FEBS Letters 581 (2007) 102–108

Characterisation of lipofuscin-like lysosomal inclusion bodies from human placenta

Bernd Schro¨dera,*, Hans-Peter Elsa¨sserb, Bernhard Schmidtc, Andrej Hasilika a Institute of Physiological Chemistry, Philipps-University Marburg, Karl-von-Frisch-Straße 1, 35032 Marburg, Germany b Institute of Cytobiology and Cytopathology, Philipps-University Marburg, Robert-Koch-Straße 6, 35032 Marburg, Germany c Department of Biochemistry II, University Go¨ttingen, Heinrich-Du¨ker-Weg 12, 37077 Go¨ttingen, Germany

Received 13 October 2006; revised 28 November 2006; accepted 5 December 2006

Available online 12 December 2006

Edited by Judit Ova´di

to 480 nm and emission between 460 and 630 nm [2]. These Abstract A structural hallmark of is heterogeneity of their contents. We describe a method for isolation of parti- pigments are frequently observed in senescent postmitotic cells culate materials from human placental lysosomes. After a like neurons or cardiomyocytes and are considered to be unde- methionine methyl ester-induced disruption of lysosomes and gradable residues of incomplete intralysosomal breakdown [3]. two density gradient centrifugations we obtained a homogeneous Protein modifications by products of lipid peroxidation and membrane fraction and another one enriched in particulate inclu- crosslinking of proteins are discussed as reasons for the resis- sions. The latter exhibited a yellow-brown coloration and con- tance towards proteinolytic degradation [4]. tained bodies lacking a delimiting membrane, which were We disrupted lysosomes prepared from human placenta characterised by a granular pattern and high electron density. using a hydrolysable substrate and isolated two particulate The lipofuscin-like inclusion materials were rich in tripeptidyl fractions, the lysosomal membranes and aggregated lysosomal peptidase I, b-glucuronidase, acid and apolipoprotein inclusions, by two subsequent density gradient centrifugations. D and contained proteins originating from diverse subcellular localisations. The fraction containing the lipofuscin-like inclusions was Here we show that human term placenta contains lipofuscin- characterised by a proteomic analysis. like lysosomal inclusions, a phenomenon usually associated with senescence in postmitotic cells. These findings imply that a sim- ple pelleting of a lysosomal lysate is not appropriate for the iso- 2. Materials and methods lation of lysosomal membranes, as the inclusions tend to be sedimented with the membranes. 2006 Federation of European Biochemical Societies. Published 2.1. Subcellular fractionation Fresh normal human term placentas were obtained from local by Elsevier B.V. All rights reserved. hospitals and transported on ice. Placental tissue was homogenised and fractionated by a combination of differential centrifugation and Keywords: Ageing; Lysosomal membrane; Lipofuscin; Percoll density gradient centrifugation as described previously [5]. Placenta; I A -enriched pool (‘‘dense pool’’) was collected from the bot- tom third of the gradient. ‘‘Dense pool’’ (30 ml) was diluted ninefold with isotonic Hepes buffer (250 mM sucrose, 10 mM Hepes-NaOH, 1 mM EDTA, pH 7.2) and centrifuged for 20 min at 11500 rpm (15800 · gmax) in an SS-34 rotor (Sorvall, Bad Homburg, Germany). The sediment was mixed with isotonic Hepes buffer and 100 mM 1. Introduction methionine methyl ester (MME) in this buffer to obtain 18 ml and a final concentration of 20 mM MME. The mixture was incubated for 30 min at room temperature to disrupt the lysosomes [6] and then ad- Lysosomes represent a degradative organelle of eukaryotic justed to the following concentrations of proteinase inhibitors: 0.5 mM cells. They are responsible for breakdown of macromolecules iodoacetamide, 10 lM leupeptin, 1 mM phenylmethanesulfonyl known to be delivered to these organelles by endo-/phagocyto- fluoride and 1 lM Z-Phe-Phe-diazomethylketone and pepstatin A. sis and autophagocytosis. Near complete degradation is The sample was split and layered over two 28 ml sucrose gradients pre- pared from solutions of 32.5% and 55% (w/v) sucrose in 10 mM Hepes- achieved by the action of more than 50 , mostly hydro- NaOH, 1 mM EDTA, pH 7.2. The gradients were centrifuged for 12 h lases [1]. Defects in these enzymes or in transporters releasing at 25000 rpm (112700 · gmax) in a SW-28 rotor (Beckman, Mu¨nchen, the hydrolytic products into the cytosol result in lysosomal Germany). Sixteen fractions, 2.3 ml each, were collected from the top. storage. Under physiological conditions intralysosomally The two fractions with the highest b- activity were stored materials are summarised under the term lipofuscin combined (‘‘sucrose gradient pool’’) and further fractionated in a self-forming iodixanol gradient. An isotonic iodixanol working solu- and have been associated with ageing [2]. Characteristic fea- tion was prepared according to manufacturer’s instructions containing tures of lipofuscin are yellow-brown coloration and autofluo- 50% (w/v) iodixanol, 42 mM sucrose, 10 mM Hepes-NaOH, 1 mM rescence with excitation wavelengths in the range from 320 EDTA, pH 7.2, which was diluted with isotonic Hepes buffer to obtain solutions with lower iodixanol concentrations. Per gradient 4.4 ml ‘‘sucrose gradient pool’’ were mixed with 5.4 ml 50% iodixanol work- *Corresponding author. Fax: +49 6421 28 66957. ing solution and 11.7 ml 10 mM Hepes-NaOH, 1 mM EDTA, pH 7.2, E-mail address: bernd.schroeder@staff.uni-marburg.de (B. Schro¨der). and transferred into an Ultracrimp (Sorvall) polyallomer tube. A cushion of 5 ml 40% (w/v) iodixanol solution was layered below and Abbreviations: 2D-PAGE, two-dimensional gel electrophoresis; CTAB, 6 ml of 12.5% (w/v) iodixanol solution and 2.5 ml isotonic Hepes buffer N-cetyl-N,N,N-trimethylammonium bromide; MME, methionine above the sample. The gradients were centrifuged in a TV860 vertical methyl ester; TPPI, rotor (Sorvall) for 2 h at 50000 rpm (236500 · gmax) and divided into

