Mechanism and Rate of Permeation of Cells by Polycyclic Aromatic Hydrocarbons*

The principal mechanism of cellular uptake of benzo(a)pyrene and other polycyclic aromatic hydrocarbons (PAH) from lipoproteins into cells is sponta- neous transfer through the aqueous phase (Plant, L., Benson, D. M., and Smith, L. C. (1985) J. Cell Biol. 100, 1295-1308). Cellular uptake of benzo(a)pyrene from low density lipoproteins followed first-order ki- netics with a rate constant that was independent of the relative lipoprotein concentrations or cell number but which was 2 orders of magnitude smaller than the rate constant for benzo(a)pyrene desorption from low den- sity lipoproteins. Moreover, identical rate constants for cellular uptake of benzo(a)pyrene were observed when the donor vehicle was high density lipoproteins, very low density lipoproteins, or single bilayer phosphatidylcholine vesicles, even though rate constants for benzo(a)pyrene transfer from these donor vehicles differed by 10-fold. When phosphatidylcholine vesicles containing benzo(a)pyrene and a nontransferable flu- orescence quencher were mixed with cells in a stopped-flow system, two kinetic components were distin- guished: a fast component with a rate constant corre-sponding to that measured for transfer of benzo(a)- pyrene out of vesicles, followed by a much slower component, with a time course approximating that measured for cellular accumulation of benzo(a)pyrene by other techniques. Rate constants for desorption of a

mechanism of cellular uptake of benzo(a)pyrene and other polycyclic aromatic hydrocarbons (PAH) from lipoproteins into cells is spontaneous transfer through the aqueous phase (Plant, A. L., Benson, D. M.,  J. Cell Biol. 100,[1295][1296][1297][1298][1299][1300][1301][1302][1303][1304][1305][1306][1307][1308]. Cellular uptake of benzo(a)pyrene from low density lipoproteins followed first-order kinetics with a rate constant that was independent of the relative lipoprotein concentrations or cell number but which was 2 orders of magnitude smaller than the rate constant for benzo(a)pyrene desorption from low density lipoproteins. Moreover, identical rate constants for cellular uptake of benzo(a)pyrene were observed when the donor vehicle was high density lipoproteins, very low density lipoproteins, or single bilayer phosphatidylcholine vesicles, even though rate constants for benzo(a)pyrene transfer from these donor vehicles differed by 10-fold. When phosphatidylcholine vesicles containing benzo(a)pyrene and a nontransferable fluorescence quencher were mixed with cells in a stoppedflow system, two kinetic components were distinguished: a fast component with a rate constant corresponding to that measured for transfer of benzo(a)pyrene out of vesicles, followed by a much slower component, with a time course approximating that measured for cellular accumulation of benzo(a)pyrene by other techniques. Rate constants for desorption of a series of PAH which contained different number of aromatic rings from phosphatidylcholine vesicles differed over a 70-fold range. First-order rate constants for cell uptake of benzo(a)pyrene and five other PAH of different molecular sizes had the same 70-fold range of values, but were 2 orders of magnitude smaller than their respective rate constants for desorption from single bilayer vesicles. In addition, activation energies for cell uptake were essentially identical to the respective activation energies for desorption of PAH from phosphatidylcholine vesicles, confirming the mechanistic similarity of the two processes.
The passive uptake of molecules by cells has been investigated extensively, with most studies concerned with the transport of water-soluble non-electrolytes (Stein, 1967). The ratelimiting step in uptake of water-soluble substances by cells involves the absorption of the hydrophilic molecule by the external leaflet of the plasma membrane bilayer and diffusion * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $Present address: Chemistry Bldg., A361, National Bureau of Standards, Gaithersburg, MD 20899. Anne L. Plant$, Roger D. Knapp, and Louis C. Smith From the DeDartments of Biochemistrv. Cell Bioloev. and Medicine, Baylor College of Medicine and The Methodist Hospital, through the membrane (Davison and Danielli, 1943). Increased hydrophobicity of molecules has been correlated with more rapid rates of uptake (Collander, 1954;Smulders and Wright, 1971;Wright and Pietras, 1974;Bindslev and Wright, 1976). Whereas this correlation has been clearly demonstrated experimentally and is well founded theoretically , the range of substances previously examined has not included extremely hydrophobic molecules. Collander (1954) tested compounds with oi1:water partition coefficients between 0.00003 and 0.3. Klocke et al. (1972) studied the rates of uptake of the 1-5-carbon fatty acids with ether:water partition coefficients between 0.1 and 100. Giorgi and Stein (1981) examined uptake rates for steroids with octano1:water partition coefficients as high as 100. Passive uptake of extremely hydrophobic fatty acids by cells has also been studied (Sallee and Dietschy, 1973;Sherrill and Dietschy, 1975). These studies have suggested that cellular uptake of large, moderately lipophilic solutes all encounter the same rate-limiting barrier at the plasma membrane.
