Biosynthesis of (R)-beta-tyrosine and its incorporation into the highly cytotoxic chondramides produced by Chondromyces crocatus

The chondramides are mixed non-ribosomal peptide/polyketide secondary metabolites produced by the myxobacterium Chondromyces crocatus Cm c5, which exhibit strong cytotoxic activity. On the basis of their striking structural similarity to the marine depsipeptides jaspamides, the chondramides have been assumed to incorporate a (R)-beta-tyrosine moiety, an expectation we confirm here. Thus, the recent sequencing of the chondramide biosynthetic gene cluster provided the opportunity to probe the shared origin of this unusual beta-amino acid. We demonstrate here that (R)-beta-tyrosine is produced directly from l-tyrosine by the aminomutase CmdF. Along with the tyrosine aminomutase SgcC4 from the C-1027 enediyne pathway, this enzyme belongs to a novel family of tyrosine aminomutases related to the ammonium lyase family of enzymes but exhibits opposite facial selectivity for the hydroxycinnamate intermediate. We also show that the adenylation (A) domain in the chondramide pathway, which activates the beta-tyrosine building block, exhibits the required preference for (R)-beta-tyrosine, further arguing against alternative origins for the moiety in the chondramides. Comparison to the (S)-beta-tyrosine specific A domain SgcC1 should enhance our understanding of the structural and stereochemical determinants guiding amino acid selection by non-ribosomal peptide synthetase multienzymes.     

Involvement of myosin in intracellular motility and cytomorphogenesis in Micrasterias

Myosin was detected on Western blots of Micrasterias denticulata extracts by use of antibodies from different sources. Inhibitors with different targets of the actomyosin system, such as the myosin ATPase-blockers N-ethylmaleimide (NEM) and 2,3-butanedione monoxime (BDM), or the myosin light chain kinase inhibitor 1-(5-iodonaphthalene-1-sulfonyl)-1H-hexhydro-1,4-diazapine (ML7), had similar effects on intracellular motility during cell development in the green alga Micrasterias, thus pointing towards a participation of myosin in these processes. The drugs markedly altered the mode of postmitotic nuclear migration, slowed down cytoplasmic streaming, changed cell pattern development and prevented normal chloroplast distribution and spreading into the growing semicell. In addition, an increase and dilatations in ER cisternae and marked morphological changes of the Golgi system were observed by transmission electron microscopy after exposure of growing cells to BDM.Neither BDM nor ML7 exhibited any effect on the distribution or arrangement of the cortical F-actin network nor on the F-actin basket around the nucleus, characteristic of untreated growing Micrasterias cells (J Cell Sci 107 (1994) 1929). This is particularly interesting since BDM caused disintegration of the microtubule system co-localized to the F-actin cage during normal nuclear migration. Together with the fact that other microtubules not connected to the F-actin system remained uninfluenced by BDM, this observation is evidence of an integrative function of myosin between the cytoskeleton elements.     

Effect of cytochalasins on F-actin and morphology of Ehrlich ascites tumor cells

Cytochalasins have been used extensively to probe the role of F-actin in different aspects of cellular function. Most of the data obtained are interpreted on the basis of the well-established depolymerizing effects of cytochalasins on F-actin preparations in vitro. However, some evidence indicates that, in intact cells, different cytochalasins can have varying effects on cell morphology and F-actin content and organization. To examine this problem in more detail, we analyzed the effects of cytochalasins on the cell morphology of and F-actin content and organization in Ehrlich ascites tumor (EAT) cells. After a 3-min exposure to 0.5 microM cytochalasin D, B, or E, F-actin content was equally reduced in all cases and this correlated with a reduction in the amount of cortical F-actin associated with the EAT cell membrane. However, only with CE was cell morphology markedly altered, with the appearance of numerous blebs. At 10 microM, blebbing was present in all conditions and the organization of cortical F-actin was disrupted. F-actin content, however, was not further reduced by this higher concentration and in CD it was identical to control levels. Exposure of EAT cells to similar concentrations of cheatoglobosin C, an analog of the cytochalasins that has little to no affinity for F-actin, resulted in a loss of F-actin content, a reduction in F-actin fluorescence, but no change in cell morphology, including a complete lack of bleb formation. Myosin II immunoreactivity, concentrated in the cortical cytoplasm colocalized with F-actin and in an area associated with the Golgi, was reduced by the high-dose cytochalasin. These results demonstrate that caution must be exercised in the use of cytochalasins to probe the role of F-actin in cellular function and that several parameters must be analyzed to obtain an accurate assessment of the effect of cytochalasin on the actin filament system.     