0014-5793/$32.00 2006 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2006.12.005 B. Schro¨der et al. / FEBS Letters 581 (2007) 102–108 103

16 fractions a` 2 ml and a final fraction of 3 ml from top to bottom. For search engine in NCBI and SwissProt protein databases. Protein further analyses fractions 1–4 and 14–16 were pooled and will be re- hits with Mascot scores higher than 75 were considered to be signifi- ferred to as ‘‘iodixanol upper and lower pools’’. From these pools cant. membranes and aggregated lysosomal inclusions were recovered by dilution of the density gradient medium with 10 mM Tris–HCl, pH 7.4, and sedimentation in a 60Ti rotor (Beckman) for 2 h at 3. Results 250000 · gmax followed by one step of washing in the same buffer under the same conditions of centrifugation. 3.1. Fractionation of placental lysosomes 2.2. Determination of activity and protein concentration The fractionation workflow is shown in Fig. 1. The ‘‘dense The activities of b-glucocerebrosidase [7] and succinate dehydroge- pool’’ recovered from the Percoll gradient was enriched 45- nase [5] were determined at 37 C using spectrophotometric methods. fold in b- and 7.5-fold in succinate dehydroge- Tripeptidyl peptidase I (TPPI) activity was assayed in the presence nase activity (n ¼ 8) as compared to the homogenate. Electron of 0.25 mM fluorogenic substrate Ala-Ala-Phe-aminomethylcoumarin, 0.1% (w/v) Triton X-100, 0.1 M sodium acetate, pH 3.5, in a total volume of 30 ll. Reaction was carried out for 20 min at 37 C and ter- minated by the addition of 70 ll 0.4 M glycine–NaOH, pH 10.4. The product, 7-amino-4-methylcoumarin, was quantified fluorometrically (kex ¼ 355 nm, kem ¼ 460 nm). Protein concentration was determined by the method of Bradford [8]. Fluorescence of isolated intralysosomal aggregates was examined in a SFM 25 spectralfluorometer (BioTek instruments, Bad Friedrichshall, Germany). At several excitation wavelengths ranging from 300 to 600 nm emission spectra were re- corded in the range from kex to 700 nm.