Polycyclic aromatic hydrocarbons (PAH)' are common environmental pollutants. The mechanism by which PAH enter cells and are distributed intracellularly has considerable importance since many of them are potent carcinogens. Benzo(a)pyrene, like other PAH, is very poorly soluble in water, -lo-' M (Mackay and Shiu, 1977), and has an octanol:H,O partition coefficient of 5,000,000 (Mackay et al. 1980). In plasma, benzo(a)pyrene partitions readily into plasma lipoproteins (Avigan, 1959;Shu and Nicols, 1979). If the inferred relationship between hydrophobicity and uptake rate holds for these compounds, then their rate of entry into cells should be extremely rapid.
The entry of benzo(a)pyrene into cells from plasma lipoproteins has been shown to be a spontaneous, nonmediated transfer process (Remsen and Shireman, 1981;Plant et al. 1985). Benzo(a)pyrene uptake by cells does not involve lipoprotein endocytosis since uptake is independent of the class of lipoproteins with which benzo(a)pyrene is originally associated and occurs a t identical rates with either live or HCHOfixed cells . Entry of benzo(a)pyrene into living cells from low density lipoproteins (LDL) occurs equally well in the absence as in the presence of receptors for lipoproteins (Remsen and Shireman, 1981).
The mechanism for equilibration of many of the noncovalent lipoprotein components among the various lipoprotein classes is spontaneous transfer through the aqueous phase Duckwitz-Peterlein and Moraal, 1978). This process involves desorption of a lipophil from the hydrophobic surface, diffusion through the aqueous phase, and readsorption into another hydrophobic compartment. Spontaneous transfer of lipophils among lipoproteins, model lipoproteins, and phospholipid vesicles has been studied extensively. The variables have been characteristics of the transferring species (Doody et al., 1980;Massey et al., 1982;Pownall et al., 1983;Plant et al., 1983;Phillips et al., 1980) and characteristics of the donor vehicle, such as lipid composition (Damen and Scherphof, 1981;Pownall et al., 1983) and vehicle size (Charlton et at., 1978;Charlton and Smith, 1982;Almgren,1980;Smith and Doody, 1981). Desorption of the molecule from a hydrophobic environment for dissolution in the aqueous phase is the rate-limiting step, whereas diffusion of the molecule and readsorption are extremely rapid.
Benzo(a)pyrene and other PAH transfer between phosphatidylcholine single bilayer vesicles and between LDL with half-times of several hundred milliseconds Plant et al., 1983). By contrast, the rate constant for benzo(a)pyrene entry into cells is about 200-fold slower than the rate constant for desorption from LDL . Since the rate of uptake of benzo(a)pyrene by cells is not limited by the rate of desorption from extracellular donor lipoproteins or vesicles, it is of interest to identify the ratelimiting step in cellular uptake.