Role of the F-actin cytoskeleton in the RVD and RVI processes in Ehrlich ascites tumor cells.

The role of the F-actin cytoskeleton in cell volume regulation was studied in Ehrlich ascites tumor cells, using a quantitative rhodamine-phalloidin assay, confocal laser scanning microscopy, and electronic cell sizing. A hypotonic challenge (160 mOsm) was associated with a decrease in cellular F-actin content at 1 and 3 min and a hypertonic challenge (600 mOsm) with an increase in cellular F-actin content at 1, 3, and 5 min, respectively, compared to isotonic (310 mOsm) control cells. Confocal visualization of F-actin in fixed, intact Ehrlich cells demonstrated that osmotic challenges mainly affect the F-actin in the cortical region of the cells, with no visible changes in F-actin in other cell regions. The possible role of the F-actin cytoskeleton in RVD was studied using 0. 5 microM cytochalasin B (CB), cytochalasin D (CD), or chaetoglobosin C (ChtC), a cytochalasin analog with little or no affinity for F-actin. Recovery of cell volume after hypotonic swelling was slower in cells pretreated for 3 min with 0.5 microM CB, but not in CD- and ChtC-treated cells, compared to osmotically swollen control cells. Moreover, the maximal cell volume after swelling was decreased in CB-treated, but not in CD- or Chtc-treated cells. Following a hypertonic challenge imposed using the RVD/RVI protocol, recovery from cell shrinkage was slower in CB-treated, but not in CD- or Chtc-treated cells, whereas the minimal cell volume after shrinkage was unaltered by either of these treatments. It is concluded that osmotic cell swelling and shrinkage elicit a decrease and an increase in the F-actin content in Ehrlich cells, respectively. The RVD and RVI processes are inhibited by 0.5 microM CB, but not by 0.5 microM CD, which is more specific for actin.     

Anti-tumor activity of the farnesyl-protein transferase inhibitors arteminolides, isolated from Artemisa

Members of the Artemisia genus are important medicinal plants found throughout the world. Arteminolides A-D (1-4), isolated from the aerial parts of Artemisia, have an inhibitory activity on farnesyl-protein transferase (FPTase; EC 2.5.1.29) in in vitro assay. This study was carried out with the purpose of validating anti-tumor effects of the compounds in human tumor cells and mouse xenograft model. The arteminolides inhibited tumor cell growth in a dose-dependent manner. Furthermore, arteminolide C (3) blocked in vivo growth of human colon and lung tumor xenograft without the loss of body weight in nude mice.     

Chemoattractant-stimulated polymorphonuclear leukocytes contain two populations of actin filaments that differ in their spatial distributions and relative stabilities

Chemoattractants stimulate actin polymerization in lamellipodia of polymorphonuclear leukocytes. We find that removal of chemoattractant results in rapid (within 10 s at 37 degrees C) and selective depolymerization of the F-actin located in lamellipodia. Addition of 10 microM cytochalasin B, in the presence of chemoattractant, also resulted in rapid and selective depolymerization of lamellar F-actin. The elevated F-actin level induced by chemoattractant rapidly returns to the level present in unstimulated cells after (a) a 10-fold decrease in chemoattractant concentration; (b) the addition of 10 microM cytochalasin B; or (c) cooling to 4 degrees C. The F-actin levels of unstimulated cells are only slightly affected by these treatments. Based on the similar effects of cytochalasin addition and chemoattractant dilution, it is likely that both treatments result in actin depolymerization from the pointed ends of filaments. Based on our results we propose that chemoattractant-stimulated polymorphonuclear leukocytes contain two distinct populations of actin filaments. The actin filaments within the lamellipodia are highly labile and in the continued presence of chemoattractant these filaments are rapidly turning over, continually polymerizing at their plus (barbed) ends, and depolymerizing at their minus ends. In contrast, the cortical F-actin filaments of both stimulated and unstimulated cells are differentially stable.     