2.3. Polyacrylamide gel electrophoresis and Western blotting Sample preparation and separation in SDS–PAGE was carried out according to Laemmli [9]. Whereas aliquots (15 ll) of fractions from sucrose gradient were subjected to SDS–PAGE directly, aliquots (150 ll) of iodixanol gradient fractions were precipitated with trichlo- roacetic acid (10% (w/v) final concentration) prior to solubilisation. Protein was visualised with silver nitrate [10]. For immunoblotting electrophoretically separated proteins were transferred onto PVDF membranes. Membranes were probed with anti-human LAMP-2 monoclonal antibody [11] or rabbit antibody raised against thiore- doxin fusion protein of C-terminal fragment of human TPPI (from Dr. E. Kominami, Juntendo University, Tokyo) and HRP-coupled secondary antibodies (Bio-Rad, Munich, Germany) that were visualised using ECL Plus Kit (GE Healthcare Life Sciences, Freiburg i. Br., Germany).

2.4. Electron microscopy For electron microscopy pooled fractions from the gradient were mixed with the same volume of a fixation solution [12] containing 2.5% glutaraldehyde, 2.5% paraformaldehyde, 0.05% (w/v) picric acid in 0.1 M cacodylate buffer (pH 7.35). Samples were incubated on ice for 1 h and subsequently sedimented at 90600 · gmax for 1 h. The pel- lets were washed three times in 0.1 M cacodylate buffer and postfixed in 2% OsO4 for 1 h. Standard procedures for dehydration and embed- ding in Epon were employed. Thin sections were stained with uranyl acetate and lead citrate and examined with a Zeiss EM109 electron microscope.

2.5. Diagonal two-dimensional PAGE and mass spectrometric identification of spots Electrophoretic separation of proteins combining a N-cetyl-N,N,N- trimethylammonium bromide (CTAB)–PAGE in the first with an SDS–PAGE in the second dimension was performed as described pre- viously [13]. The samples were preincubated in the presence of 50 mM Tris–HCl, pH 8.5, 1% (w/v) CTAB, 30 mM DTT for 10 min at 60 C. Then 2 volumes of 4.5 M urea, 1% (w/v) CTAB, 96 mM KOH, 141 mM acetic acid, pH 5.1, 7.5% (v/v) glycerol were added and the samples were heated for 5 min at 60 C. Protein was stained with colloidal Coomassie brilliant blue according to Neuhoff et al. [14]. To identify proteins, gel plugs (2 mm diameter) were excised and sub- jected to tryptic in gel digestion according to standard protocols. Pep- tides were applied manually to a MALDI sample support precoated with a-cyano-4-hydroxycinnamic acid and mass spectra acquired with an Ultraflex I MALDI-TOF/TOF mass spectrometer (Bruker - ics, Bremen, Germany) in an automated mode as described by Jahn et al. [15]. For each sample a peptide mass fingerprint and four MS/MS spectra were recorded that were used for a combined database Fig. 1. Workflow for purification of lysosomal membranes and search with BioTools 3.0 software (Bruker Daltonics) and the Mascot inclusions. 104 B. Schro¨der et al. / FEBS Letters 581 (2007) 102–108 microscopic examination of ‘‘dense pool’’ confirmed lysosomes and mitochondria to be the major constituents (Fig. 4A). The contents of lysosomes varied from homogeneously electron dense lysosomal matrices with a dotted and finely granulated pattern to amorphous clumps of high electron density and transparent vacuolar inclusions. After incubating the ‘‘dense pool’’ lysosomes with methionine methyl ester the disrupted membranes were sepa- rated from mitochondria in a sucrose gradient. In a represen- tative gradient shown in Fig. 2 the sublysosomal membrane vesicles characterised by the presence of b-glucocerebrosidase activity (Fig. 2A) and LAMP-2 protein (Fig. 2B) were enriched in fraction 5 (mean density 1.15 g/cm3). The mitochondria (succinate dehydrogenase activity) band at a higher density with a maximum in fraction 9. If the substrate-induced lyso- somal disruption was omitted, the distributions of lysosomes and mitochondria overlapped resulting in poor separation of these two organelles (not shown). The activity of the soluble lysosomal matrix enzyme b-hexosaminidase peaked in fraction 3 (not shown) indicating that after MME-induced disruption lysosomal membranes and soluble contents could be separated from each other. Fractions 5 and 6, containing the bulk of lysosomal mem- branes with little mitochondrial contamination, were pooled (‘‘sucrose gradient pool’’) and further fractionated using an iodixanol density gradient. As shown for a representative separation (Fig. 3) a sigmoidal density profile with a shallow middle region was obtained (Fig. 3A). The majority of sublyso- somal membrane vesicles was floating into the top fractions