MATERIALS AND METHODS
Digital Imaging Fluorescence Microscopy-The digital imaging fluorescence microscopy system has been described . Briefly, the system consists of a Leitz Divert fluorescence microscope with a mercury arc excitation source, a Hammamatsu C-1000 silicon target video camera, a Grinnell274 image processor, and an LSI 11/ 23 minicomputer (Lab Datex) with 256K RAM and a 40 Mbyte hard disc. Living cells were maintained on the microscope stage in a Bionique chamber (Corning, Lake Placid, NY) at 37 "C in a 5% CO, atmosphere and were monitored for viability by time-lapse video recording of phase-dense lysosome movement (Willingham and Pasbenzo(a)pyrene-labeled LDL or HDL was added to cells on the stage tan, 1978). To initate a kinetic experiment, medium containing through ports in the chamber lid. Benzo(a)pyrene fluorescence intensity was quantified during 1-s exposures to excitation light with a 10nm band-pass filter centered at 365 nm. Emitted light was observed with a filter that transmitted wavelengths greater than 400 nm. For each exposure, fluorescence intensity of each 0.065-pm2 pixel in the field was digitized and assigned a value between 0 and 255. Neutral density filters were used to attenuate excitation light so that photobleaching was eliminated . Cells treated with 2% HCHO were examined by digital fluorescence microscopy in a similar manner. Under the experimental conditions employed, only intracellular fluorescence was a significant contributor to the total fluorescence intensity of the images.
Cell Culture-Normal human fibroblasts and P388D1 murine macrophages were cultured as previously described . Fibroblasts were grown on 4.4-cm2 No.1 glass coverslips in tissue culture dishes and transferred to the Bionique chamber for microscopy experiments. Other kinetic experiments were performed with suspensions of P388D1 murine macrophages which were cultured in RPMI 1640 medium and 10% fetal calf serum. P388DI macrophages in dishes were suspended by rinsing with cold medium and collected by centrifugation. These cultures have been previously shown not to metabolize benzo(a)pyrene during these experiments .
Filter Kinetic Assays-For efflux experiments, cells were first incubated 20 min with [G3-H]benzo(a)pyrene-labeled LDL. Cells were separated from the benzo(a)pyrene-labeled LDL by centrifugation, washed once in serum-free medium, and resuspended. Efflux was initiated by addition of unlabeled LDL. To initiate cellular uptake, [G-3H]benzo(a)pyrene was added to 2 ml of RPMI 1640 medium or standard buffer containing 2 X lo6 P388D1 macrophages. Cell suspensions were stirred continually with a magnetic bar. Aliquots of the suspensions were removed at various times, added to 5 ml of standard buffer containing 5 mg ml" bovine serum albumin, and rapidly filtered under reduced pressure through glass-fiber filters (Whatman). Cells and associated benzo(a)pyrene were trapped by the filters, which were counted in scintillation fluid without further processing. For filtration experiments using aqueous solutions of benzo(a)pyrene, bovine serum albumin was omitted from the filter buffer solution.
Fluorometric Kinetic Assays-P388D1 cells were suspended at a cell density of 1-3 X lo6 cells ml" in either RPMI 1640 medium or standard buffer and stirred continuously in a temperature-regulated cuvette holder. Reactions were initiated by the addition of POPC vesicles containing 0.3 mol % PAH and 3 mol % DNP-DOA. The increase in fluorescence intensity, which was the result of PAH transfer from quenched vesicles to cells, was measured in an SLM 8000 spectrofluorometer, using the following excitation and emission wavelength settings, respectively, benzoperylene, 336 and 408 nm; 1,2,3,4-dibenzanthracene, 300 and 400 nm; benzopyrene, 365 and 406 nm; 9-phenylanthracene, 387 and 420 nm; 3,4-benzophenanthrene, 318 and 400 nm; and pyrene, 338 and 395 nm.
Stopped-flow Experiments-Lipoproteins or vesicles containing PAH were mixed rapidly with vesicles containing DNP-DOA in a Durrum-Gibson stopped-flow instrument described previously (Plant et al., 1983). Benzo(a)pyrene-labeled LDL (0.025 mg of protein ml-'), VLDL (0.05 mg of protein ml-'), and HDL (0.48 mg of protein m1-l) were mixed with 2 mM POPC containing 60 p M DNP-DOA. Other PAH, 6 p~ in 0.2 mM POPC vesicles, were mixed with 2 mM POPC vesicles containing quencher. P388D1 cells (2 X lo6 cells ml-I) were mixed with 0.2 mM POPC vesicles containing 6 PM benzo(a)pyrene and 60 PM DNP-DOA. Mixing time of the instrument was about 4 ms. Fluorescence intensities measured at millisecond time intervals were stored in 2048 channels in a Biomation 805 waveform recorder. For all kinetic assays, cells were examined by microscopy after the reactions. There was no observable cell debris, and the cells appeared to be structurally intact after mixing.
