Biochimica et Biophysics Acta, 1043 (1990) 19-26 Elsevier
BBALIP
19
53343
CTP:cholinephosphate cytidylyltransferase in human and rat lung: Association in vitro with cytoskeletal actin Alan N. Hunt, I. Colin S. Normand and Anthony D. Postle Child Health, Faculty of Medicine, Southampton General Hospital, Southampton (U.K.) (Received
Key words:
CTP:Cholinephosphate
cytidylyltransferase;
6 October
1989)
Actin; Cytoskeletal (Human lung)
interaction;
Poly(ethyleneglyco1);
Enzyme activation;
CTPxholinephosphate cytidylyltransferase activities were compared in saline homogenates of immature fetal (15-16 weeks gestation) and adult human lung. There were no differences in subcellular enzyme distribution, in V,,, activity, or in the phosphatidylglycerol-mediated stimulation of soluble enzyme activity. These results provide no support for a developmental tram&cation of cytidylyltransferase from a cytosolic to a microsomal location in human lung, such as that proposed to accompany the maturation of pulmonary surfactant phosphatidylcholine biosynthesis in rat. Soluble cytidylyltransferase activity from human but not rat lung was increased after manipulation in vitro. Resolution of human H form (> lo3 kDa) and L form (200 kDa) enzyme by gel filtration led to an activity increase of 200%. Incubation at 37 o C for 2 h increased soluble enzyme recovery, although prior centrifugal removal of generated actin-rich aggregates was necessary in adult lung fractions. In contrast, 85% of soluble rat lung cytidylyltransferase was actin aggregate-associated after incubation. The apparent heteroassociation of rat and human lung enzyme with actin in the presence of poly(ethylene glycol) at 4” C strongly suggested close in vitro and potential in vivo linkage. A partial co-purification of adult human lung cytidylyltransferase with actin was also consistent with this idea. We propose that some reported cytidylyltransferase translocation phenomena may be mediated by cytoskeletal interactions in vitro.
Introduction
dependent
on the composition
of homogenising
buffers
[5,61. A regulatory role for CTP:cholinephosphate cytidylyltransferase (cytidylyltransferase) (EC 2.7.7.15) in the CDPcholine pathway of mammalian phosphatidylcholine (PC) biosynthesis is well established [l]. The claims of a rate-limiting function [l] imply that its control strength [2] may be near unity. Tissue fractionation following a variety of pathway flux manipulations have shown apparent enzyme redistribution between soluble and particulate subcellular compartments [3]. A regulatory translocation between an active, membraneassociated form and a less active soluble enzyme has been proposed [3]; a characteristic typical of enzymes described as ambiquitous [4]. The distribution of cytidylyltransferase between soluble and particulate fractions of liver and lung homogenates is critically
Abbreviations: PC, phosphatidylcholine; PG. phosphatidylglycerol.
PEG,
poly(ethyleneglyco1);
Correspondence: A.N. Hunt, Child Health, Faculty of Medicine, Southampton General Hospital, Tremona Road, Southampton, SO9 4XY, U.K. 00052760/90/$03.50
0 1990 Elsevier Science Publishers
B.V. (Biomedical
Translocation of brain hexokinase, the prototype ambiquitous enzyme [4], has recently been re-evaluated using rapid fractionation techniques, which failed to demonstrate any redistribution other than that of artefactual origin [7]. The significance of the apparent translocation of pulmonary cytidylyltransferase, including that accompanying the maturation of the surfactant system [S], has likewise been questioned [9] following observed flux-independent redistributions [lo]. Moreover, reported flux-dependent increases in the soluble, phospholipid-stimulated activity in the absence of redistribution phenomena are difficult to reconcile with the translocation hypothesis [ll]. PC synthesis in phagocytic cells, by both CDPcholine and N-methylation pathways, can be regulated by the state of microfilament and microtubule assembly, respectively [12]. Together with the recent demonstration of CDPcholine pathway compartmentalisation [13], this suggests some functional association of elements of the cytoskeleton with the enzymes of PC synthesis. We are unaware of similar observations with other non-phagocytic cells. The subcellular distribution of cytoskeletal Division)
20 components in vitro is also sensitive to fractionation conditions [14]; changes in such conditions can consequently affect associations of enzymes with cytoskeletal fractions in tissue homogenates. Such a variation in distribution is well documented for a number of glycolytic enzymes [15]. Several candidate ambiquitous enzymes have strong associations both in vivo and in vitro with elements of the cytoskeleton. These include phosphofructokinase [16,17], glucose-6-phosphate dehydrogenase [18] and Ca*+-phospholipid-dependent protein kinase C [19]. Cytoskeletal binding can result in changes both in K, and V,,,,, parameters [17,18]. Demonstrations of enzyme-cytoskeleton interactions in vitro are problematic due to the non-physiological dilution accompanying tissue fractionation. The large space-filling polymer poly(ethylene glycol) (PEG) has been employed to increase effective protein concentration in vitro and promote polymerisations and heteroassociations known to occur in vivo [20]. PEG has recently been used with pure and semipurified glycolytic enzyme preparations to show and evaluate enzyme/ actin and enzyme/enzyme associations at effective protein concentrations close to those encountered in vivo
1211. In this study, properties of cytidylyltransferase have been compared between fetal (15-16 weeks gestation) human lung, morphologically comparable to d18 rat lung [22], and adult human lung in order to assess any maturational variation in the parameters. Similarities and differences between human and rat lung enzyme have also been evaluated. These properties include subcellular distribution, extent of lipid stimulation and potential cytoskeletal interactions. Materials
Chemicals and biochemicals were obtained from BDH Chemicals (Poole, U.K.) or Sigma (Poole, U.K.). [ Me-‘4C]Choline phosphate was obtained from Amersham International (U.K.). Chromatography media were purchased from Pharmacia (Uppsala, Sweden). Fetal human lung was obtained from therapeutic abortion at 15-16 weeks gestation, and nonmalignant adult lung biopsies were taken during surgical procedure. Fetal human lung-derived fibroblasts were maintained in monolayer culture [23] and harvested in logarithmic phase growth at passage number 5-7. Rat lungs, female Wistar (150-200 g) were blanched in situ and lavaged (6 x 15 ml) with ice-cold saline to remove pulmonary surfactant [24]. Methods
Lungs and fibroblasts were homogenised in buffer A: 150 mM NaCl/SO mM Tris-HCl/l mM EDTA/l mM DTT/l mM NaN,/200 PM phenylmethylsulphonyl
fluoride (pH 7.4) at 4°C. Tissue was homogenised (20%, w/v) using an IKA-Ultra Turrax T18 homogeniser (Janke and Kunkel) and fibroblast homogenates (3-4%, w/v) were prepared by sonication on ice (3 x 20 s, 14 pm amplitude) using a Soniprep 150 with exponential microprobe (MSE Instruments, Crawley, U.K.). Initial centrifugation at 10000 x g for 10 min yielded a first pellet containing tar-like material in adult human lungs; this pellet exhibited negligible cytidylyltransferase activity in any tissue and was discarded. The supernatant was centrifuged at 100000 x g for 60 min to provide a membrane-rich particulate fraction and a soluble fraction. Lipid floating at the surface was removed by filtration through glass wool. Cytidylyltransferase assays monitored the conversion of [ Me-‘4C]choline phosphate (4726 dpm/nmol) to [ Me-‘4C]CDPcholine over 20 min at 37 o C. CDPcholine was isolated by charcoal binding [6], modified as detailed for use with 1.5 ml microfuge tubes. The final assay mix (90 ~1) contained enzyme, 60 mM TrisHC1/120 mM NaCl/l.4 mM EDTA/2.2 mM DTT/4 mM MgCl/4 mM CTP (pH 7.4) f 0.25 mM PG. Egg yolk phosphatidylglycerol (PG, Sigma) was sonicated directly into the assay buffer before use (3 x 20 s, 14 pm). The reaction was started with the addition of [Me-‘4C]choline phosphate (10 ~1. 