Fig. 3. Fractionation of ‘‘sucrose gradient pool’’ in a self-forming iodixanol gradient. (A) Fractions of an iodixanol gradient were assayed for activities of b-glucocerebrosidase (b-Gluc, m) and TPPI (j) and protein concentration ($). One arbitrary unit corresponds to a b-glucocerebrosidase activity of 1 mU/ml, a TPPI activity of 7.5% of the total activity in all gradient fractions and a protein concentration of 10 lg/ml. The refractive index (—) in the fractions demonstrates the sigmoidal profile of the density gradient. In (B) and (C) protein from aliquots of gradient fractions was subjected to SDS–PAGE. LAMP-2 and TPPI (C) were detected by Western blotting and protein by staining with silver nitrate (B). The open and closed arrowheads indicate the positions of LAMP-2 and TPPI.

1–4 (‘‘iodixanol upper pool’’) and separated from the bulk of protein that was recovered in fractions 14–16 (‘‘iodixanol lower pool’’) and reproducibly associated with a conspicuous Fig. 2. Separation of ‘‘dense pool’’ in a sucrose gradient after yellow band. SDS–PAGE analysis of the gradient fractions re- substrate-induced lysis. (A) The activities of b-glucocerebrosidase vealed major differences in protein composition and band pat- (b-Gluc, m), succinate dehydrogenase (Succ DH, s) and TPPI (j) tern between the two pools (Fig. 3B). The open and were determined in fractions of a representative gradient. TPPI activity closed arrowheads indicate the positions of LAMP-2 and was expressed as percentage of the total activity of all gradient fractions. (B) Aliquots of the fractions were separated in SDS–PAGE. TPPI, respectively, as examples of proteins being enriched Proteins were transferred to a PVDF membrane and probed with in fractions of ‘‘iodixanol upper’’ and ‘‘lower pool’’. The antibodies as indicated in the figure. pooled materials were examined by electron microscopy. The B. Schro¨der et al. / FEBS Letters 581 (2007) 102–108 105

3.2. Characterisation of the lysosomal inclusions The sedimentable materials from ‘‘iodixanol lower pool’’ were separated in a diagonal 2D-PAGE using strong ionic detergents in both dimensions (Fig. 5). Altogether 34 different proteins were identified as listed in Table 1. Among these, seven are marked in italics referring to their prominent pres- ence in the ‘‘iodixanol upper pool’’ (not shown). Therefore they are considered to originate from lysosomal membranes rather than from the inclusion bodies. LAMP-1 and -2, although equally abundant in the lysosomal membrane as LIMP-2, were not identified. Due to size and modifications the number of tryptic peptides detectable by MALDI-TOF- MS is much lower for these proteins than for LIMP-2 making unequivocal identification of small amounts of LAMP-1 and -2 difficult with the proteomic technology used. The remaining 27 proteins on the list are likely constituents of the inclusions. Ten of them originate from mitochondria, four from the ER, three from cytosol, two are plasma proteins and four are primarily associated with granules in phagocytic leukocytes. Despite some variations in spot pattern between preparations from different placentas the major constituents of this material were three lysosomal matrix proteins: TPPI, b-glucuronidase and acid ceramidase. Another reproducibly abundant protein was apolipoprotein D. Although stringent conditions for dena- turing and solubilising the proteins had been applied, the 2D pattern revealed a number of diffusely stained areas lacking clear delineation in addition to the major spots. The main spot of TPPI (spot-No. 83) migrated as expected for the mature enzyme which has a molecular weight of 46 kDa [16].

Fig. 4. Electron microscopic examination of different pools. The ‘‘dense’’ (A), ‘‘iodixanol upper’’ (B) and ‘‘iodixanol lower pools’’ (C) were fixed and processed for electron microscopy as described. Magnification: 40000·; Arrows: a membrane sheet adhering to a granulated body; Bar = 200 nm.

