatic mammalian cells. Nucleoplasmin was isolated from Xenopus oocytes and labeled fluorescently. Control experiments ascertained that the labeled protein had retained its normal ability to be accumulated in the nucleus. It was confirmed that many of the observat ions concerning nucleocytoplasmic transport of nucleoplasmin in other cellular systems also hold for somatic mammalian cells, at least on a qualitative level. The quantitation of nucleocytoplasmic transport kinetics and intracellular mobility suggested that nuclear uptake consists of two distinct processes, nuclear envelope permeation and nuclear accumulation.

Among those general aspects concerning the intracellular transport of nucleoplasmin which were shown to hold also for somatic mammalian cells are: the nuclear accumulation of nucleoplasmin after cytoplasmic injection, the sensitivity of nuclear accumulation to temperature reduction and ATP depletion, the inability of the tryptic nucleoplasmin core fragment to permeate the nuclear envelope, the targeting of large proteins to

the nucleus by conjugation with nucleoplasmin, the absence of tight binding to cellular structures in the case of nucleoplasmin and its core fragment. However, certain quantitative differences do exist among different systems. For instance, in Xenopus oocytes, nuclear accumulation of nucleoplasmin takes several hours for completion [15]. In HTC cells, a steady state is established within 10 min after injection. This difference may be simply a matter of nuclear dimensions. The influx rate constant, k, of a spherical vesicle with radius r is related to the permeability coefficient P by k = (3/r)P. Thus, for a given permeability coefficient the rate constant is inversely proportional to the vesicle radius. The nuclear radii of Xenopus oocytes and HTC cells differ by a factor of 100 (e.g., 300/~m/3/~m) which approximately accounts for the differences in accumulation times.

Among those aspects revealed by microfluoro- metric measurements are: the quantitation of intracellular mobility in terms of effective diffusion coefficients and mobile fractions, the measurement of nuclear uptake kinetics as a function of concentration, temperature and ATP depletion, the discrimination between nucleocytoplasmic equilization and nuclear accumulation and the measurement of equilization kinetics.

The measured effective intracellular diffusion coefficients deserve some attention. In the electron microscope the nucleoplasmin pentamer was visualized as a disc of 75 A in diameter [18] suggesting a rather compact molecular organization. A molecule of that size and shape is expected

430

to have a diffusion coefficient in water of approx. 50 /xm2/s. In HTC cells, the apparent diffusion coefficient of nucleoplasmin was found to be approx. 2 /~m2/s at 37°C. We have previously studied the mobility of dextrans in HTC cells [9]. It seems that intracellular diffusion of dextrans is largely determined by viscosity and, to a smaller extent, by steric restriction. Binding of dextrans to structural elements may be negligible. Among the dextrans studied the 40-kDa species had a Stokes' radius of 47 A which is somewhat larger than that of the nucleoplasmin pentamer. The diffusion coefficient of the 40-kDa dextran was 4.5 /.tmZ/s in the cytoplasm and 8.0/~m2/s in the nucleus at 37°C; i.e., substantially larger than that of nucleoplasmin. Similarly, the intracellular diffusion coefficient of R-phycoerythrin (240 kDa) was approx. 4.5 #m2/s . This suggests that the intracellular mobility of nucleoplasmin cannot be ex- plained by viscosity effects and steric restriction alone. In addition, other processes have to be invoked, e.g., the formation of rather large homo- or heteropolymers a n d / o r weak interaction with structural elements.

The molecular mechanisms by which karyophilic proteins are taken up by the nucleus are still unresolved. Based on a literature survey and on our own studies we have recently suggested [4] a scheme to account, on a more general level, for the currently known facts. The scheme, referred to as the signal modulation model, assumes that nucleoplasmin contains a signal for nuclear envelope permeation. If the signal is in an exposed, accessible conformation, it can bind to the pore complex. Binding triggers a conformational change in the pore complex which permits permeation of nucleoplasmin and its release on the opposite surface. This mechanism facilitates permeation. It does not necessarily require metabolic energy and does not impose a directionality on the transport process. Accumulation in the nucleus is thought to depend on a separate energy-requiring process localized in the nucleus, e.g., a chemical modification of nucleoplasmin which renders the signal inacessible and traps nucleoplasmin in the nucleus. The results of the present study are compatible with the signal modulation model. The observed dependence of nuclear uptake on concentration, temperature and ATP depletion supports the idea

that nuclear envelope permeation and nuclear accumulation are distinct processes. Also, it may be emphasized that the kinetics of nuclear envelope permeation were directly measured, independently of accumulation.