The kinetics of chemotactic peptide-induced change in F-actin content, F-actin distribution, and the shape of neutrophils

Formyl-met-leu-phe (fMLP) induces actin assembly in neutrophils; the resultant increase in F-actin content correlates with an increase in the rate of cellular locomotion at fMLP concentrations less than or equal to 10(-8) M (Howard, T.H., and W.H. Meyer, 1984, J. Cell Biol., 98:1265-1271). We studied the time course of change in F-actin content, F-actin distribution, and cell shape after fMLP stimulation. F-actin content was quantified by fluorescence activated cell sorter analysis of nitrobenzoxadiazole-phallacidin-stained cells (Howard, T.H., 1982, J. Cell Biol., 95(2, Pt. 2:327a). F-actin distribution and cell shape were determined by analysis of fluorescence photomicrographs of nitrobenzoxadiazole-phallacidin-stained cells. After fMLP stimulation at 25 degrees C, there is a rapid actin polymerization that is maximal (up to 2.0 times the control level) at 45 s; subsequently, the F-actin depolymerizes to an intermediate F-actin content 5-10 min after stimulation. The depolymerization of F-actin reflects a true decrease in F-actin content since the quantity of probe extractable from cells also decreases between 45 s and 10 min. The rate of actin polymerization (3.8 +/- 0.3-4.4 +/- 0.6% increase in F-actin/s) is the same for 10(-10) - 10(-6) M fMLP and the polymerization is inhibited by cytochalasin D. The initial rate of F-actin depolymerization (6.0 +/- 1.0-30 +/- 5% decrease in F-actin/min) is inversely proportional to fMLP dose. The F-actin content of stimulated cells at 45 s and 10 min is greater than control levels and varies directly with fMLP dose. F-actin distribution and cell shape also vary as a function of time after stimulation. 45 s after stimulation the cells are rounded and F-actin is diffusely distributed; 10 min after stimulation the cell is polarized and F-actin is focally distributed. These results indicate that actin polymerization and depolymerization follow fMLP stimulation in sequence, the rate of depolymerization and the maximum and steady state F-actin content but not the rate of polymerization are fMLP dose dependent, and concurrent with F-actin depolymerization, F-actin is redistributed and the cell changes shape.     

Cortical microtubules and cortical microfilaments in the green alga,Micrasterias pinnatifida

Summary Cortical microtubules and cortical microfilaments were visualized in cells ofMicrasterias pinnatifida treated by freeze-substitution, and the pattern of their distribution was reconstructed from serial sections. Most cortical microtubules accompanied the long microfilaments that ran parallel to the microtubules. Cortical microfilaments not accompanied by the microtubules were also found. They were short and slightly curved. Both types of cortical microfilament were not grouped into bundles, and were 6–7 nm in diameter, a value that corresponds to the diameter of filaments of F-actin.     

Induction of anti-actin drug resistance in Tetrahymena

Both cytochalasin D and latrunculin B reversibly inhibited Tetrahymena phagocytosis at concentrations similar to those effective in mammalian systems, even though ciliate actins are known to be highly divergent from mammalian actins. Overnight exposure to relatively low (0.25 microM) concentrations of latrunculin B induced resistance in Tetrahymena to the inhibitory effects of that drug, as well as cross-resistance to cytochalasin D. However, much higher (> 30 microM) concentrations of cytochalasin D were required for induction of cross-resistance to latrunculin B. Anti-actin drug resistance in Tetrahymena may involve a general multidrug resistance mechanism and/or specific feedback regulation of F-actin assembly and stability.     

Association of Serotonin, Dopamine, or Noradrenaline with an Actin-Like Component in Pheochromocytoma (PC12) Cells

A rat pheochromocytoma (PC12) cell line was used to examine the possibility that 5-hydroxytryptamine (serotonin), 3,4-dihydroxyphenylethylamine (dopamine), or noradrenaline may be associated with cytoplasmic actin, as was suggested by previous in vitro binding studies on an actin-like protein from rat brain synaptosomes. When PC12 cells were incubated with [3H]serotonin. [3H]dopamine, or [3H]noradrenaline for 30 min at 37 degrees C, approximately 2-4% of the radioactivity present in the cells was found to be associated with a high-molecular-weight (actin-like) component in supernatant fractions. Evidence relating this monoamine binding component to actin filaments includes: (a) its strong absorption by myosin filaments at low ionic strength: (b) a decrease in its affinity for myosin in the presence of 1 mM ATP, which lowers the affinity of authentic actin for myosin: (c) displacement of bound [3H]serotonin from it by DNase I, which binds strongly to actin and which inhibits [3H]serotonin binding to actin in vitro; (d) an increase in its binding of each monoamine (by 25-40%) after PC12 cells were preincubated with 10 microM cytochalasin B (a drug that induces depolymerization of F-actin). These findings suggest that serotonin, dopamine, or noradrenaline may associate with actin filaments in vivo.     