‘‘iodixanol upper pool’’ (Fig. 4B) represented a rather homogeneous preparation of sublysosomal membrane vesicles. The morphological examination of ‘‘iodixanol lower pool’’ (Fig. 4C) revealed the presence of electron-dense bodies. These were heterogeneous in shape and their granular ultrastructure resembled that of the lysosomal contents seen in ‘‘dense pool’’ lysosomes (Fig. 4A). The majority of these bodies was free of a delimiting membrane. Residual membrane sheets adhering to these bodies were seen sporadically (Fig. 4C, arrows). The pres- ence of membrane vesicles in the lower pool in addition to the aggregates was consistent with the detection of b-glucocerebro- Fig. 5. Preparative diagonal 2D-PAGE of lysosomal inclusion bodies. sidase (Fig. 3A) and LAMP-2 (Fig. 3C) in these fractions. The The sediment of ‘‘iodixanol lower pool’’ after ultracentrifugation and particulate material of presumably intralysosomal origin obvi- washing (180 lg protein) was subjected to a diagonal 2D-PAGE. ously co-localised with the lysosomal membranes in the sucrose Protein was visualised by staining with colloidal Coomassie brilliant gradient pool. Then, under the isoosmotic conditions of the blue. Numbers refer to proteins listed in Table 1. In the second dimension the sample proteins migrated with a lag as compared to the iodixanol gradient, the properties of both species were suffi- molecular weight marker. The spots Nos. 161 and 164 with an ciently different to allow separation (mean densities of ‘‘iodix- abnormal rapid migration in the second dimension were not observed anol upper and lower pools’’: 1.06 g/cm3 and 1.12 g/cm3). in similar separations and are considered an artefact. 106 B. Schro¨der et al. / FEBS Letters 581 (2007) 102–108

Table 1 Proteins identified in lysosomal inclusion bodies Protein Acc.-No.a MWb (kDa) Spot No.c Acid ceramidase Q13510 44.7 43, 104 Actin, beta P60709 41.7 71 ADP/ATP 2 (SLC25A5) P05141 32.8 126 ADP/ATP translocase 3 (SLC25A6) P12236 32.7 126 ADP-ribosylation factor-like 10 C Q9NVJ2 21.5 154 Alkaline phosphatase P05187 58.0 32 Alpha-2-macroglobulin P01023 163.3 16, 19, 85 Apolipoprotein D P05090 21.3 145 ATP synthase alpha chain, mitochondrial P25705 59.7 135 ATP synthase beta chain, mitochondrial P06576 56.6 66 Azurocidin (CAP 37) P20160 26.9 121, 128, 129 Beta-galactosidase P16278 76.1 41 Beta-glucocerebrosidase P04062 59.7 34 Beta-glucuronidase P08236 74.7 2, 6, 9, 14, 164 Calpain-6 (calpamodulin) Q9Y6Q1 74.6 28 G P08311 28.8 137, 161 Creatine kinase, mitochondrial P12532 47.0 77 Cytochrome P450 11A1, mitochondrial P05108 60.2 69 Cytochrome P450 19A1 P11511 57.8 135 Dihydrolipoamide succinyltransferase component of 2-oxoglutarate dehydrogenase P36957 48.6 124 Dolichyl-diphosphooligosasaccharide-protein-glycosylransferase 48 kDa subunit P39656 48.8 67 -regulated protein, 78 kDa P11021 72.3 23 Glyceraldehyde-3-phosphate dehydrogenase P04406 35.9 90 Heat shock protein 60 kDa (HSP60) P10809 61.1 45 LIMP-2 Q14108 54.2 25 Monoamine oxidase type A P21397 59.7 30, 34, 40, 100 (leukocyte ) P24158 27.8 121, 124, 129, 131 Myeloperoxidase P05164 83.9 49 Steryl-sulfatase P08842 65.5 37 Tripeptidyl peptidase I O14773 61.2 83, 141, 149, 152, 158 V0-ATPase d P61421 40.3 63 V1-ATPase A P38606 68.3 25 V1-ATPase B P21281 56.5 56, 58 VDAC 1 (porin 31 HL/HM) P21796 30.6 92 Proteins printed in italics are of higher abundance in ‘‘iodixanol upper pool’’ and likely constituents of the lysosomal membranes. aDatabase accession numbers refer to Swiss Prot database http://www.expasy.org. bMolecular mass values refer to those in the database. cSpot numbers refer to the positions shown in Fig. 4.