The signal modulation model assumes that nucleoplasmin occurs in two variants, only one of which can permeate the nuclear envelope. In Xenopus ocytes, nucleoplasmin is highly concentrated in the nucleus. Nucleoplasmin from this cell type should, therefore, represent, for the most part, the impermeable variant. This assumption agrees with the observation that nucleoplasmin was strictly confined to the nucleus after nuclear injection. However, the assumption seems to con- tradict all those observations showing that nucleoplasmin is taken up by nuclei after cytoplasmic injection. The model could only account for that observation if the cytoplasm harbored a mechanism which is able to revert the postulated intra-nuclear modification and to turn the impermeable into the permeable variant. This prediction provides the basis for a critical test of the signal modulation model yet to be performed.

Acknowledgements

Dr. W. Haase (Max-Planck-Institut fiir Bio- physik, Frankfurt, F.R.G.) was of great help in the histological part of the work and allowed us to use his cryomicrotom. We are indebted to Dr. W.W. Franke (Deutsches Krebsforschungszentrum, Hei- delberg, F.R.G.) for drawing our attention to karyophilic proteins. Dr. J.A. Kleinschmidt (Deutsches Krebsforschungszentrum, Heidelberg, F.R.G.) kindly discussed with us the isolation of nucleoplasmin. Drs. M. Trendelenburg and H. Steinbeisser (Deutsches Krebsforschungszentrum, Heidelberg, F.R.G.) taught us how to inject Xeno- pus oocytes and suggested how to prepare sections. Support by the Deutsche Forschungsge- meinschaft is gratefully acknowledged.

References

1 Franke, W.W. and Scheer, U. (1974) in The Cell Nucleus (Bush, H., ed.), Vol. 1, pp. 219-347, Academic Press, New York

2 Bonner, W.M. (1978) in The Cell Nucleus (Bush, H., ed.) Vol. 7, pp. 97-148, Academic Press, New York

3 De Robertis, E.M. (1983) Cell 32, 1021-1025 4 Peters, R. (1986) Biochim. Biophys. Acta 864, 305-359 5 Feldherr, C.M. and Feldherr, A.B. (1960) Nature (London)

185, 250-251 6 Paine, P.L. and Feldherr, C.M. (1972) Exptl. Cell Res. 74,

81-98 7 Paine, P.L., Moore, L.C. and Horowitz, S.B. (1975) Nature

(London) 254, 109-114 8 Peters, R. (1984) EMBO J. 3, 1831-1836 9 Lang, I., Scholz, M. and Peters, R. (1986) J. Cell Biol. 102,

1183-1190 10 Bonner, W.M. (1975) J. Cell Biol. 64, 431-437 11 De Robertis, E.M., Longthorne, R.F. and Gurdon, J.B.

(1978) Nature (London) 272, 254-256 12 Laskey, R.A., Mills, A.D. and Morris, N.R. (1977) Cell 10,

237-243 13 Mills, A.D., Laskey, R.A., Black, P. and De Robertis, E.M.

(1980) J. Mol. Biol. 139, 561-568 14 Dabauvalle, M.-C. and Franke, W.W. (1982) Proc. Natl.

Acad. Sci. USA 79, 5302-5306 15 Dingwall, C., Sharnick, S.V. and Laskey, R.A. (1982) Cell

30, 449-458 16 Feldherr, C.M., Cohen, R.J. and Ogburn, J.A. (1983) J. Cell

Biol. 96, 1486-1490 17 Laskey, R.A., Honda, B.M., Mills, A.D. and Finch, J.T.