Changes in cell morphology and actin organization during heat shock in Dictyostelium discoideum: does HSP70 play a role in acquired thermotolerance?

In response to heat shock (34°C, 30 min), cell morphology and actin organization in Dictyostelium discoideum are drastically changed. Loss of pseudopodia and disappearance of F-actin-containing structures were observed by using fluorescence microscopy. These changes were paralleled by a rapid decrease of the F-actin content measured by a TRITC-phalloidin binding assay. The effects of heat shock on cell morphology and actin organization are transient: After heat shock (34°C) or during a long-term heat treatment (30°C), cell morphology, F-actin patterns and F-actin content recovered/adapted to a state which is characteristic for untreated cells. Because F-actin may be stabilized by increased amounts of heat shock proteins, their response and interaction with F-actin was analyzed. After a 1 h heat treatment (34°C), the major heat shock protein of D. discoideum (HSP70) showed maximally increased synthesis rates and levels. During recovery from a 34°C shock or during a continuous heat treatment at 30°C, the HSP70 content first increased and then declined slowly toward normal levels. Pre-treatment of cells with a short heat shock of 30 min at 34°C stabilized the F-actin content when the cells were exposed to a second heat shock. Furthermore, a transient colocalization of HSP70 and actin was observed at the beginning of heat treatment (30°C) using immunological detection of HSP70 in the cytoskeletal actin fraction.     

A major constituent of green tea,EGCG, inhibits the growth of a human cervical cancer cell line,CaSki cells,through apoptosis,G(1) arrest,and regulation of gene expression

A constituent of green tea, (-)-epigallocatechin-3-gallate (EGCG) has been known to possess antiproliferative properties. In this study, we investigated the anticancer effects of EGCG in human papillomavirus (HPV)-16 associated cervical cancer cell line, CaSki cells. The growth inhibitory mechanism(s) and regulation of gene expression by EGCG were also evaluated. EGCG showed growth inhibitory effects in CaSki cells in a dose-dependent fashion, with an inhibitory dose (ID)(50) of approximately 35 microM. When CaSki cells were further tested for EGCG-induced apoptosis, apoptotic cells were significantly observed after 24 h at 100 microM EGCG. In contrast, an insignificant induction of apoptotic cells was observed at 35 microM EGCG. However, cell cycles at the G1 phase were arrested at 35 microM EGCG, suggesting that cell cycle arrests might precede apoptosis. When CaSki cells were tested for their gene expression using 384 cDNA microarray, an alteration in the gene expression was observed by EGCG treatment. EGCG downregulated the expression of 16 genes over time more than twofold. In contrast, EGCG upregulated the expression of four genes more than twofold, suggesting a possible gene regulatory role of EGCG. This data supports that EGCG can inhibit cervical cancer cell growth through induction of apoptosis and cell cycle arrest as well as regulation of gene expression in vitro. Furthermore, in vivo antitumor effects of EGCG were also observed. Thus, EGCG likely provides an additional option for a new and potential drug approach for cervical cancer patients.     

Effect of jasplakinolide on the growth, encystation, and actin cytoskeleton of Entamoeba histolytica and Entamoeba invadens