Furthermore, TPPI was detected in spots in the 30 kDa area placenta, an organ continuously growing in the course of a possibly indicating proteinolytic breakdown. pregnancy, contains significant amounts of lipofuscin-like lysosomal inclusions as shown in this study what is usually 3.3. Tripeptidyl peptidase I in the course of lysosomal considered as a hallmark of ageing. fractionation Two studies have analysed the proteome of lipofuscin iso- In both gradients the profile of the enzyme activity lated from retina pigment epithelium. Schutt et al. were able (Figs. 2A, 3A) correlated well with the signals from immune to identify 60 proteins by mass spectrometry after separation detection (Figs. 2B, 3C). In the sucrose gradient TPPI was in conventional 2D-PAGE [19]. Warburton et al. detected 41 distributed similar to LAMP-2. However, in the iodixanol proteins taking a combined approach with 1D-SDS and 2D- gradient the bulk of TPPI was separated from the membrane PAGE [20]. Only six of the 27 proteins that were detected in marker and recovered in the ‘‘iodixanol lower pool’’ contain- intralysosomal aggregates in the present study (actin, a- and ing the lysosomal inclusions. The maximum of TPPI (fraction b-chains of mitochondrial ATP synthase, glucose-regulated 15) was different from that of lysosomal membrane markers protein, glyceraldehyde-3-phosphate dehydrogenase and porin (fraction 14) also being present in this part of the gradient. VDAC 1) have also been reported by Schutt et al. [19,20]. Merely three of the proteins (actin, b-chain of mitochondrial ATP synthase, TPPI) overlap with the list by Warburton 4. Discussion et al. [20]. A common finding is that proteins originating from varied subcellular locations are detected. The differences be- Upon disruption of placental lysosomes, particulate lipofus- tween the results may reflect the dissimilarities regarding the cin-like materials were released that could be separated from tissue examined. membranes. These materials were not autofluorescent but The diagonal 2D-PAGE separation shown in Fig. 5 demon- exhibited a yellow-brown coloration. There have been strated proteins with apparent molecular weights between contradictory reports about the existence of lipofuscin in 20 kDa and 110 kDa. High molecular weight proteins were human placenta based solely on the microscopic detection obviously underrepresented in the inclusion bodies. Proteins of autofluorescent pigments [17,18]. Rather surprisingly, below 20 kDa were not reliably separated. Subunit c of the B. Schro¨der et al. / FEBS Letters 581 (2007) 102–108 107 mitochondrial ATP synthase as well as saposins which are [3] Terman, A., Gustafsson, B. and Brunk, U.T. (2006) The major constituents of the storage material in several types of lysosomal-mitochondrial axis theory of postmitotic aging and neuronal ceroid lipofuscinoses [21] would not have been cell death. Chem. Biol. Interact. [4] Burcham, P.C. and Kuhan, Y.T. (1997) Diminished susceptibility resolved if they had been present in the lysosomal inclusion to after protein modification by the lipid peroxidation bodies. The apparently smeary separation of the stained product malondialdehyde: inhibitory role for crosslinked and protein (Fig. 5) might indicate that the lysosomal inclusions noncrosslinked adducted proteins. Arch. Biochem. Biophys. 340, are heterogeneous mixtures of modified proteins unlikely to 331–337. [5] Diettrich, O., Gallert, F. and Hasilik, A. (1996) Purification of be separated into discrete spots and identified in completeness. lysosomal membrane proteins from human placenta. Eur. J. Cell Those identified may represent the most prominent com- Biol. 69, 99–106. ponents of an undegradable residue and lysosomal matrix [6] Goldman, R. and Kaplan, A. (1973) Rupture of rat liver proteins that tend to associate with this residue. lysosomes mediated by L- esters. Biochim. Biophys. It seems likely that autophagocytosed organelles represent Acta 318, 205–216. [7] Gatt, S. (1969) b-Glucosidase from bovine brain. Methods the origin of the reported intralysosomal aggregates. Auto- Enzymol. 14, 152–155. phagocytosis has been observed in human syncytiotrophoblast [8] Bradford, M.M. (1976) A rapid and sensitive method for the [22] and lysosomes have been postulated to play a role in mem- quantitation of microgram quantities of protein utilizing the brane turnover during differentiation of cytotrophoblast to principle of protein-dye binding. Anal. Biochem. 