16.6 mM, 2 ~Ci/~mol). The reaction was stopped by placing the tubes in boiling water for 2 min. Acid-washed charcoal (500 ~1) [25] was then added. Bound CDPcholine was separated by centrifugation (12 000 x g for 1 min) followed by 4 X 1 ml washes with distilled water. CDPcholine was eluted with ethanol/water/O.880 ammonia (60: 37 : 3, v/v) (2 X 1 ml, 90°C for 20 min) and radioactivity was determined by liquid scintillation in Cocktail T (BDH Chemicals). 1 mU of activity was defined as the conversion of 1 nmol choline phosphate to CDPcholine per min at 37°C. The reaction velocity was constant over 30 min and proportional to enzyme activity up to 20 mu/ml. Gel filtration on Sepharose 6B (90 X 2.5 cm) in buffer A at 4” C was monitored at 280 nm, using a flow-rate of 40 ml/h and 10 ml fraction collection. Cytidylyltransferse assays of column fractions were made singly f0.25 mM PG. Polyethylene glycol 6000 (BDH, mol.wt. 6000-7500) was prepared as a stock solution (10% or 20%) in buffer A and stored at 4 o C. Human lung actin was prepared by ammonium sulphate fractionation, gel filtration and ion-exchange chromatography as previously described [26] and its identity was confirmed by M,, amino acid composition and Western blot analysis with chicken anti-muscle actin. Discontinuous SDS-PAGE was performed on 10% slab gels with 6% stacking gel according to the method
21
PG or equivalent concentration of total lung lipid (results not shown). PG stimulation of soluble rat lung cytidylyltransferase was not as great as the 5-fold seen previously [25]. Some tissue lipid contamination seemed likely despite careful lavage to remove surfactant lipids. Delipidation with 10 vol. isopropyl ether appeared to confirm this with enzyme activity declining to 30% of non-delipidated fractions, while treatment with PG or whole lung lipid restored maximum activity. Measurement of lipid phosphorus in the soluble fractions of both adult and fetal human lung revealed a 13.5-fold difference in phospholipid concentration at 0.63 f. 0.19 mM (n = 4) and 46.6 k 8.5 PM (n = 4) respectively. In the case of the fetal enzyme, this level of endogenous lipid contamination seemed too small to account for the lack of difference in response to PG. The full extent of the phospholipid stimulation effect could not be accurately assessed, however, since delipidation of human soluble fractions using isopropyl ether irreversibly removed all measureable cytidylyltransferase activity. The apparent requirement of freshly isolated soluble human lung enzyme for some lipid to express activity was in contrast to the partially purified H and L forms, which needed no exogenous lipid and which in the case of L form enzyme fractions contained less than 1 I_LMlipid phosphorus. Particulate cytidylyltransferase activities ((10000 x g
of Laemmli [27] and stained with Coomassie Brilliant Blue. Protein was determined using a modified FolinCoicalteau reagent [28] with deoxycholate/ trichloroacetic acid co-precipitation. BSA (1 mg/ml) was used as a reference standard. The phospholipid phosphorus of lipid extracts [29] was determined according to the method of Bartlett [30]. Results Subcellular
distribution of human lung cytidylyltransferase
In common with most reports of cytidylyltransferase distribution in mammalian tissues, fetal and adult human lung enzyme activity was found associated with both soluble and particulate fractions of saline homogenates (Table I). The proportion of cytidylyltransferase activity in the soluble fraction of human lung homogenates was comparable for fetal and adult tissue; in each case the soluble (100 000 x g for 60 min supernatant) fraction contained about 66% of measurable activity. Rat lung cytidylyltransferase distribution following fractionation under similar conditions was also comparable. The extent of lipid stimulations of soluble human enzyme in the presence of 0.25 mM PG were similar (Table I), with no greater activity observed at 1.25 mM
TABLE
I
Cytidylyltransferase
activities in soluble and particulate fractions
of human and adult rat lung homogenates
Fresh fetal and adult human and adult rat lung was homogenised and separated into soluble and particulate components and cytidylyltransferase activities (* PC) determined as described in methods. After ageing at 37 o C for 2 h, soluble activity was measured before and after removal of aggregated protein by a 2nd 100000X 8 for 60 mm centrifugation. * Significantly different (P < 0.005) from initial soluble activity. Cytidylyltransferase
Fetal lung Without PC + 0.25 mM PC Distribution (% Total) Adult lung Without PC + 0.25 mM PC Distribution (% Total) Adult rat lung Without PC + 0.25 mM PC Distribution (% Total)
activity
(mU/mg
protein + S.E.)
particulate fraction
soluble fraction initial soluble
aged soluble
2nd supernatant
1.65 + 0.17 1.65 f 0.29 (n =16) 33.2 +4.8%
0.36 + 0.05 0.80 k 0.06 (n = 25) 66.8 *4.8%
* 0.65 f 0.08 * 1.48 * 0.26 (n=S) _
* 0.68 f 0.26 * 1.46 k 0.26 (n=6) _
1.58k0.39 1.88*0.44 (n = 5) 33.9 f7.5W
0.23 + 0.03 0.58 * 0.07 (n = 10) 66.1 *7.5%
0.48 f 0.20 0.67kO.15 (n = 3) _
* 1.23 + 0.30 * 1.3OkO.28 (n = 3)
3.21+ 0.92 3.05 k 0.61 (n = 8) 31.5 *3.3x
1.56 k 0.08 2.01*0.11 (n = 8) 68.5 *3.3%
1.88+_0.17 2.03 + 0.21 (n = 8) _
* 0.29 * 0.04 * 0.34+0.07 (n = 6) _
22 for 10 min) - (100000 X g for 60 min pellet)) were identical for fetal and adult human lung, and not stimulated by exogenous PG in rat or human. We were concerned about the possible effects of protein degradation post-mortem on the measurement of active cytidyiyltransferase, prompted by a comparison of enzyme activities in adult lung samples obtained at autopsy with those provided from surgical excisions. Delays inherent in the receipt of autopsy tissue, in the order of 12-36 h, apparently caused a 75% decrease in cytidylyltransferase activity, from 0.23 mU/mg protein to 0.05 mU/mg protein measured in the absence of PG. Any changes in cytidylyltransferase activity due to anoxia and post-mortem changes accompanying the typical 1 h delay between surgical excision and tissue processing could not be directly quantified. A partial control, however, was provided by cytidylyltransferase measurements in fetal human lung-derived fibroblasts. When processed without delay, fibroblast enzyme distribution was similar (74% soluble) and specific activity higher (4.8 mU/mg) than those measured in whole lung and there was no significant change in either parameter when fibroblasts were kept at 4°C under a nitrogen atmosphere for 12 h. Gel filtration of soluble cyti&Iyltransferase A representative gel filtration elution profile of soluble adult human lung cytidylyltransferase activity is shown in Fig. 1. The two peaks of activity observed agree with numerous previous reports for mammalian tissues [l] and were coincident with our own fractionation of rat lung enzyme. High M,, H form, activity was associated with void volume (V,) fractions, while a lower M,, L form, eluted ahead of haemoglobin with fractions of about 200 kDa. Unlike rat lung cytidylyltransferase, the sum of H and L forms from both soluble fetal and adult human lung fractions was consistently 200% greater than the total activity applied to
TABLE
H form
I
Il.
__~I___. 200
.L 300 Elution
400 vol
-.
(ml)
Fig. 1. Gel filtration of soluble adult human lung cytidylyltransferase. A 100000~ g for 60 min su~rnatant from adult human lung homogenate in buffer A (15 ml. 80 mU cytidyiyltransfera~) was applied to a Sepharose 6B column (90X 2.5 cm) and eluted in buffer A al 4OC. Fractions (10 ml) were collected and assayed for cytidylyltransferase it 0.25 mM PG.
the column (Fig. 1, Table II). Prior ageing by incubation at 37°C for 2 h (see below) increased the recovery of L form enzyme from both fetal and adult human lung, although adult lung H form enzyme activity fell. Moreover, it seems likely that additional enzyme activity was lost during the 12 h chromatographic separation, for although H form activity was stable at 4°C for 24 h, L form activity declined by more than 50% over the same period. It appeared that some inhibitor component was resolved from enzyme-containing fractions during chromatographic separation. Surprisingly, lipid stimulation was not required for activity measurement in either resolved fraction.