Kinetic Data Analysis-Stopped-flow data were analyzed on-line with parallel interface between an Apple I1 Plus microcomputer and the Biomation waveform recorder. All kinetic data were analyzed by nonlinear least squares regression with respect to a monoexponential function (Wentworth, 1965). The analytical program involved a reiterative procedure and stringent criteria for convergence. Results of the analysis were evaluated by comparative plots of data points and the fitted function. In addition, a rigorous statistical analysis was performed for every reaction, including standard deviation of data points from the fitted line, standard deviations for rate constant, initial intensity and final intensity values, and covariance of errors in these parameters. Kinetic data which were not collected by stopped-flow methods were entered from the keyboard and analyzed with the same program.

RESULTS
Rates of Benzo(a)pyrene Desorption from Extracellular Donors-The rate constant for the entry of benzo(a)pyrene into cells from LDL was measured by three independent experimental methods  and was about 100 times smaller than the rate constant for desorption of benzo(a)pyrene from LDL, which was determined by stopped-flow methods (Table I). Furthermore, only one rate constant for cell uptake of benzo(a)pyrene was obtained, even when the diameter of the donor vehicle was different. Rate constants

Rate constants for transfer of benzo(a)pyrene into cells from lipoproteins and vesicles
The rate constants for benzo(a)pyrene transfer from lipoproteins (LP) and POPC vesicles to POPC vesicles were measured by stoppedflow fluorometry. The rate constants for cellular uptake of benzo(a)pyrene were measured by digital fluorescence imaging microscopy and fluorometry, and isotopically with [G-3H]benzo-(a)pyrene .   for benzo(a)pyrene transfer from HDL, LDL, VLDL, and POPC vesicles as donors to identical acceptors differed as much as 10-fold. The slower interparticle transfer was a t least 50-fold faster than cellular uptake (Table I).
Hypotheses to Be Tested-The lack of a relationship of the rate constants for benzo(a)pyrene uptake by cells and the rate constants for transfer of benzo(a)pyrene from donor vehicles clearly excluded desorption of benzo(a)pyrene from the extracellular hydrophobic carrier as the rate-limiting step. Subsequent experiments were designed to eliminate five other possible explanations which could account for the observed kinetics of benzo(a)pyrene entry into cells, as follows. 1) The observed kinetics might involve a collisional event between donor vehicles and cells. 2) The observed kinetics might be dominated by transfer processes between the extracellular donors to give an observed rate constant that is smaller than the actual rate constant for cellular uptake.
3) The observed kinetics would be restricted by boundary water layers associated with the cell membrane surface such that diffusion of LDL to the vicinity of the cell membrane was rate-limiting. 4) The rate-limiting kinetic process could be diffusion of benzo(a)pyrene from the outer to the inner leaflet of the plasma membrane. 5 ) The observed kinetics could be determined by the desorption of benzo(a)pyrene from the inner leaflet of the plasma membrane to the interior of the cells.
Elimination of Transfer Involving Collisional Complexes-If a collisional complex were formed between the extracellular donor and the cell, cellular uptake of benzo(a)pyrene would be pseudo-first-order and not a true first-order reaction. To examine this possibility, rate constants for benzo(a)pyrene uptake into cells were measured over 4 orders of magnitude of [G-3H]benzo(a)pyrene-labeled LDL concentration (Table  11). Identical rate constants were obtained under all conditions. The results of nonlinear regression analysis of these data are shown on the same scale in Fig. 1

FIG. 1. Uptake of benzo(a)pyrene by cells in the presence of increasing concentrations of [G-SH]benzo(a)pyrene-labeled
LDL. Suspensions (2 X lo6 P388D1 macrophages) were stirred constantly at 20 "C in 2 ml of buffer containing 40 mM Tris, 0.15 M NaCl, and 0.3 mM EDTA, pH 7.4. Small volumes of LDL containing 0.2 pg of benzo(a)pyrene/mg of LDL protein were added to final LDL concentrations of 0.5 (O), 5 (0), 50 (A), or 500 (A) pg m1-l. As described under "Materials and Methods," aliquots were taken at the times indicated and rapidly filtered to trap cells on glass-fiber filters, which wme counted for [G-3H]benzo(a)pyrene. Solid lines represent analysis of the data with respect to a monoexponential function.
does not require a collision event between the donor vehicle and the cell surface.