The effect of jasplakinolide. an actin-polymerizing and filament-stabilizing drug, on the growth, encystation, and actin cytoskeleton of Entamoeba histolytica and Entamoeba invadens was examined. Jasplakinolide inhibited the growth of E. histolytica strain HM-1:IMSS and E. invadens strain IP-1 in a concentration-dependent manner, the latter being more resistant to the drug. The inhibitory effect of jasplakinolide on the growth of E. histolytica trophozoites was reversed by removal of the drug after exposure to 1 microM for 1 day. Encystation of E. invadens as induced in vitro was also inhibited by jasplakinolide. Trophozoites exposed to jasplakinolide in encystation medium for 1 day did not encyst after removal of the drug, whereas those exposed to the drug in growth medium for 7 days did encyst without the drug. The process of cyst maturation was unaffected by jasplakinolide. Large round structures were formed in trophozoites of both amoebae grown with jasplakinolide; these were identified as F-actin aggregates by staining with fluorescent phalloidin. Accumulation in trophozoites of both amoebae of actin aggregates was observed after culture in jasplakinolide. Also, E. invadens cysts formed from trophozoites treated with jasplakinolide contained the actin aggregate. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis and immunoblot analysis revealed that the jasplakinolide treatment led to an increase in the proportion of F-actin associated with formation of the aggregate. The results suggest that aggregates are formed from the cortical flow of F-actin filaments, and that these filaments would normally be depolymerized but are artificially stabilized by jasplakinolide binding.     

Rupestrines A-D,alkaloids from the aerial parts of Corydalis rupestris

Phytochemical investigation of the dichloromethane extract of the dried aerial parts of Corydalis rupestris (Papaveraceae) resulted in the identification of four new isoquinoline alkaloids rupestrines A-D and one known isoquinoline alkaloid, namely, stylopine. The structures of these compounds were characterized by extensive spectroscopic methods including 1D- (¹H and ¹³C) and 2D NMR experiments (COSY, HSQC, HMBC, and NOESY) as well as HRESIMS analyses. In addition, the absolute configurations of rupestrines A-D were determined using modified Mosher’s method. Cytotoxic effects of alkaloids and their interaction with albumin were also investigated in this study.     

Latrunculin B-induced plant dwarfism: Plant cell elongation is F-actin-dependent

Marine macrolides latrunculins are highly specific toxins which effectively depolymerize actin filaments (generally F-actin) in all eukaryotic cells. We show that latrunculin B is effective on diverse cell types in higher plants and describe the use of this drug in probing F-actin-dependent growth and in plant development-related processes. In contrast to other eukaryotic organisms, cell divisions occurs in plant cells devoid of all actin filaments. However, the alignment of the division planes is often distorted. In addition to cell division, postembryonic development and morphogenesis also continue in the absence of F-actin. These experimental data suggest that F-actin is of little importance in the morphogenesis of higher plants, and that plants can develop more or less normally without F-actin. In contrast, F-actin turns out to be essential for cell elongation. When latrunculin B was added during germination, morphologically normal Arabidopsis and rye seedlings developed but, as a result of the absence of cell elongation, these were stunted, resembling either genetic dwarfs or environmental bonsai plants. In conclusion, F-actin is essential for the plant cell elongation, while this F-actin-dependent cell elongation is not an essential feature of plant-specific developmental programs.     

Evidence for a gelsolin-rich, labile F-actin pool in human polymorphonuclear leukocytes.

Filamentous (F) actin is a major cytoskeletal element in polymorphonuclear leukocytes (PMNs) and other non-muscle cells. Exposure of PMNs to agonists causes polymerization of monomeric (G) actin to F-actin and activates motile responses. In vitro, all purified F-actin is identical. However, in vivo, the presence of multiple, diverse actin regulatory and binding proteins suggests that all F-actin within cells may not be identical. Typically, F-actin in cells is measured by either NBDphallacidin binding or as cytoskeletal associated actin in Triton-extracted cells. To determine whether the two measures of F-actin in PMNs, NBDphallacidin binding and cytoskeletal associated actin, are equivalent, a qualitative and quantitative comparison of the F-actin in basal, non-adherent endotoxin-free PMNs measured by both techniques was performed. F-actin as NBDphallacidin binding and cytoskeletal associated actin was measured in cells fixed with formaldehyde prior to cell lysis and fluorescent staining (PreFix), or in cells lysed with Triton prior to fixation (PostFix). By both techniques, F-actin in PreFix cells is higher than in PostFix cells (54.25 +/- 3.77 vs. 23.5 +/- 3.7 measured as mean fluorescent channel by NBDphallacidin binding and 70.3 +/- 3.5% vs. 47.2 +/- 3.6% of total cellular actin measured as cytoskeletal associated actin). These results show that in PMNs, Triton exposure releases a labile F-actin pool from basal cells while a stable F-actin pool is resistant to Triton exposure. Further characterizations of the distinct labile and stable F-actin pools utilizing NBDphallacidin binding, ultracentrifugation, and electron microscopy demonstrate the actin released with the labile pool is lost as filament. The subcellular localization of F-actin in the two pools is documented by fluorescent microscopy, while the distribution of the actin regulatory protein gelsolin is characterized by immunoblots with anti-gelsolin. Our studies show that at least two distinct F-actin pools coexist in endotoxin-free, basal PMNs in suspension: 1) a stable F-actin pool which is a minority of total cellular F-actin, Triton insoluble, resistant to depolymerization at 4 degrees C, gelsolin-poor, and localized to submembranous areas of the cell; and 2) a labile F-actin pool which is the majority of total cellular F-actin, Triton soluble, depolymerizes at 4 degrees C, is gelsolin-rich, and distributed diffusely throughout the cell. The results suggest that the two pools may subserve unique cytoskeletal functions within PMNs, and should be carefully considered in efforts to elucidate the mechanisms which regulate actin polymerization and depolymerization in non-muscle cells.     