72, 248–254. [9] Laemmli, U.K. (1970) Cleavage of structural proteins during the syncytiotrophoblast [23]. Furthermore trophoblastic cells have assembly of the head of bacteriophage T4. Nature 227, 680–685. been shown to perform phagocytosis [24,25] which may be [10] Heukeshoven, J. and Dernick, R. (1988) Improved silver staining important for their invasion into the maternal endometrium procedure for fast staining in PhastSystem Development Unit. I. and the extensive tissue remodelling that is characteristic of Staining of sodium dodecyl sulfate gels. Electrophoresis 9, 28–32. early placenta development [25,26]. Thus, the observed intra- [11] Radons, J., Faber, V., Buhrmester, H., Volker, W., Horejsi, V. and Hasilik, A. (1992) Stimulation of the biosynthesis of lysosomal aggregates may also represent residues from cells lactosamine repeats in glycoproteins in differentiating U937 cells and materials internalised by phagocytosis during this phase. and its suppression in the presence of NH4Cl. Eur. J. Cell Biol. 57, TPPI has been reported to behave like a soluble lysosomal 184–192. enzyme in brain extracts [27]. The enzyme has been purified [12] Ito, S. and Karnovsky, M.J. (1968) Formaldehyde–glutaralde- hyde fixative containing trinitro compounds. J. Cell Biol. 39, to homogeneity from different species and tissues and shown 168a–169a. to aggregate and oligomerise under acidic conditions [28–30]. [13] Schro¨der, B. and Hasilik, A. (2006) A protocol for combined Defects in TPPI cause the late infantile type of neuronal ceroid delipidation and subfractionation of membrane proteins using lipofuscinosis (NCL) [31]. The enzyme has been demonstrated organic solvents. Anal. Biochem. 357, 144–146. to interact with the CLN5 protein, reportedly an integral mem- [14] Neuhoff, V., Arold, N., Taube, D. and Ehrhardt, W. (1988) Improved staining of proteins in polyacrylamide gels including brane protein [32]. This may implicate an association of TPPI isoelectric focusing gels with clear background at nanogram with the lysosomal membrane which, however, was not de- sensitivity using Coomassie Brilliant Blue G-250 and R-250. tected in this study. It has been reported that the TPPI content Electrophoresis 9, 255–262. of human brain tissue increases with age [33]. This finding may [15] Jahn, O., Hesse, D., Reinelt, M. and Kratzin, H.D. (2006) Technical innovations for the automated identification of gel- be explained by an accumulation of lipofuscin-like storage separated proteins by MALDI-TOF mass spectrometry. Anal. materials binding TPPI. Bioanal. Chem. 386, 92–103. According to current protocols lysosomal membranes are [16] Lin, L., Sohar, I., Lackland, H. and Lobel, P. (2001) The human isolated by sedimenting disrupted lysosomes prepared by CLN2 protein/tripeptidyl-peptidase I is a that freeze-thawing or sonication. By this means soluble lysosomal autoactivates at acidic pH. J. Biol. Chem. 276, 2249–2255. [17] Parmley, T.H., Gupta, P.K. and Walker, M.A. (1981) Aging matrix enzymes are removed but particulate lysosomal inclu- pigments in term human placenta. Am. J. Obstet. Gynecol. 139, sions are enriched along with the membranes. Fractionation 760–766. of membranes and intralysosomal aggregates can be achieved [18] Haigh, M., Chawner, L.E. and Fox, H. (1984) The human by combining a specific substrate-induced lysis of lysosomes placenta does not contain lipofuscin pigment. Placenta 5, 459– 464. followed by a separation using density gradients. This [19] Schutt, F., Ueberle, B., Schnolzer, M., Holz, F.G. and Kopitz, J. methodology should facilitate studies on lysosomal inclusions (2002) Proteome analysis of lipofuscin in human retinal pigment from other sources and tissues. epithelial cells. FEBS Lett. 528, 217–221. [20] Warburton, S., Southwick, K., Hardman, R.M., Secrest, A.M., Acknowledgements: The authors thank the Bundesministerium fu¨r Grow, R.K., Xin, H., Woolley, A.T., Burton, G.F. and Thulin, Bildung und Forschung for support (031U104B) and Eva Becker, C.D. (2005) Examining the proteins of functional retinal lipofus- Karlheinz Burk, Thomas Stein, Heinrich Kaiser, Traudel Jarosch, Bri- cin using proteomic analysis as a guide for understanding its gitte Agricola and Nicole Eiselt for technical assistance. We also thank origin. Mol. Vis. 11, 1122–1134. Olaf Jahn and Hartmut Kratzin, for access to the mass spectrometric [21] Dawson, G. and Cho, S. (2000) Batten’s disease: clues to neuronal facilities of the Max-Planck-Institute of Experimental Medicine, protein catabolism in lysosomes. J. Neurosci. Res. 60, 133–140. Go¨ttingen. [22] Nelson, D.M. (1996) Apoptotic changes occur in syncytiotropho- blast of human placental villi where fibrin type fibrinoid is deposited at discontinuities in the villous trophoblast. Placenta 17, 387–391. References [23] Contractor, S.F., Banks, R.W., Jones, C.J. and Fox, H. (1977) A possible role for placental lysosomes in the formation of villous [1] de Duve, C. (1983) Lysosomes revisited. Eur. J. Biochem. 137, syncytiotrophoblast. Cell Tissue Res. 178, 411–419. 391–397. [24] Amarante-Paffaro, A., Queiroz, G.S., Correa, S.T., Spira, B. and [2] Seehafer, S.S. and Pearce, D.A. (2006) You say lipofuscin, we say Bevilacqua, E. (2004) Phagocytosis as a potential mechanism for ceroid: defining autofluorescent storage material. Neurobiol. microbial defense of mouse placental trophoblast cells. Repro- Aging 27, 576–588. duction 128, 207–218. 108 B. Schro¨der et al. / FEBS Letters 581 (2007) 102–108