Luck of Influence of the Concentration of Extracellular Donors-Rate constants for spontaneous transfer depend on whether or not the donor and acceptor vehicles are identical. In heterologous systems in which the donor and acceptor vehicles are different, the rate constant is influenced by the relative numbers of the donor and acceptor vehicles if rate constants for desorption for both vehicles are similar (Almgren, 1980;Bojesen, 1982). Thus, the observed rate constant should contain contributions from both the forward transfer involving donor X and the reverse transfer reaction involving donor Y according to the following expression.
Acceptable accuracy in the measurements depends on the quantity of benzo(a)pyrene taken up by the cells. Consequently, the number of donor LDL was always several orders of magnitude greater than the number of cells. Even at the lowest concentration of LDL used, 0.5 pg of LDL protein ml-', there were 6.9 x 10" LDL particles, compared to 1 x lo6 cells. A more suitable basis for comparison is that of surface areas; the ratio of LDL to cell surface area ranged from 0.034 to 34 from the lowest to highest LDL concentration. For this calculation, LDL and macrophages were assumed to have spherical shapes with diameters of 0.023 and 30 pm, respectively.
If additional unlabeled LDL were introduced into the reaction mixture, the contribution of a competing reverse transfer reaction to the observed rate constant should be greatly enhanced. Table I11 and Fig. 2 show the results of experiments in which increasing amounts of unlabeled LDL were added to cell suspensions immediately before addition of [G-3H] benzo(a)pyrene-labeled LDL to initiate cellular uptake. Even though concentrations of unlabeled LDL were increased, the rate constants under all conditions were identical.
To determine the rate constants for benzo(a)pyrene efflux, cells were preloaded with [G-3H]benzo(a)pyrene and then mixed with unlabeled LDL to initiate efflux. Aliquots of 1 X IO6 cells were mixed with 5, 50, or 500 pg of LDL protein Rate constants for bento(a)pyrene uptake as a function of unlabeled LDL concentration Rate constants were measured by the filtration protocol. The transfer was initiated with 0.5 pg ml" [G-3H]benzo(a)pyrene-labeled LDL as described under "Materials and Methods."  50 (A), or 500 (A) pg ml" unlabeled LDL and were stirred constantly at 21 "C. At time 0,0.5 pg ml" LDL containing [G-3H]benzo(a)pyrene was added. Aliquots were taken at time intervals and filtered as described under "Materials and Methods." Solid lines represent analysis of data with respect to a monoexponential function.

TABLE IV
Rate constants for efflux of benzo(a)pyrene from cells Rate constants were measured by the filtration protocol. The cells were previously labeled with [G-3H]ben~~(a)pyrene as described under "Materials and Methods." benzo(a)pyrene (Fig. 3). When movement of benzo(a)pyrene was diffusion-limited, the rate constant for this reaction was identical to that obtained in the presence of LDL. Thus, the relatively slow rate constant for benzo(a)pyrene entry into cells was independent of the lipoprotein carrier and not related to possible effects of diffusion barriers on movement of either benzo(a)pyrene or LDL.
