The oleic acid complexes of proteolytic fragments of a-lactalbumin display apoptotic activity Serena Tolin 1 , Giorgia De Franceschi 1 , Barbara Spolaore 1 , Erica Frare 1 , Marcella Canton 2 , Patrizia Polverino de Laureto 1 and Angelo Fontana 1 1 CRIBI Biotechnology Centre, University of Padua, Italy 2 Department of Experimental Biomedical Sciences, University of Padua, Italy Introduction a-Lactalbumin (a-LA) is a small, acidic, Ca 2+ -bind- ing protein involved in the biosynthesis of lactose, being a component of the lactose synthase complex [1]. For a few decades, a-LA has been the subject of intensive structural investigations, and it has served as a model system in many protein folding studies [2,3]. The 123 residue chain of a-LA is organized into a discontinuous a-helical domain composed of residues 1–39 and 81–123, and a small b-domain comprising the rest of the polypeptide chain [4,5] (Fig. 1). A noteworthy property of a-LA is its ability to adopt a partly folded or molten globule (MG) state under a variety of conditions, including low pH. This state, lacking the specific interactions of the ter- tiary structure of the native protein, but maintaining a high degree of secondary structure, has been exten- sively analyzed with a variety of techniques and approaches [6–9]. Keywords apoptosis; HAMLET; oleic acid; protein fragments; a-lactalbumin Correspondence P. Polverino de Laureto, CRIBI Biotechnology Centre, University of Padua, Viale G. Colombo 3, 35121 Padua, Italy Fax: +39 049 827 6159 Tel: +39 049 827 6157 E-mail: patrizia.polverinodelaureto@unipd.it (Received 7 August 2009, revised 9 October 2009, accepted 27 October 2009) doi:10.1111/j.1742-4658.2009.07466.x The complexes formed by partially folded human and bovine a-lactalbumin with oleic acid (OA) have been reported to display selective apoptotic activity against tumor cells. These complexes were named human (HAMLET) or bovine (BAMLET) alpha-lactalbumin made lethal to tumor cells. Here, we analyzed the OA complexes formed by fragments of bovine a-lactalbumin obtained by limited proteolysis of the protein. Speci- fically, the fragments investigated were 53–103 and the two-chain fragment species 1–40 ⁄ 53–123 and 1–40 ⁄ 104–123, these last being the N-terminal fragment 1–40 covalently linked via disulfide bridges to the C-terminal fragment 53–123 or 104–123. The OA complexes were obtained by mixing the fatty acid and the fragments in solution (10-fold and 15-fold molar excess of OA over protein fragment) or by chromatography of the frag- ments loaded onto an OA-conditioned anion exchange column and salt- induced elution of the OA complexes. Upon binding to OA, all fragments acquire an enhanced content of a-helical secondary structure. All OA com- plexes of the fragment species showed apoptotic activity for Jurkat tumor cells comparable to that displayed by the OA complex of the intact pro- tein. We conclude that the entire sequence of the protein is not required to form an apoptotic OA complex, and we suggest that the apoptotic activity of a protein–OA complex does not imply specific binding of the protein. Abbreviations a-LA, a-lactalbumin; BAMLET, bovine a-lactalbumin made lethal to tumor cells; CAC, critical aggregate concentration; HAMLET, human a-lactalbumin made lethal to tumor cells; MG, molten globule; OA, oleic acid; [h], mean residue ellipticity; TNS, 6-(p-toluidino)-2- naphthalenesulfonate. FEBS Journal 277 (2010) 163–173 ª 2009 The Authors Journal compilation ª 2009 FEBS 163 An interesting property of a-LA is its capacity to interact with membranes and lipid bilayers [10–14], as well as fatty acids [15]. In particular, a complex formed by Ca 2+ -depleted a-LA in its partly folded state with oleic acid (OA) has been extensively studied. This OA complex, named human a-LA made lethal to tumor cells (HAMLET), was initially isolated from human milk, and was shown to selectively induce apoptosis in tumor and immature cells, but not in healthy cells [16–19]. It was proposed that the condi- tions required to induce HAMLET formation in vivo are those of the stomach of the breastfed child, where low pH may partially unfold a-LA by releasing its protein-bound Ca 2+ [19–21], and free OA can be produced by lipases that hydrolyze milk triglycerides [22,23]. HAMLET can also be prepared in vitro,by application of the apo form of a-LA to an anion exchange column equilibrated with OA and elution of the OA complex with high salt concentration, followed