II
The effects of ageing on the recovery and H/L
f mm distribution
of sofubfe cytidy fy ftransferase jrom human lung
Cytidylyltransferase activities were determined in soluble fractions (15 ml) from tissue homogenates immediateIy following initial isolation at 100000X g for 60 min and after ageing at 37O C for 2 h. Distribution between H and L forms of enzyme were assessed by gel filtration on Sepharose 6B (90 x 2.5 cm in buffer A). For comparative purposes, recoveries were standardised against the maximum initial activity determined in the presence of 0.25 mM PG (100%). Each value represents the mean of two to four experiments. Cytidylyltransferase applied column
to 6B
activity
(% fresh activity+0.25
mM PG)
H form activity
L form activity
total recovered activity
Fetal human Soluble enzyme Aged2hat37OC
100 185
123 104
177 266
300 370
Adult human Soluble enzyme Aged2hat37OC
100 116
154 28
148 223
302 251
23 Effect of storage or incubation on soluble lung cytidylyltransferase Previous reports of liver cytidylyltransferase have also demonstrated considerable latent enzyme activity which could be revealed by a variety of procedures. Soluble rat liver cytidylyltransferase showed a 6-fold increasing following storage at 4°C for 5 days [31] or incubation at 20” C for 8 h [32]. The possibility that a similar process of ageing by storage might also reveal latent soluble human lung cytidylyltransferase activity was investigated using a combination of ageing protocols and centrifugation. Ageing at 4°C for 5 days resulted in only slight increases in soluble fetal and adult cytidylyltransferase to 0.47 and 0.40 mU/mg without PG and 0.93 and 0.62 mU/mg, respectively, in the presence of 0.25 mM PG, with no change in rat lung enzyme activity. In contrast, at 37” C for 2 h, a more pronounced effect was seen (Table I). Cytidylyltransferase activity in soluble fetal lung fractions increased almost 2-fold following incubation, whether measured in the presence or absence of PG. Ageing produced no significant change in soluble adult lung cytidylyltransferase in human or rat. An increase in turbidity was noted to accompany all ageing experiments at 37°C and its removal by a second centrifugation at 100000 x g for 60 min yielded a clear supernatant (2nd supernatant) and a protein-rich pellet. The principal protein component of this pellet from both human and rat, characterised below by SDSPAGE was actin, suggesting that the increased turbidity might reflect a soluble actin aggregation. The pellets contained approx. 5% of all initially soluble protein. Cytidylyltransferase activity in the 2nd supernatant from fetal lung was unchanged following centrifugation, but was significantly stimulated in the corresponding adult lung fraction suggesting the removal of an inhibitory component (Table I). The most dramatic effect was seen with rat lung enzyme, however, which lost 85% activity during this centrifugation step. An apparently
small activity associated with the actin-rich particulate fractions from human ageing experiments and a somewhat larger activity from those with rat lung could not be accurately quantified due to problems of effective dispersion. The co-purification of actin with H form cytidylyltransferase SDS-PAGE analysis of V, fractions from gel filtration of the soluble fraction of adult human lung disclosed the presence of a dominant band of 43 kDa which co-migrated with rabbit muscle actin (Fig. 2). We confirmed the identity of this protein as human lung actin by amino acid composition and Western blot analysis as described previously [26]. Assay of fractions eluted from the ion exchange chromatography of lung actin [26], showed a partial co-purification of cytidylyltransferase with this cytoskeletal component; 62% of recovered cytidylyltransferase activity was found in actin-containing fractions, while 38% was unretarded by DE 23 and not associated with actin. cytidylyltransferase co-eluted with this partially purified lung F-actin in V, fractions upon subsequent Sepharose 6B gel filtration (results not shown). Attempts to co-solubilise cytidylyltransferase and actin using SDS, previously shown to solubilise lung actin aggregates [26], were unsuccessful due to the irreversible inhibition of human lung enzyme under these conditions. This contrasts sharply with the rat liver enzyme which is apparently more stable in the presence of SDS and may be converted between H and L forms by this treatment [31]. Poly(ethyleneglycol)-dependent associations with lung actin The well-documented ability of PEG to promote both the aggregation of muscle actin and interactions of hetero-associated proteins [20,33], led us to use it in evaluation of potential actin/cytidylyltransferase interactions in vitro. Addition of PEG 6000 up to a final
.Actin
Fig. 2. SDS-PAGE homogenate. Tracks
analysis of void volume fractions from Sepharose 6B gel filtration of the soluble fraction 2-11 contained protein from column fractions 19-30 of Fig. 1 and tracks 1 and 14 contained The 10% polyacrylamide gel was stained with Coomassie Brilliant Blue.
from rabbit
an adult human lung muscle actin standard.
24
80
0
/
I 1
I
1-0 2
[PEG
6oocij
(7.)
4
Fig. 3. The effect of incubation of soluble rat lung cytidylyltransferase with PEG 6000 at concentrations up to 10%. Aliquots of soluble rat lung cytidylyltransferase (500 ~1) were adjusted to the required PEG concentration by mixture with stock PEG 6000 (10% or 20%) and allowed to stand for 2 h at 4°C. Supernatant and pellet fractions from centrifugation (12000x g for 2 min), were then assayed for cytidylyltransferase in the presence of 0.25 mM PC after re-suspension of pellets to the equivalent volume of supernatant with buffer A.
concentrations of 10% revealed that soluble lung fractions from human and rat incubated with 4% PEG precipitated more than 90% of soluble actin within 30 min at 4°C. Concentrations above 4% produced progressively greater recovery of other PEG-generated insoluble proteins. The activity of cytidylyltransferase in these mixtures before centrifugation was unaffected by the presence of PEG. We investigated soluble/particulate enzyme distribution, by centrifugation, following a 2 h incubation of soluble rat lung enzyme between 1% and 10% PEG at 4 o C, Fig. 3. Maximum recovery of particulate rat lung enzyme was achieved at 3% PEG with a corresponding
TABLE
fall in soluble enzyme when compared with PEG-free controls. At 3% PEG, 64% activity was recovered with the actin-rich pellet. The supernatant cytidylyltransferase appeared partially inhibited (22.5% of initial activity). When adjusted to 6% PEG and incubated for a further 30 min, however, the remaining activity became insoluble and the resuspended pellet contained 54% of initial activity. All the protein and cytidylyltransferase activity present in the 3% PEG co-precipitate was recovered in void volume fractions upon subsequent gel filtration. Partially purified rat lung L form cytidylyltransferase was not rendered insoluble by any PEG concentration tested up to 10%. The behaviour of the initially soluble human lung cytidylyltransferase in the presence of PEG 6000 differed from that of the rat, Table III. After incubation with 3% PEG, 38% of fetal and 3.3% of adult lung enzyme were associated with the insoluble pellet. The fetal 3% PEG supernatant contained negligible activity, but when adjusted to 6% PEG, produced a second, smaller actin-rich pellet containing 153% of initial soluble activity. In contrast, the adult 3% PEG supernatant contained 70% initial activity and this was rendered insoluble following adjustment to 6% PEG. As with rat lung enzyme fetal and adult human L form, cytidylyltransferase separated by gel filtration remained soluble at PEG concentrations up to 6%. Discussion
Analysis of the subcellular distribution of fetal and adult human lung cytidylyltransferase resolved soluble and particulate forms common to enzyme isolated from other mammalian sources [l]. With no significant varia-
III
Precrpltation
of soluble human and rub lung cytidylyltransferase
by PEG 6000
Aliquots of soluble cytidylyltransferase of purified H and L forms (500 ~1) were adjusted to 3 or 6% PEG by addition of 10% PEG stock. After 2 h at 4O C. pellets and supernatants were separated by centrifugation (12000 x g for 2 min). Activities in the presence of 0.25 mM PC are expressed as a fraction of uncentrifuged controls. Each value represents the mean of two to four experiments. Cytidylyltransferase
activity (W initial soluble enzyme)
3% PEG
L form enzyme
H form enzyme
soluble enzyme 3-6% PEG
3% PEG
6% PEG
3% PEG
6% PEG
Adult human Pellet Supernatant
3 70
80 0
1 100
93 0
0 100
0 100
Fetal human Pellet Supernatant
38 0
153 0