Effect of non-esterified fatty acids on bovine granulosa cell steroidogenesis and proliferation in vitro

In high-yielding dairy cows, the negative energy balance (NEB) during the first weeks post partum may influence dominant follicle growth and steroidogenesis. Since non-esterified fatty acid (NEFA) concentrations are elevated during NEB and are shown to be toxic for several cell types, we investigated the individual and combined effects of the three main NEFA's on granulosa cell proliferation and steroidogenesis in vitro. Granulosa cells from large follicles were cultured for two days in serum free medium in the presence of palmitic (C16:0) (PA), stearic (C18:0) (SA) and/or oleic acid (C18:1) (OA). Addition of 150, 300 or 500 microM of PA and SA inhibited cell proliferation (P<0.05) while OA only elicited such an effect at 500 microM (P<0.01). In the combination treatment (150 microM of each fatty acid), cell numbers were also reduced (P<0.01). These inhibitory effects on cell number are partly due to the induction of apoptosis by these NEFA's, as was demonstrated by annexin V-FITC/propidium iodide staining of the granulosa cells. Oestradiol-17beta production was stimulated by all doses of PA, by 300 and 500 microM of SA and by 500 microM of OA (P<0.05). Combined treatment with 150 microM of each fatty acid also stimulated oestradiol-17beta production per 10(4) cells (P<0.05). We can conclude that PA, SA and to a lesser degree OA modulate granulosa cell proliferation and steroidogenesis in vitro. These effects may be involved in the occurrence of ovarian dysfunction during the postpartum period in high-yielding dairy cows.     

Stimulation of regulatory volume decrease (RVD) by isolated bovine articular chondrocytes following F-actin disruption using latrunculin B

Articular chondrocytes are exposed to significant changes in extracellular osmolarity during normal joint activity, which can lead to changes in cell volume and metabolism of the extracellular matrix (ECM). Chondrocytes can respond to cell swelling/shrinking by volume regulatory pathways, but the signalling pathways are poorly understood although a role for the cytoskeleton is frequently implicated. Here, we have investigated the effects of disruption of the chondrocyte F-actin cytoskeleton on the recovery of cell volume by RVD. The cytoskeleton was perturbed using the relatively specific agent latrunculin B (5 microM; 30 min) and loss of F-actin integrity quantified using fluorescent phalloidin-labelling and confocal laser scanning microscopy (CLSM). Imaging of isolated chondrocytes labelled with Fura-2 to measure the fluorescence associated with cell volume changes, showed that the extent of hypo-osmotic swelling was unaffected by latrunculin B treatment. Two categories of the chondrocyte RVD response were observed: 'fast' RVD where at 3 min post-osmotic challenge there was a recovery in cell fluorescence of >or=80%, whereas other cells exhibited 'slow' RVD. Latrunculin B increased the proportion of chondrocytes demonstrating 'fast' RVD by approximately 10 fold and reduced those cells showing 'slow' RVD. An inhibitor of chondrocyte RVD (REV 5901) had no significant effect on the integrity of the cytoskeleton showing that the RVD response could be inhibited independent of the state of the F-actin cytoskeleton. These results suggest that the intact cortical F-actin cytoskeleton has a restraining effect on the RVD response of isolated bovine articular chondrocytes.