by dialysis and lyophilization [18]. It seems that the formation of the OA complex relies on the fact that the apo form of a-LA is more hydrophobic than the holo form, and thus is prone to bind the hydrophobic fatty acid [20]. The conformational features of a-LA in the OA complex are those of a protein MG, and upon binding to the protein, OA can probably stabilize this altered protein conformation [21]. HAMLET-like com- plexes with similar biological activity can be obtained with a-LA from different species, including bovine, equine, porcine, ovine and caprine a-LA [24]. The OA complex of bovine a-LA was named bovine a-LA made lethal to tumor cells (BAMLET) [21]. Moreover, it was also shown that a-LA mutants with amino acid replacements at the level of the Ca 2+ -binding loop were capable of producing active OA complexes [21]. Despite the intensive research on HAMLET and BAMLET, the molecular features of the active OA complex in terms of protein ⁄ fatty acid stoichiometry and monomeric ⁄ oligomeric state of the protein in the complex are still not clarified, and are a matter of debate in the current literature [20,25–27]. The molecu- lar mechanism of interaction and physicochemical properties of the OA complex are not understood, and neither are the mechanisms and cellular events involved in the toxicity of HAMLET. It was shown by using labeled a-LA that the OA complex is able to 14053 104 1 40 53 103 1 123 H1 5–11 h1b 23–34 S1 S2 S3 H2 h2 86–98 105–110 h3c H3 H4 123 123 (1−40/53−123)/OA NaCl molarity ( ) 0.0 0.5 1.0 BAMLET (53−103)/OA 0.0 0.5 1.0 (1−40/104−123)/OA Volume (mL) 0 20 40 60 80 0 20 40 60 80 020406080020406080 Relative absorbance ( ___ ) α-LA EDTA EDTA EDTA EDTA Aggregates a A B c b d Fig. 1. (A) Top: scheme of the secondary structure of the 123 residue chain of a-LA [4]. The four a-helices (H1–H4) are indicated by large boxes, and the corresponding chain segments are given above them. The three b-strands (S1, 41–44; S2, 47–50; S3, 55–56) and the 3 10 helices (h1b, 18–20; h2, 77–80; h3c, 115–118) are indicated by small boxes. Bottom: schematic representation of the three a-LA fragments investigated. The four disulfide bridges (6–120, 28–111, 61–77, and 73–91) are represented by thin lines, and the gray box indicates the segment encompassing the Ca 2+ -binding loop. (B) Preparation of the OA complexes of a-LA (a) and its fragments (b–d) by chromatography on an OA-conditioned anion exchange col- umn [18]. The protein material was applied to a DEAE-Trisacryl M column conditioned with OA, and the OA complexes were eluted with a gradient of 1 M NaCl (dashed lines). The solid bars indicate the fractions of the effluent from the column that were collected for further studies. Oleic acid complexes of a-lactalbumin fragments S. Tolin et al. 164 FEBS Journal 277 (2010) 163–173 ª 2009 The Authors Journal compilation ª 2009 FEBS bind to tumor cells and accumulate in the cell nuclei [28]. These authors identified specific histone proteins as nuclear targets for HAMLET. However, it was also shown that a-LA in the absence of OA can interact with histones and charged, disordered poly-a-amino acids (i.e. poly-Lys and poly-Arg). This a-LA interac- tion was shown to be driven by electrostatic forces [29,30]. Therefore, the active species in the cell may be the whole protein–fatty acid complex, the protein alone, or even the OA by itself. In this last case, the fatty acid aggregation state should be considered, as it can be significantly influenced by the presence of a protein in the same solution. In this study, we analyzed the propensity of three fragment species of bovine a-LA, obtained by limited proteolysis of the protein [9,31,32], to bind OA and to form biologically active OA complexes. As shown in Fig. 1, the fragments have different structural charac- teristics, with fragment 1–40 ⁄ 104–123 encompassing three of the four a-helices of the native protein, frag- ment 53–103 containing the chain segment that binds Ca 2+ in the native protein, and fragment 1–40 ⁄ 53–123 being able to adopt, at neutral pH, an MG conforma- tion resembling that of the MG conformation adopted by intact a-LA at pH 2.0 [31,32]. The conformational properties of the OA complexes of these fragments were analyzed by far-UV CD measurements, and it was shown that the fragments acquire an enhanced content of a-helical secondary structure upon binding OA. The physical and aggregation state of OA at physiological pH was analyzed