1 100
75 0
0 100
0 100
Adult rat Pellet Supernatant
64 23
54 0
79 0
0 0
0 100
0 100
25 tion in the proportions of the enzyme found in each fraction, a regulatory translocation similar to that accompanying maturation of the rat surfactant system [8] could not be confirmed. While human lung matures earlier in gestation than that of the rat making direct comparison difficult, fetal human lung at 15-16 weeks is both immature with respect to surfactant synthesis and comparable morphologically to rat lung at d18 [22]. Type II pneumocytes, the site of surfactant synthesis, comprise only about 16% of adult lung parenchyma [34] and changes within these cells alone may have been masked by the lack of overall tissue response. Equally, a transient redistribution in the perinatal period cannot be discounted. An earlier study of neonatal human lung cytidylyltransferase in unfractionated homogenates without lipid activation did not address this latter possibility [35]. A regulatory role for lipid stimulation of soluble lung cytidylyltransferase [9] was not established, since no significant variation in PG activation of soluble activity was seen between mature and immature lung. Lipid was present in soluble fractions, but our inabiIity to delipidate soluble human lung enzyme while retaining biological activity precluded further evaluation. The lack of a lipid requirement for L form activity, however, appears to argue against this possibility. The measurement of latent soluble cytidylyltransferase activity cast serious doubt on the extrapolation of fresh a~tivity/distribution figures. Clearly, this additional activity showed almost 90% of human lung enzyme to be soluble in vitro. A major change accompanying the disclosure of this activity was the aggregation of cytoskeletal actin. Rat lung actin aggregation under similar conditions did not reveal any additional cytidylyltransferase activity. Both cytidylyltransferase and actin distribution between soluble and particulate subcellular fractions are critically dependent on the ionic composition of the homogenising medium [5,6,14]. The ionic strengths of aqueous intracellular compartments, while unknown, may be very low [36]. Coupled with high-protein concentrations in vivo, this might favour the existence of particulate enzyme [5,6]. Ionic strength and dilution effects during tissue fractionation may artificially create translocation phenomena. Interestingly, buffer systems that purport to preserve and demonstrate phosphorylation effects on the reportedly soluble enzyme [32,37], are remarkably similar to the optimum conditions for fibroblast soluble G-actin recovery [14]. The use of PEG controlled one variable by effecting high-protein concentration after tissue fractionation. Aggregation of actin-rich cytoskeletal elements by PEG also rendered rat lung cytidylyltransferase insoluble in a concentration dependent fashion. Human lung enzyme/actin co-precipitation required a greater PEG concentration, perhaps reflecting a slightly weaker en-
zyme-actin interaction. At 6% however, the PEG concentration needed to co-precipitate actin and cytidylyltransferase is much smaller than the lo%-148 used to demonstrate actin/glycolytic enzyme interactions [21]. The greater apparent association of human lung enzyme with actin-rich fractions following PEG exposure compared with that after incubation at 37”C, may result from PEG’s mitigation of dilution effects. The PEGpromoted insolubility appeared to require the presence of actin, since L form cytidylyltransferase which lacked actin remained soluble in the presence of 6% PEG. The apparent heteroassociation of cytidylyltransferase and lung actin at high effective protein concentrations, similar to those that exist intracelIularly, raises the possibility of an in vivo linkage. Over 30 examples of enzyme/cytoskeleton associations have been described 1381, including many regulatory enzymes [17,39944]. It is known that hormones and growth factors regulate cytoskeletal assembly [45-481 and potentially the activities of associated enzymes, providing a possible framework for regulation of cytidylyltransferase. The significance of our observations to the situation in vivo is unclear. Recently, however, channeling of PC precursors in cultured glioma cells has been shown [13], implying a functional compartmentation of the Kennedy pathway enzymes. Together with increasing evidence of a highly organised aqueous cytoplasm [36,49,50], these results indicate that the possible existance of CDPcholine pathway multienzyme complexes, associated with the cytoskeleton in vivo, merits further investigation. Acknowledgement This work was supported by a project grant from the Children Nationwide Medical Research Fund. References 1 Pelech,
2 3 4 5 6 I 8 9 10 11
S.L. and Vance, D.E. (1984) Biochim. Biophys. Acta 779, 217-251. Kacser, H. and Porteous, J.W. (1987) Trends Biochem. Sci. 12, 5-14. Vance, D.E. and Pelech, S.L. (1984) Trends B&hem. Sci. 9, 75-77. Wilson, J.E. (1980) Curr. Top. Cell. Reg. 16, l-44. Schneider, W.C. (1963) J. Biol. Chem. 238, 3572-3578. Stern, W., Kovac, C. and Weinhold, P.A. (1976) B&him. Biophys. Acta 441, 280-293. Kyriazi, H.T. and Basford, R.E. (1986) Arch. Biochem. Biophys. 248, 253-271. Weinhold, P.A., Rounsifer. M.E., Williams, SE., Brubaker, P.G. and Feldman, D.A. (1984) J. Biol. Chem. 259, 10315-10321. Rooney, S.A. (7985) Am. Rev. Resp. Dis. 131, 439-460. Tesan, M., Ancheschi, M.M. and Bleasdale, J.E. (1985) Biochem. J. 232. 705-713. Chu, A.J. and Rooney, S.A. (1985) Biochim. Biophys. Acta 835, 132-140.
26 12 Pike, M.C., Kredich, N.M. and Snyderman, R. (1980) Cell 20, 313-379. 13 George, T.P., Morash, SC., Cook, H.W., Byers, D.M., Palmer, F.B.Q.C. and Spence, M.W. (1989) B&him. Biophys. Acta 1004, 283-291. 14 Bray, D. and Thomas, C. (1975) J. Mol. Biol. 105, 527-544. 15 Clarke, F.M. and Morton, D.J. (1982) B&hem. Biophys. Res. Commun. 109, 388-393. 16 Luther, M.A. and Lee, J.E. (1986) J. Biol. Chem. 261, 1753-1759. 17 Choate, G.L., Lan, L. and Mansour, T.E. (1985) J. Biol. Chem. 260, 4815-4822. 18 Swezy, R.R. and Epel, D. (1986) J. Cell. Biol. 103, 1509-1515. 19 Wolf, M. and Sahyoun, N. (1986) J. Biol. Chem. 261,13327-13332. 20 Ingham, K.C. (1984) Methods Enzymol. 104, 351-356. 21 Walsh, J.L. and Knull, H.R. (1987) Biochim. Biophys. Acta 952, 83-91. 22 Farrell, P.M. (1982) Lung Development: Biological and Clinical Perspectives, Vol. 1 (Farrell, P.M., ed.), pp. 13-25. Academic Press, New York. 23 Caeser, P.A., Wilson, S.J.. Normand, I.C.S. and Postle, A.D. (1988) Biochem. J. 253, 451-457. 24 Frosolono, M.F. (1982) Lung Development: Biological and Clinical Perspectives, Vol. 1 (Farrell, P.M., ed.), pp. 111-134, Academic Press, New York. 25 Feldman, D.A., Dietrich, J.W. and Weinhold, P.A. (1980) Biochim. Biophys. Acta 620, 603-611. 26 Postle, A.D., Hunt, A.N. and Normand, I.C.S. (1985) B&him. Biophys. Acta 837, 305-313. 27 Laemmli, U.K. (1970) Nature 227, 680-685. 28 Peterson, G.L. (1983) Methods Enzymol. 91, 95-119. 29 Bligh, E.G. and Dyer, W.S. (1959) Can. J. Biochem. 37, 911-923. 30 Bartlett, G.R. (1959) J. Biol. Chem. 234, 466-468. 31 Choy, P.C., Lim, P.H. and Vance, D.E. (1977) J. Biol. Chem. 252, 7673-7677.
32 Pelech, S.L. and Vance, D.E. (1982) J. Biol. Chem. 257, 14198-14202. 33 Tellam, R.L., Scully, M.J. and Nicol, L.W. (1983) B&hem. J. 213. 651-659. 34 Crapo, J.D., Barry, B.E., Gehr, P., Bachofen, M. and Weibel, E.R. (1982) Am. Rev. Resp. Dis. 125, 740-745. 35 Thorn, M.L. and Zachman, R.D. (1975) Pediat. Res. 9, 201-205. 36 Clegg, J.S. (1984) Am. J. Physiol. 246, R133-R151. 37 Radika, K. and Possmayer, F. (1985) Biochem. J. 232, 8333840. 38 Hunt, A.N. (1987) PhD Thesis, University of Southampton. 39 Volpe, J. (1979) J. Biol. Chem. 254, 2568-2571. 40 Dubose, D.A., Shepro, D. and Hechtman, H.B. (1987) Life Sci. 40, 447-453. 41 Reitman, J., Haines, K.A., Rich, A.M., Cristello, P., Giedd, K.N. and Weiseman, G. (1986) J. Immunol. 136, 1027-1032. 42 Al-Mohamma, F.A., Ohishi, I. and Hallett, M.B. (1987) FEBS Lett. 219, 40-44. 43 Leiser, M., Rubin, C.S. and Erlichman, J. (1986) J. Biol. Chem. 261, 1904-1908. 44 Ohta, Y., Nishida, E. and Sakai, H. (1986) FEBS Lett. 208, 423-426. 45 Ohta, Y., Akiyama, T., Nishida, E. and Sakai, H. (1987) FEBS Lett. 222, 305-310. 46 Rao, K.M.K., Betschart, J.N. and VirJi, M.A. (1985) Biochem. J. 230, 790-794. 47 Kadowaki, T., Koyasu, S., Nishida, E., Sakai, H., Takaku, F., Yahara, I. and Kasuga, M. (1986) J. Biol. Chem. 261,16141-16147. 48 Amsterdam, A., Rotmensch, S. and Ben-Ze’ev, A. (1989) Trends Biochem. Sci. 14, 377-382. 49 Srere, P.A. (1987) Annu. Rev. Biochem. 56, 89-124. 50 Srivastava, D.K. and Bernhard, S.A. (1987) Annu. Rev. Biophys. Biophys. Chem. 16, 1755204.