Benzo(a)pyrene uptake was examined directly by measuring the constant for transfer of benzo(a)pyrene from phosphatidylcholine vesicles to the plasma membrane of cells. POPC vesicles containing benzo(a)pyrene and the nonexchangeable fluorescence quencher DNP-DOA were mixed rapidly with P388D1 macrophages in a stopped-flow system (Fig. 4). When the reaction was followed at the faster time base of 5 ms/ channel, two kinetic components were distinguished. A rapid increase that preceded the slow reaction occurred with a rate constant of 61 min", about the same as the rate constant measured for transfer of benzo(a)pyrene between vesicles, 72 min-l. The fast reaction contributed approximately 10% of the total increase in benzo(a)pyrene fluorescence. This quantity apparently represented benzo(a)pyrene associated with the plasma membrane-associated benzo(a)pyrene. When the increase in fluorescence intensity was followed at a time base of 100 ms/channel for approximately 3 min, a portion of a much slower reaction was observed. Therefore, the rate of transfer of benzo(a)pyrene to the plasma membranes of cells The fluorescence spectrum of the solution indicated an absence of crystalline benzo(a)pyrene. Macrophages (1.4 X lo4 cells) were suspended in 2 ml of the solution and stirred constantly at 26 "C. Aliquots of 100 pl were removed at indicated times and filtered as described under "Materials and Methods." A rate constant of 0.4 min" was determined by nonlinear regression analysis with respect to a monoexponential function. ml-', which corresponded to 6 X lo'*, 6 X and 6 X lo1* acceptors, respectively. The measured rate constants were essentially identical to those measured for uptake, about 0.2 min" (Table IV). The rate constants for efflux were also independent of extracellular lipoprotein concentrations. These data suggest that competitive uptake between extracellular LDL and the plasma membrane for benzo(a)pyrene is not the rate-limiting process in the entry of benzo(a)pyrene into cells. FIG. 4. Stopped-flow kinetics of benzo(a)pyrene transfer to cells. Macrophages (2 X lo6 cells ml") were mixed rapidly at 26 'C with an equal volume of 0.2 mM POPC vesicles containing 6 p~ DNP-DOA and 6 PM benzo(a)pyrene. Fluorescence intensity was measured at 5-ms intervals for the first half of the trace and at 100ms intervals for the second half. The solid line represents the analysis of the first half of the trace with respect to a monoexponential function with a rate constant of 61.2 min-I.

Absence of Boundary
is not determined by the rate at which benzo(a)pyrene desorbs from the donor vehicle and is apparently not impeded by a diffusion barrier surrounding the cells. Diffusion of the molecule through the aqueous phase and its readsorption by the acceptor vesicle are too fast to measure on a millisecond time scale (Charlton et al., 1976).
Activation Energies for Benzo(a)pyrene Transfer-The temperature dependence for uptake of benzo(a)pyrene by cells is shown in Fig. 5. The activation energies for uptake and efflux are 17.6 and 18.0 kcal mol", respectively. The magnitude of the activation energy for uptake of benzo(a)pyrene by cells is similar to that obtained from stopped-flow measurement of spontaneous transfer of benzo(a)pyrene between single bilayer vesicles, 11.7 kcal mol" ( Table V). These activation energies correspond to those reported for diffusion of molecules through phospholipid bilayers (Davison and Danielli, 1943). Activation energies of about 4 kcal mol" for fluorescence polarization of perylene in membranes suggest a relatively small temperature dependence for PAH movement in membranes (Rudy and Gitler, 1972). Therefore, diffusion of benzo(a)pyrene appears not to be the rate-limiting step for cellular uptake.
Effect of Molecular Size on PAH Transfer Rates-Rate constants for transfer of PAH between single bilayer vesicles are inversely proportional to their molecular surface areas (Plant et al., 1983). Increases in hydrophobic surface area of the transferring molecules and the decreased rate constants for intervesicular transfer are the consequence of reduced partitioning into the aqueous phase from the hydrocarbon domain. If a desorption process were to be rate-limiting in the  1.8 11.3 9.0 12.0 cellular uptake of PAH, then an inverse relationship between molecular size of PAH and rate of cell uptake would be expected. A series of PAH of various molecular sizes were examined to compare the rate constants for spontaneous transfer between single bilayer model membrane vesicles and the rate constants for uptake of this series of PAH by cells. Fig. 6 showed a strong positive correlation between the rate constants measured in these two systems, although rate constants for cellular uptake were 2 orders of magnitude smaller than rate constants for intermembrane transfer. This observation confirmed that the rate-limiting step in PAH permeation of cells was a desorption process which was kinetically independent of desorption from the extracellular donor and that the rate constant for permeation of PAH, was inversely proportional to the hydrophobicity of the compound.

DISCUSSION
A physiologically relevant mode of presentation of benzo(a)pyrene to cells is as a noncovalent component of lipoproteins rather than as microcrystalline dispersions (Kocan et al., 1983;Lakowicz et al., 1980). Studies in vivo (Grubbs and Moon, 1973) and in vitro (Shu and Nichols, 1981) have shown that dibenzanthracene and benzo(a)pyrene partition readily into plasma lipoproteins. Cell uptake of benzo-(a)pyrene has been shown to occur independently of lipopro-

FIG. 6. Comparison of rate constants for transfer of PAH between POPC vesicles and for transfer from vesicles to cells.