(1978) Nature (London) 275, 416-420 18 Earnshaw, W.C., Honda, B.M. and Laskey, R.A. (1980)

Cell 21, 373-383 19 Krohne, G. and Franke, W.W. (1980) Proc. Natl. Acad. Sci.

USA 77, 1034-1038 20 Krohne, G. and Franke, W.W. (1980) Exptl. Cell Res. 129,

167-189 21 Laskey, R.A. and Earnshaw, W.C. (1980) Nature (London)

286, 763-767 22 Dingwall, C. (1985) Trends Biochem. Sci. 6, 64-66 23 Lanford, R.E. and Butel, J.S. (1984) Cell 37, 801-813 24 Kalderon, D., Richardson, W.D., Markham, A.F. and

Smith, A.E. (1984) Nature (London) 311, 33-38 25 Smith, A.E., Kalderon, D., Roberts, B.L., College, W.H.,

Edge, M., Gillett, P., Markham, A., Paucha, E. and Richardson, W.D. (1985) Proc. R. Soc. Lond. B 226, 43-58

26 Goldfarb, D.S., Gariepy, J., Schoolnik, G. and Kornberg, R.D. (1986) Nature (London) 322, 641-644

431

27 Lanford, R.E., Kanda, P. and Kennedy, R.C. (1986) Cell 46, 575-582

28 Hall, M.N., Hereford, L. and Herskowitz, I. (1984) Cell 36, 1057-1065

29 Silver, P.A., Keegan, L.P.°and Ptashne, M. (1984) Proc. Natl. Acad. Sci. USA 81, 5951-5955

30 Davey, J., Dimmock, N.J. and Colman, A. (1985) Cell 40, 667-675

31 Moreland, R.B., Nam, H.G., Hereford, L.M. and Fried, H.M. 1985) Proc. Natl. acad. Sci. USA 82, 6561-6565

32 Richardson, W.D., Roberts, B.L. and Smith, A.E. (1986) Cell 44, 77-85

33 Wychowski, C., Benichou, D. and Girard, M. (1986) EMBO J. 5, 2569-2576

34 Dingwall, C., Dilworth, S.M., Black, S.J. Kearsey, S.E., Cox, L.S. and Laskey, R.A. (1986) EMBO J. 6, 69-74

35 Kleinschmidt, J.A., Dingwall, C., Maier, G. and Franke, W.W. (1986) EMBO J. 5, 3547-3552

36 Schulz, B. and Peters, R (1986) in Nucleocytoplasmic Transport (Peters, R. and Trendelenburg, M., eds.), pp. 171-184, Springer-Verlag, Heidelberg

37 Goldman, M. (1968) Fluorescent Antibody Methods, pp. 1-303, Academic Press, New York

38 Laemmli, U.K. (1970) Nature (London) 227, 680-685 39 Smith, B.J. (1984) in Methods in Molecular Biology (Walker,

J.M., eds.), Vol. 1, pp. 119-127, Humana Press, Clifton 40 Oi, V.T., Glazer, A.N. and Stryer, L. (1982) J. Cell Biol. 93,

981-995

41 Kronick, M.N. and Grossman, P.D. (1983) Clin. Chem. 29, 1582-1586

42 Newmeyer, D.D., Lucocq, J.M., BiJrglin, T.R. and De Robertis, E.M. (1986) EMBO J. 5, 501-510

43 Wyllie, A.H., Gurdon, J.B. and Price, J. (1977) Nature (London) 268, 150-152

44 Feldherr, C.M., Kallenbach, E. and Schultz, N. (1984) J. Cell Biol. 99, 2216-2222

45 Lang, I. and Peters, R. (1984) in Information and Energy Transduction in Biological Membranes (Bolis, C.L., Helmreich, E.J.M. and Passow, H., eds.), pp. 377-385, Allan R. Liss, New York

46 Sugawa, H., Uchida, T., Yoneda, Y., Ishiura, M. and Okada, Y. (1985) Exptl. Cell Res. 159, 410-418

nucleocytoplasmic protein traffic in single mammalian cells studied by fluorescence microphotolysis

Documents