BBALIP
19
53343
CTP:cholinephosphate cytidylyltransferase in human and rat lung: Association in vitro with cytoskeletal actin Alan N. Hunt, I. Colin S. Normand and Anthony D. Postle Child Health, Faculty of Medicine, Southampton General Hospital, Southampton (U.K.) (Received
Key words:
CTP:Cholinephosphate
cytidylyltransferase;
6 October
1989)
Actin; Cytoskeletal (Human lung)
interaction;
Poly(ethyleneglyco1);
Enzyme activation;
CTPxholinephosphate cytidylyltransferase activities were compared in saline homogenates of immature fetal (15-16 weeks gestation) and adult human lung. There were no differences in subcellular enzyme distribution, in V,,, activity, or in the phosphatidylglycerol-mediated stimulation of soluble enzyme activity. These results provide no support for a developmental tram&cation of cytidylyltransferase from a cytosolic to a microsomal location in human lung, such as that proposed to accompany the maturation of pulmonary surfactant phosphatidylcholine biosynthesis in rat. Soluble cytidylyltransferase activity from human but not rat lung was increased after manipulation in vitro. Resolution of human H form (> lo3 kDa) and L form (200 kDa) enzyme by gel filtration led to an activity increase of 200%. Incubation at 37 o C for 2 h increased soluble enzyme recovery, although prior centrifugal removal of generated actin-rich aggregates was necessary in adult lung fractions. In contrast, 85% of soluble rat lung cytidylyltransferase was actin aggregate-associated after incubation. The apparent heteroassociation of rat and human lung enzyme with actin in the presence of poly(ethylene glycol) at 4” C strongly suggested close in vitro and potential in vivo linkage. A partial co-purification of adult human lung cytidylyltransferase with actin was also consistent with this idea. We propose that some reported cytidylyltransferase translocation phenomena may be mediated by cytoskeletal interactions in vitro.
Introduction
dependent
on the composition
of homogenising
buffers
[5,61. A regulatory role for CTP:cholinephosphate cytidylyltransferase (cytidylyltransferase) (EC 2.7.7.15) in the CDPcholine pathway of mammalian phosphatidylcholine (PC) biosynthesis is well established [l]. The claims of a rate-limiting function [l] imply that its control strength [2] may be near unity. Tissue fractionation following a variety of pathway flux manipulations have shown apparent enzyme redistribution between soluble and particulate subcellular compartments [3]. A regulatory translocation between an active, membraneassociated form and a less active soluble enzyme has been proposed [3]; a characteristic typical of enzymes described as ambiquitous [4]. The distribution of cytidylyltransferase between soluble and particulate fractions of liver and lung homogenates is critically
Abbreviations: PC, phosphatidylcholine; PG. phosphatidylglycerol.
PEG,
poly(ethyleneglyco1);
Correspondence: A.N. Hunt, Child Health, Faculty of Medicine, Southampton General Hospital, Tremona Road, Southampton, SO9 4XY, U.K. 00052760/90/$03.50
0 1990 Elsevier Science Publishers
B.V. (Biomedical
Translocation of brain hexokinase, the prototype ambiquitous enzyme [4], has recently been re-evaluated using rapid fractionation techniques, which failed to demonstrate any redistribution other than that of artefactual origin [7]. The significance of the apparent translocation of pulmonary cytidylyltransferase, including that accompanying the maturation of the surfactant system [S], has likewise been questioned [9] following observed flux-independent redistributions [lo]. Moreover, reported flux-dependent increases in the soluble, phospholipid-stimulated activity in the absence of redistribution phenomena are difficult to reconcile with the translocation hypothesis [ll]. PC synthesis in phagocytic cells, by both CDPcholine and N-methylation pathways, can be regulated by the state of microfilament and microtubule assembly, respectively [12]. Together with the recent demonstration of CDPcholine pathway compartmentalisation [13], this suggests some functional association of elements of the cytoskeleton with the enzymes of PC synthesis. We are unaware of similar observations with other non-phagocytic cells. The subcellular distribution of cytoskeletal Division)
20 components in vitro is also sensitive to fractionation conditions [14]; changes in such conditions can consequently affect associations of enzymes with cytoskeletal fractions in tissue homogenates. Such a variation in distribution is well documented for a number of glycolytic enzymes [15]. Several candidate ambiquitous enzymes have strong associations both in vivo and in vitro with elements of the cytoskeleton. These include phosphofructokinase [16,17], glucose-6-phosphate dehydrogenase [18] and Ca*+-phospholipid-dependent protein kinase C [19]. Cytoskeletal binding can result in changes both in K, and V,,,,, parameters [17,18]. Demonstrations of enzyme-cytoskeleton interactions in vitro are problematic due to the non-physiological dilution accompanying tissue fractionation. The large space-filling polymer poly(ethylene glycol) (PEG) has been employed to increase effective protein concentration in vitro and promote polymerisations and heteroassociations known to occur in vivo [20]. PEG has recently been used with pure and semipurified glycolytic enzyme preparations to show and evaluate enzyme/ actin and enzyme/enzyme associations at effective protein concentrations close to those encountered in vivo
1211. In this study, properties of cytidylyltransferase have been compared between fetal (15-16 weeks gestation) human lung, morphologically comparable to d18 rat lung [22], and adult human lung in order to assess any maturational variation in the parameters. Similarities and differences between human and rat lung enzyme have also been evaluated. These properties include subcellular distribution, extent of lipid stimulation and potential cytoskeletal interactions. Materials
Chemicals and biochemicals were obtained from BDH Chemicals (Poole, U.K.) or Sigma (Poole, U.K.). [ Me-‘4C]Choline phosphate was obtained from Amersham International (U.K.). Chromatography media were purchased from Pharmacia (Uppsala, Sweden). Fetal human lung was obtained from therapeutic abortion at 15-16 weeks gestation, and nonmalignant adult lung biopsies were taken during surgical procedure. Fetal human lung-derived fibroblasts were maintained in monolayer culture [23] and harvested in logarithmic phase growth at passage number 5-7. Rat lungs, female Wistar (150-200 g) were blanched in situ and lavaged (6 x 15 ml) with ice-cold saline to remove pulmonary surfactant [24]. Methods
Lungs and fibroblasts were homogenised in buffer A: 150 mM NaCl/SO mM Tris-HCl/l mM EDTA/l mM DTT/l mM NaN,/200 PM phenylmethylsulphonyl
fluoride (pH 7.4) at 4°C. Tissue was homogenised (20%, w/v) using an IKA-Ultra Turrax T18 homogeniser (Janke and Kunkel) and fibroblast homogenates (3-4%, w/v) were prepared by sonication on ice (3 x 20 s, 14 pm amplitude) using a Soniprep 150 with exponential microprobe (MSE Instruments, Crawley, U.K.). Initial centrifugation at 10000 x g for 10 min yielded a first pellet containing tar-like material in adult human lungs; this pellet exhibited negligible cytidylyltransferase activity in any tissue and was discarded. The supernatant was centrifuged at 100000 x g for 60 min to provide a membrane-rich particulate fraction and a soluble fraction. Lipid floating at the surface was removed by filtration through glass wool. Cytidylyltransferase assays monitored the conversion of [ Me-‘4C]choline phosphate (4726 dpm/nmol) to [ Me-‘4C]CDPcholine over 20 min at 37 o C. CDPcholine was isolated by charcoal binding [6], modified as detailed for use with 1.5 ml microfuge tubes. The final assay mix (90 ~1) contained enzyme, 60 mM TrisHC1/120 mM NaCl/l.4 mM EDTA/2.2 mM DTT/4 mM MgCl/4 mM CTP (pH 7.4) f 0.25 mM PG. Egg yolk phosphatidylglycerol (PG, Sigma) was sonicated directly into the assay buffer before use (3 x 20 s, 14 pm). The reaction was started with the addition of [Me-‘4C]choline phosphate (10 ~1. 