Rate constants were measured at 37 "C by stopped-flow for transfer between vesicles (ordinate) and by spectrofluorometry for transfer into cells (abscissa). The compounds, in order of increasing rate constants, were benzoperylene (structure not shown), 1,2,3,4-dibenzanthracene, benzo(a)pyrene; 9-phenylanthracene, 3,4-benzophenanthrene, and pyrene. The correlation coefficient was 0.994, and the slope was 412. tein endocytosis (Remsen and Shireman, 1981;Plant et al., 1985) and the size of the donor vehicle . Fig. 7 depicts PAH uptake into cells as an equilibration process involving three compartments: the extracellular lipoprotein or vesicle, the cell plasma membrane, and the intracellular lipid membranes and compartments. The experimental results, taken together, indicate that the rate-determining step in the permeation of cells by PAH is transferred out of the plasma membrane and that k,,,, the rate constant for desorption of the PAH from the plasma membrane, is the first-order rate constant that is measured experimentally. Since the rate of desorption of benzo(a)pyrene from extracellular LDL or vesicles is very fast, a steady state is rapidly achieved between benzo(a)pyrene in extracellular donors and benzo(a)pyrene in the plasma membrane. Steady state plasma membrane concentrations of benzo(a)pyrene are reached in milliseconds, with a rate constant which is determined by the rate constant for benzo(a)pyrene desorption out of the donor vehicle. Rate constants for benzo(a)pyrene permeation by cells are 2 orders of magnitude smaller than rate constants for desorption of benzo(a)pyrene from lipoproteins or single bilayer vesicles.
Differences in rate constants for desorption of lipophils from cell plasma membranes compared to POPC vesicles could in part be due to differences in lipid composition and packing constraints. Differences in rate constants for desorption of lipophils from model membranes have been measured as a function of matrix lipid composition (Damen and Scherphof, 1981;Poznansky and Czekanski, 1979;Phillips and Rothblat, 1985). However, the slow rate constant for transfer from cell plasma membranes is probably primarily due to the effect of the small radius of curvature of these membranes compared to smaller cells, vesicles, and lipoproteins. Rate constants for lipophil desorption from different lipoprotein classes have been found to be inversely proportional to the radius of the lipoprotein particles (Charlton and Smith, 1982;Almgren, 1980;Smith and Doody, 1981). Bojesen (1982) reported a 10-fold larger rate constant for transfer of cholesterol from plasma lipoproteins to red blood cells than from red blood cells to plasma. The rate constant for transfer of benzo(a)pyrene from red blood cell membranes to LDL (data not shown) is 1.07 min-', or 7-fold faster than the rate constant for efflux from a fibroblast or P388D, macrophage. The diameter of red blood cells is about 5-10 pm, compared to diameters of about 30-50, 0.03, and 0.023 pm for P388D1 macrophages, POPC vesicles, and LDL, respectively. The passive transfer mechanism described here for cellular uptake of PAH is analogous to cellular uptake of cholesterol (Phillips and Rothblat, 1985) and fatty acids (Walter and Gutknecht, 1984). These data suggest that cellular uptake of all relatively soluble lipophils is dominated by a passive mechanism. The relatively slow rate of benzo(a)pyrene uptake by cells suggests that the permeation of tissues by PAH and other passively transferring lipophils may be extremely slow when many layers of cells are involved and depends on the content of intracellular lipid.
The observed inverse relationship between hydrophobicity for a series of PAH and their permeation into cells is directly opposed to conclusions drawn from cell permeation studies involving relatively water-soluble compounds. For water-sol-uble molecules, diffusion of the hydrophobic molecule across the hydrophobic interior of the plasma membrane is ratelimiting. By contrast, for a very lipophilic molecule, desorption from the interior surface of the plasma membrane is slow, compared to other events. Thus, rate constants for membrane permeation of a series of homologous compounds that differ in hydrophobicity should increase with increasing hydrophobicity of the permeant, but only to the point at which the insolubility of the molecule in the aqueous phase limits the rate of desorption of the permeant from the inner leaflet of the plasma membrane.