[25] Hata, T., Ohkawa, K., Tomita, M. and Kishino, M. (1981) (2001) Rat tripeptidyl peptidase I: molecular cloning, functional Phagocytosis of human cytotrophoblast cell invading into decid- expression, tissue localization and enzymatic characterization. ual tissue in early stage of gestation. Acta Obstet. Gynaecol. Jpn. Biol. Chem. 382, 1715–1725. 33, 537–544. [30] Page, A.E., Fuller, K., Chambers, T.J. and Warburton, M.J. [26] Choy, M.Y. and Manyonda, I.T. (1998) The phagocytic activity (1993) Purification and characterization of a tripeptidyl peptidase of human first trimester extravillous trophoblast. Hum. Reprod. I from human osteoclastomas: evidence for its role in bone 13, 2941–2949. resorption. Arch. Biochem. Biophys. 306, 354–359. [27] Sohar, I., Sleat, D.E., Jadot, M. and Lobel, P. (1999) Biochemical [31] Golabek, A.A. and Kida, E. (2006) Tripeptidyl-peptidase I in characterization of a lysosomal protease deficient in classical late health and disease. Biol. Chem. 387, 1091–1099. infantile neuronal ceroid lipofuscinosis (LINCL) and develop- [32] Vesa, J., Chin, M.H., Oelgeschlager, K., Isosomppi, J., DellAn- ment of an enzyme-based assay for diagnosis and exclusion of gelica, E.C., Jalanko, A. and Peltonen, L. (2002) Neuronal ceroid LINCL in human specimens and animal models. J. Neurochem. lipofuscinoses are connected at molecular level: interaction of 73, 700–711. CLN5 protein with CLN2 and CLN3. Mol. Biol. Cell 13, 2410– [28] Junaid, M.A., Wu, G. and Pullarkat, R.K. (2000) Purification and 2420. characterization of bovine brain lysosomal pepstatin-insensitive [33] Kida, E., Golabek, A.A., Walus, M., Wujek, P., Kaczmarski, W. proteinase, the product deficient in the human late-infantile and Wisniewski, K.E. (2001) Distribution of tripeptidyl peptidase neuronal ceroid lipofuscinosis. J. Neurochem. 74, 287–294. I in human tissues under normal and pathological conditions. J. [29] Du, P.G., Kato, S., Li, Y.H., Maeda, T., Yamane, T., Yamam- Neuropathol. Exp. Neurol. 60, 280–292. oto, S., Fujiwara, M., Yamamoto, Y., Nishi, K. and Ohkubo, I.