16.6 mM, 2 ~Ci/~mol). The reaction was stopped by placing the tubes in boiling water for 2 min. Acid-washed charcoal (500 ~1) [25] was then added. Bound CDPcholine was separated by centrifugation (12 000 x g for 1 min) followed by 4 X 1 ml washes with distilled water. CDPcholine was eluted with ethanol/water/O.880 ammonia (60: 37 : 3, v/v) (2 X 1 ml, 90°C for 20 min) and radioactivity was determined by liquid scintillation in Cocktail T (BDH Chemicals). 1 mU of activity was defined as the conversion of 1 nmol choline phosphate to CDPcholine per min at 37°C. The reaction velocity was constant over 30 min and proportional to enzyme activity up to 20 mu/ml. Gel filtration on Sepharose 6B (90 X 2.5 cm) in buffer A at 4” C was monitored at 280 nm, using a flow-rate of 40 ml/h and 10 ml fraction collection. Cytidylyltransferse assays of column fractions were made singly f0.25 mM PG. Polyethylene glycol 6000 (BDH, mol.wt. 6000-7500) was prepared as a stock solution (10% or 20%) in buffer A and stored at 4 o C. Human lung actin was prepared by ammonium sulphate fractionation, gel filtration and ion-exchange chromatography as previously described [26] and its identity was confirmed by M,, amino acid composition and Western blot analysis with chicken anti-muscle actin. Discontinuous SDS-PAGE was performed on 10% slab gels with 6% stacking gel according to the method
21
PG or equivalent concentration of total lung lipid (results not shown). PG stimulation of soluble rat lung cytidylyltransferase was not as great as the 5-fold seen previously [25]. Some tissue lipid contamination seemed likely despite careful lavage to remove surfactant lipids. Delipidation with 10 vol. isopropyl ether appeared to confirm this with enzyme activity declining to 30% of non-delipidated fractions, while treatment with PG or whole lung lipid restored maximum activity. Measurement of lipid phosphorus in the soluble fractions of both adult and fetal human lung revealed a 13.5-fold difference in phospholipid concentration at 0.63 f. 0.19 mM (n = 4) and 46.6 k 8.5 PM (n = 4) respectively. In the case of the fetal enzyme, this level of endogenous lipid contamination seemed too small to account for the lack of difference in response to PG. The full extent of the phospholipid stimulation effect could not be accurately assessed, however, since delipidation of human soluble fractions using isopropyl ether irreversibly removed all measureable cytidylyltransferase activity. The apparent requirement of freshly isolated soluble human lung enzyme for some lipid to express activity was in contrast to the partially purified H and L forms, which needed no exogenous lipid and which in the case of L form enzyme fractions contained less than 1 I_LMlipid phosphorus. Particulate cytidylyltransferase activities ((10000 x g
of Laemmli [27] and stained with Coomassie Brilliant Blue. Protein was determined using a modified FolinCoicalteau reagent [28] with deoxycholate/ trichloroacetic acid co-precipitation. BSA (1 mg/ml) was used as a reference standard. The phospholipid phosphorus of lipid extracts [29] was determined according to the method of Bartlett [30]. Results Subcellular
distribution of human lung cytidylyltransferase
In common with most reports of cytidylyltransferase distribution in mammalian tissues, fetal and adult human lung enzyme activity was found associated with both soluble and particulate fractions of saline homogenates (Table I). The proportion of cytidylyltransferase activity in the soluble fraction of human lung homogenates was comparable for fetal and adult tissue; in each case the soluble (100 000 x g for 60 min supernatant) fraction contained about 66% of measurable activity. Rat lung cytidylyltransferase distribution following fractionation under similar conditions was also comparable. The extent of lipid stimulations of soluble human enzyme in the presence of 0.25 mM PG were similar (Table I), with no greater activity observed at 1.25 mM
TABLE
I
Cytidylyltransferase
activities in soluble and particulate fractions
of human and adult rat lung homogenates
Fresh fetal and adult human and adult rat lung was homogenised and separated into soluble and particulate components and cytidylyltransferase activities (* PC) determined as described in methods. After ageing at 37 o C for 2 h, soluble activity was measured before and after removal of aggregated protein by a 2nd 100000X 8 for 60 mm centrifugation. * Significantly different (P < 0.005) from initial soluble activity. Cytidylyltransferase
Fetal lung Without PC + 0.25 mM PC Distribution (% Total) Adult lung Without PC + 0.25 mM PC Distribution (% Total) Adult rat lung Without PC + 0.25 mM PC Distribution (% Total)
activity
(mU/mg
protein + S.E.)
particulate fraction
soluble fraction initial soluble
aged soluble
2nd supernatant
1.65 + 0.17 1.65 f 0.29 (n =16) 33.2 +4.8%
0.36 + 0.05 0.80 k 0.06 (n = 25) 66.8 *4.8%
* 0.65 f 0.08 * 1.48 * 0.26 (n=S) _
* 0.68 f 0.26 * 1.46 k 0.26 (n=6) _
1.58k0.39 1.88*0.44 (n = 5) 33.9 f7.5W
0.23 + 0.03 0.58 * 0.07 (n = 10) 66.1 *7.5%
0.48 f 0.20 0.67kO.15 (n = 3) _
* 1.23 + 0.30 * 1.3OkO.28 (n = 3)
3.21+ 0.92 3.05 k 0.61 (n = 8) 31.5 *3.3x
1.56 k 0.08 2.01*0.11 (n = 8) 68.5 *3.3%
1.88+_0.17 2.03 + 0.21 (n = 8) _
* 0.29 * 0.04 * 0.34+0.07 (n = 6) _
22 for 10 min) - (100000 X g for 60 min pellet)) were identical for fetal and adult human lung, and not stimulated by exogenous PG in rat or human. We were concerned about the possible effects of protein degradation post-mortem on the measurement of active cytidyiyltransferase, prompted by a comparison of enzyme activities in adult lung samples obtained at autopsy with those provided from surgical excisions. Delays inherent in the receipt of autopsy tissue, in the order of 12-36 h, apparently caused a 75% decrease in cytidylyltransferase activity, from 0.23 mU/mg protein to 0.05 mU/mg protein measured in the absence of PG. Any changes in cytidylyltransferase activity due to anoxia and post-mortem changes accompanying the typical 1 h delay between surgical excision and tissue processing could not be directly quantified. A partial control, however, was provided by cytidylyltransferase measurements in fetal human lung-derived fibroblasts. When processed without delay, fibroblast enzyme distribution was similar (74% soluble) and specific activity higher (4.8 mU/mg) than those measured in whole lung and there was no significant change in either parameter when fibroblasts were kept at 4°C under a nitrogen atmosphere for 12 h. Gel filtration of soluble cyti&Iyltransferase A representative gel filtration elution profile of soluble adult human lung cytidylyltransferase activity is shown in Fig. 1. The two peaks of activity observed agree with numerous previous reports for mammalian tissues [l] and were coincident with our own fractionation of rat lung enzyme. High M,, H form, activity was associated with void volume (V,) fractions, while a lower M,, L form, eluted ahead of haemoglobin with fractions of about 200 kDa. Unlike rat lung cytidylyltransferase, the sum of H and L forms from both soluble fetal and adult human lung fractions was consistently 200% greater than the total activity applied to
TABLE
H form
I
Il.
__~I___. 200
.L 300 Elution
400 vol
-.
(ml)
Fig. 1. Gel filtration of soluble adult human lung cytidylyltransferase. A 100000~ g for 60 min su~rnatant from adult human lung homogenate in buffer A (15 ml. 80 mU cytidyiyltransfera~) was applied to a Sepharose 6B column (90X 2.5 cm) and eluted in buffer A al 4OC. Fractions (10 ml) were collected and assayed for cytidylyltransferase it 0.25 mM PG.
the column (Fig. 1, Table II). Prior ageing by incubation at 37°C for 2 h (see below) increased the recovery of L form enzyme from both fetal and adult human lung, although adult lung H form enzyme activity fell. Moreover, it seems likely that additional enzyme activity was lost during the 12 h chromatographic separation, for although H form activity was stable at 4°C for 24 h, L form activity declined by more than 50% over the same period. It appeared that some inhibitor component was resolved from enzyme-containing fractions during chromatographic separation. Surprisingly, lipid stimulation was not required for activity measurement in either resolved fraction.
II
The effects of ageing on the recovery and H/L
f mm distribution
of sofubfe cytidy fy ftransferase jrom human lung
Cytidylyltransferase activities were determined in soluble fractions (15 ml) from tissue homogenates immediateIy following initial isolation at 100000X g for 60 min and after ageing at 37O C for 2 h. Distribution between H and L forms of enzyme were assessed by gel filtration on Sepharose 6B (90 x 2.5 cm in buffer A). For comparative purposes, recoveries were standardised against the maximum initial activity determined in the presence of 0.25 mM PG (100%). Each value represents the mean of two to four experiments. Cytidylyltransferase applied column
to 6B
activity
(% fresh activity+0.25
mM PG)
H form activity
L form activity
total recovered activity
Fetal human Soluble enzyme Aged2hat37OC
100 185
123 104
177 266
300 370
Adult human Soluble enzyme Aged2hat37OC
100 116
154 28
148 223
302 251
23 Effect of storage or incubation on soluble lung cytidylyltransferase Previous reports of liver cytidylyltransferase have also demonstrated considerable latent enzyme activity which could be revealed by a variety of procedures. Soluble rat liver cytidylyltransferase showed a 6-fold increasing following storage at 4°C for 5 days [31] or incubation at 20” C for 8 h [32]. The possibility that a similar process of ageing by storage might also reveal latent soluble human lung cytidylyltransferase activity was investigated using a combination of ageing protocols and centrifugation. Ageing at 4°C for 5 days resulted in only slight increases in soluble fetal and adult cytidylyltransferase to 0.47 and 0.40 mU/mg without PG and 0.93 and 0.62 mU/mg, respectively, in the presence of 0.25 mM PG, with no change in rat lung enzyme activity. In contrast, at 37” C for 2 h, a more pronounced effect was seen (Table I). Cytidylyltransferase activity in soluble fetal lung fractions increased almost 2-fold following incubation, whether measured in the presence or absence of PG. Ageing produced no significant change in soluble adult lung cytidylyltransferase in human or rat. An increase in turbidity was noted to accompany all ageing experiments at 37°C and its removal by a second centrifugation at 100000 x g for 60 min yielded a clear supernatant (2nd supernatant) and a protein-rich pellet. The principal protein component of this pellet from both human and rat, characterised below by SDSPAGE was actin, suggesting that the increased turbidity might reflect a soluble actin aggregation. The pellets contained approx. 5% of all initially soluble protein. Cytidylyltransferase activity in the 2nd supernatant from fetal lung was unchanged following centrifugation, but was significantly stimulated in the corresponding adult lung fraction suggesting the removal of an inhibitory component (Table I). The most dramatic effect was seen with rat lung enzyme, however, which lost 85% activity during this centrifugation step. An apparently
small activity associated with the actin-rich particulate fractions from human ageing experiments and a somewhat larger activity from those with rat lung could not be accurately quantified due to problems of effective dispersion. The co-purification of actin with H form cytidylyltransferase SDS-PAGE analysis of V, fractions from gel filtration of the soluble fraction of adult human lung disclosed the presence of a dominant band of 43 kDa which co-migrated with rabbit muscle actin (Fig. 2). We confirmed the identity of this protein as human lung actin by amino acid composition and Western blot analysis as described previously [26]. Assay of fractions eluted from the ion exchange chromatography of lung actin [26], showed a partial co-purification of cytidylyltransferase with this cytoskeletal component; 62% of recovered cytidylyltransferase activity was found in actin-containing fractions, while 38% was unretarded by DE 23 and not associated with actin. cytidylyltransferase co-eluted with this partially purified lung F-actin in V, fractions upon subsequent Sepharose 6B gel filtration (results not shown). Attempts to co-solubilise cytidylyltransferase and actin using SDS, previously shown to solubilise lung actin aggregates [26], were unsuccessful due to the irreversible inhibition of human lung enzyme under these conditions. This contrasts sharply with the rat liver enzyme which is apparently more stable in the presence of SDS and may be converted between H and L forms by this treatment [31]. Poly(ethyleneglycol)-dependent associations with lung actin The well-documented ability of PEG to promote both the aggregation of muscle actin and interactions of hetero-associated proteins [20,33], led us to use it in evaluation of potential actin/cytidylyltransferase interactions in vitro. Addition of PEG 6000 up to a final
.Actin
Fig. 2. SDS-PAGE homogenate. Tracks
analysis of void volume fractions from Sepharose 6B gel filtration of the soluble fraction 2-11 contained protein from column fractions 19-30 of Fig. 1 and tracks 1 and 14 contained The 10% polyacrylamide gel was stained with Coomassie Brilliant Blue.
from rabbit
an adult human lung muscle actin standard.
24
80
0
/
I 1
I
1-0 2
[PEG
6oocij
(7.)
4
Fig. 3. The effect of incubation of soluble rat lung cytidylyltransferase with PEG 6000 at concentrations up to 10%. Aliquots of soluble rat lung cytidylyltransferase (500 ~1) were adjusted to the required PEG concentration by mixture with stock PEG 6000 (10% or 20%) and allowed to stand for 2 h at 4°C. Supernatant and pellet fractions from centrifugation (12000x g for 2 min), were then assayed for cytidylyltransferase in the presence of 0.25 mM PC after re-suspension of pellets to the equivalent volume of supernatant with buffer A.
concentrations of 10% revealed that soluble lung fractions from human and rat incubated with 4% PEG precipitated more than 90% of soluble actin within 30 min at 4°C. Concentrations above 4% produced progressively greater recovery of other PEG-generated insoluble proteins. The activity of cytidylyltransferase in these mixtures before centrifugation was unaffected by the presence of PEG. We investigated soluble/particulate enzyme distribution, by centrifugation, following a 2 h incubation of soluble rat lung enzyme between 1% and 10% PEG at 4 o C, Fig. 3. Maximum recovery of particulate rat lung enzyme was achieved at 3% PEG with a corresponding
TABLE
fall in soluble enzyme when compared with PEG-free controls. At 3% PEG, 64% activity was recovered with the actin-rich pellet. The supernatant cytidylyltransferase appeared partially inhibited (22.5% of initial activity). When adjusted to 6% PEG and incubated for a further 30 min, however, the remaining activity became insoluble and the resuspended pellet contained 54% of initial activity. All the protein and cytidylyltransferase activity present in the 3% PEG co-precipitate was recovered in void volume fractions upon subsequent gel filtration. Partially purified rat lung L form cytidylyltransferase was not rendered insoluble by any PEG concentration tested up to 10%. The behaviour of the initially soluble human lung cytidylyltransferase in the presence of PEG 6000 differed from that of the rat, Table III. After incubation with 3% PEG, 38% of fetal and 3.3% of adult lung enzyme were associated with the insoluble pellet. The fetal 3% PEG supernatant contained negligible activity, but when adjusted to 6% PEG, produced a second, smaller actin-rich pellet containing 153% of initial soluble activity. In contrast, the adult 3% PEG supernatant contained 70% initial activity and this was rendered insoluble following adjustment to 6% PEG. As with rat lung enzyme fetal and adult human L form, cytidylyltransferase separated by gel filtration remained soluble at PEG concentrations up to 6%. Discussion
Analysis of the subcellular distribution of fetal and adult human lung cytidylyltransferase resolved soluble and particulate forms common to enzyme isolated from other mammalian sources [l]. With no significant varia-
III
Precrpltation
of soluble human and rub lung cytidylyltransferase
by PEG 6000
Aliquots of soluble cytidylyltransferase of purified H and L forms (500 ~1) were adjusted to 3 or 6% PEG by addition of 10% PEG stock. After 2 h at 4O C. pellets and supernatants were separated by centrifugation (12000 x g for 2 min). Activities in the presence of 0.25 mM PC are expressed as a fraction of uncentrifuged controls. Each value represents the mean of two to four experiments. Cytidylyltransferase
activity (W initial soluble enzyme)
3% PEG
L form enzyme
H form enzyme
soluble enzyme 3-6% PEG
3% PEG
6% PEG
3% PEG
6% PEG
Adult human Pellet Supernatant
3 70
80 0
1 100
93 0
0 100
0 100
Fetal human Pellet Supernatant
38 0
153 0
1 100
75 0
0 100
0 100
Adult rat Pellet Supernatant
64 23
54 0
79 0
0 0
0 100
0 100
25 tion in the proportions of the enzyme found in each fraction, a regulatory translocation similar to that accompanying maturation of the rat surfactant system [8] could not be confirmed. While human lung matures earlier in gestation than that of the rat making direct comparison difficult, fetal human lung at 15-16 weeks is both immature with respect to surfactant synthesis and comparable morphologically to rat lung at d18 [22]. Type II pneumocytes, the site of surfactant synthesis, comprise only about 16% of adult lung parenchyma [34] and changes within these cells alone may have been masked by the lack of overall tissue response. Equally, a transient redistribution in the perinatal period cannot be discounted. An earlier study of neonatal human lung cytidylyltransferase in unfractionated homogenates without lipid activation did not address this latter possibility [35]. A regulatory role for lipid stimulation of soluble lung cytidylyltransferase [9] was not established, since no significant variation in PG activation of soluble activity was seen between mature and immature lung. Lipid was present in soluble fractions, but our inabiIity to delipidate soluble human lung enzyme while retaining biological activity precluded further evaluation. The lack of a lipid requirement for L form activity, however, appears to argue against this possibility. The measurement of latent soluble cytidylyltransferase activity cast serious doubt on the extrapolation of fresh a~tivity/distribution figures. Clearly, this additional activity showed almost 90% of human lung enzyme to be soluble in vitro. A major change accompanying the disclosure of this activity was the aggregation of cytoskeletal actin. Rat lung actin aggregation under similar conditions did not reveal any additional cytidylyltransferase activity. Both cytidylyltransferase and actin distribution between soluble and particulate subcellular fractions are critically dependent on the ionic composition of the homogenising medium [5,6,14]. The ionic strengths of aqueous intracellular compartments, while unknown, may be very low [36]. Coupled with high-protein concentrations in vivo, this might favour the existence of particulate enzyme [5,6]. Ionic strength and dilution effects during tissue fractionation may artificially create translocation phenomena. Interestingly, buffer systems that purport to preserve and demonstrate phosphorylation effects on the reportedly soluble enzyme [32,37], are remarkably similar to the optimum conditions for fibroblast soluble G-actin recovery [14]. The use of PEG controlled one variable by effecting high-protein concentration after tissue fractionation. Aggregation of actin-rich cytoskeletal elements by PEG also rendered rat lung cytidylyltransferase insoluble in a concentration dependent fashion. Human lung enzyme/actin co-precipitation required a greater PEG concentration, perhaps reflecting a slightly weaker en-
zyme-actin interaction. At 6% however, the PEG concentration needed to co-precipitate actin and cytidylyltransferase is much smaller than the lo%-148 used to demonstrate actin/glycolytic enzyme interactions [21]. The greater apparent association of human lung enzyme with actin-rich fractions following PEG exposure compared with that after incubation at 37”C, may result from PEG’s mitigation of dilution effects. The PEGpromoted insolubility appeared to require the presence of actin, since L form cytidylyltransferase which lacked actin remained soluble in the presence of 6% PEG. The apparent heteroassociation of cytidylyltransferase and lung actin at high effective protein concentrations, similar to those that exist intracelIularly, raises the possibility of an in vivo linkage. Over 30 examples of enzyme/cytoskeleton associations have been described 1381, including many regulatory enzymes [17,39944]. It is known that hormones and growth factors regulate cytoskeletal assembly [45-481 and potentially the activities of associated enzymes, providing a possible framework for regulation of cytidylyltransferase. The significance of our observations to the situation in vivo is unclear. Recently, however, channeling of PC precursors in cultured glioma cells has been shown [13], implying a functional compartmentation of the Kennedy pathway enzymes. Together with increasing evidence of a highly organised aqueous cytoplasm [36,49,50], these results indicate that the possible existance of CDPcholine pathway multienzyme complexes, associated with the cytoskeleton in vivo, merits further investigation. Acknowledgement This work was supported by a project grant from the Children Nationwide Medical Research Fund. References 1 Pelech,
2 3 4 5 6 I 8 9 10 11
S.L. and Vance, D.E. (1984) Biochim. Biophys. Acta 779, 217-251. Kacser, H. and Porteous, J.W. (1987) Trends Biochem. Sci. 12, 5-14. Vance, D.E. and Pelech, S.L. (1984) Trends B&hem. Sci. 9, 75-77. Wilson, J.E. (1980) Curr. Top. Cell. Reg. 16, l-44. Schneider, W.C. (1963) J. Biol. Chem. 238, 3572-3578. Stern, W., Kovac, C. and Weinhold, P.A. (1976) B&him. Biophys. Acta 441, 280-293. Kyriazi, H.T. and Basford, R.E. (1986) Arch. Biochem. Biophys. 248, 253-271. Weinhold, P.A., Rounsifer. M.E., Williams, SE., Brubaker, P.G. and Feldman, D.A. (1984) J. Biol. Chem. 259, 10315-10321. Rooney, S.A. (7985) Am. Rev. Resp. Dis. 131, 439-460. Tesan, M., Ancheschi, M.M. and Bleasdale, J.E. (1985) Biochem. J. 232. 705-713. Chu, A.J. and Rooney, S.A. (1985) Biochim. Biophys. Acta 835, 132-140.
26 12 Pike, M.C., Kredich, N.M. and Snyderman, R. (1980) Cell 20, 313-379. 13 George, T.P., Morash, SC., Cook, H.W., Byers, D.M., Palmer, F.B.Q.C. and Spence, M.W. (1989) B&him. Biophys. Acta 1004, 283-291. 14 Bray, D. and Thomas, C. (1975) J. Mol. Biol. 105, 527-544. 15 Clarke, F.M. and Morton, D.J. (1982) B&hem. Biophys. Res. Commun. 109, 388-393. 16 Luther, M.A. and Lee, J.E. (1986) J. Biol. Chem. 261, 1753-1759. 17 Choate, G.L., Lan, L. and Mansour, T.E. (1985) J. Biol. Chem. 260, 4815-4822. 18 Swezy, R.R. and Epel, D. (1986) J. Cell. Biol. 103, 1509-1515. 19 Wolf, M. and Sahyoun, N. (1986) J. Biol. Chem. 261,13327-13332. 20 Ingham, K.C. (1984) Methods Enzymol. 104, 351-356. 21 Walsh, J.L. and Knull, H.R. (1987) Biochim. Biophys. Acta 952, 83-91. 22 Farrell, P.M. (1982) Lung Development: Biological and Clinical Perspectives, Vol. 1 (Farrell, P.M., ed.), pp. 13-25. Academic Press, New York. 23 Caeser, P.A., Wilson, S.J.. Normand, I.C.S. and Postle, A.D. (1988) Biochem. J. 253, 451-457. 24 Frosolono, M.F. (1982) Lung Development: Biological and Clinical Perspectives, Vol. 1 (Farrell, P.M., ed.), pp. 111-134, Academic Press, New York. 25 Feldman, D.A., Dietrich, J.W. and Weinhold, P.A. (1980) Biochim. Biophys. Acta 620, 603-611. 26 Postle, A.D., Hunt, A.N. and Normand, I.C.S. (1985) B&him. Biophys. Acta 837, 305-313. 27 Laemmli, U.K. (1970) Nature 227, 680-685. 28 Peterson, G.L. (1983) Methods Enzymol. 91, 95-119. 29 Bligh, E.G. and Dyer, W.S. (1959) Can. J. Biochem. 37, 911-923. 30 Bartlett, G.R. (1959) J. Biol. Chem. 234, 466-468. 31 Choy, P.C., Lim, P.H. and Vance, D.E. (1977) J. Biol. Chem. 252, 7673-7677.
32 Pelech, S.L. and Vance, D.E. (1982) J. Biol. Chem. 257, 14198-14202. 33 Tellam, R.L., Scully, M.J. and Nicol, L.W. (1983) B&hem. J. 213. 651-659. 34 Crapo, J.D., Barry, B.E., Gehr, P., Bachofen, M. and Weibel, E.R. (1982) Am. Rev. Resp. Dis. 125, 740-745. 35 Thorn, M.L. and Zachman, R.D. (1975) Pediat. Res. 9, 201-205. 36 Clegg, J.S. (1984) Am. J. Physiol. 246, R133-R151. 37 Radika, K. and Possmayer, F. (1985) Biochem. J. 232, 8333840. 38 Hunt, A.N. (1987) PhD Thesis, University of Southampton. 39 Volpe, J. (1979) J. Biol. Chem. 254, 2568-2571. 40 Dubose, D.A., Shepro, D. and Hechtman, H.B. (1987) Life Sci. 40, 447-453. 41 Reitman, J., Haines, K.A., Rich, A.M., Cristello, P., Giedd, K.N. and Weiseman, G. (1986) J. Immunol. 136, 1027-1032. 42 Al-Mohamma, F.A., Ohishi, I. and Hallett, M.B. (1987) FEBS Lett. 219, 40-44. 43 Leiser, M., Rubin, C.S. and Erlichman, J. (1986) J. Biol. Chem. 261, 1904-1908. 44 Ohta, Y., Nishida, E. and Sakai, H. (1986) FEBS Lett. 208, 423-426. 45 Ohta, Y., Akiyama, T., Nishida, E. and Sakai, H. (1987) FEBS Lett. 222, 305-310. 46 Rao, K.M.K., Betschart, J.N. and VirJi, M.A. (1985) Biochem. J. 230, 790-794. 47 Kadowaki, T., Koyasu, S., Nishida, E., Sakai, H., Takaku, F., Yahara, I. and Kasuga, M. (1986) J. Biol. Chem. 261,16141-16147. 48 Amsterdam, A., Rotmensch, S. and Ben-Ze’ev, A. (1989) Trends Biochem. Sci. 14, 377-382. 49 Srere, P.A. (1987) Annu. Rev. Biochem. 56, 89-124. 50 Srivastava, D.K. and Bernhard, S.A. (1987) Annu. Rev. Biophys. Biophys. Chem. 16, 1755204.