by fluorescence and turbidimetric analyses. It was shown that the fragments, as well as the entire protein, depress the critical concentration for aggregate formation [critical aggregate concentration (CAC)] of OA and induce the formation of small and water-soluble OA aggregates. All OA complexes dis- played apoptotic activity for tumor cells, and the extent of their activity was comparable to that observed with the OA complex of the intact protein, i.e. BAMLET. Our results indicate that the entire 123 residue chain of a-LA is not required for forming a cytotoxic OA com- plex, and raise the possibility that the cell-damaging effects of the various OA complexes could result from an enhanced solubility of the otherwise poorly soluble and inherently toxic fatty acid [33]. Results Preparation of complexes of a-LA fragments with OA Two procedures were followed to prepare the OA–fragment complexes, namely by simple mixing the two components in solution, or by chromatography using an OA-conditioned anion exchange column, as described by Svensson et al. [18] for the preparation of HAMLET. The two procedures were used here, as it is not clear whether a mixing procedure results in a less active or inactive complex [18,20,34]. Instead, we [35] and others [26,36,37] have shown that it is indeed pos- sible to prepare an OA complex displaying similar structural properties to HAMLET or BAMLET, i.e. to an OA complex prepared by chromatography. Nev- ertheless, here we preferred to use and compare both procedures in preparing the OA complexes. Bovine a-LA and its fragments were loaded on an anion exchange column conditioned with OA. The chromatographic profile obtained with intact a-LA was similar to that previously reported [18]. Salt and EDTA were eluted first from the column. The free protein was eluted from the column at low salt concen- tration, whereas the OA complex was eluted at  1 m salt. The three fragments strongly bound to the OA-conditioned matrix, and their OA complexes could be eluted by high salt (Fig. 1B). The amounts of pro- tein fragment in the eluted OA complex, calculated on the basis of the material loaded onto the column, were  50% for fragments 1–40 ⁄ 53–123 and 1–40 ⁄ 104–123, and  25% for fragment 53–103, as estimated from UV absorption measurements. This indicated that a proportion of the protein fragment material remained bound to the column. In the case of fragment 53–103, aggregated species were eluted at a higher retention time than that of the OA–fragment 53–103 complex. Aggregation of the fragment was deduced from the turbidity of the last eluted fraction (Fig. 1Bd). This would account for the low recovery of soluble OA complex of fragment 53–103. For the sake of compari- son, the OA complexes were also prepared in solution by direct mixing of the a-LA fragments with OA at a molar ratio of 1 : 10 or 1 : 15 (see below). Conformational properties of protein fragment complexes with OA The conformational properties of the OA complexes formed by a-LA fragments prepared by chromatogra- phy or by direct mixing in solution were analyzed by far-UV CD spectroscopy in NaCl ⁄ P i (pH 7.4). Figure 2A shows the far-UV CD spectra of fragment 1–40 ⁄ 104–123 in the presence of increasing concentra- tions of OA. The CD spectrum of this fragment appeared to be that of a largely disordered polypep- tide, but upon addition of OA the spectrum acquired the characteristics of a-helical secondary structure. It is of interest that, in the presence of OA (protein ⁄ fatty S. Tolin et al. Oleic acid complexes of a-lactalbumin fragments FEBS Journal 277 (2010) 163–173 ª 2009 The Authors Journal compilation ª 2009 FEBS 165 acid molar ratio of 1 : 10), the CD spectrum of frag- ment 1–40 ⁄ 104–123 was quite similar to that of the corresponding OA complex prepared by chromato- graphy, implying that the OA complexes prepared by the two procedures displayed similar conformational features. Analogous conformational effects of OA binding were observed with fragment 53–103 in the presence of 15 equivalents of OA (Fig. 2B). Thus, frag- ment species 1–40 ⁄ 104–123 and 53–103 both appeared to be rather disordered in solution at pH 7.4, but upon binding OA they acquired a folded structure character- ized by a significant content of a-helical structure, as the OA complexes displayed far-UV CD spectra with the typical minima of ellipticity at 208 and 222 nm of the a-helical secondary structure [38]. The fragment species 1–40 ⁄ 53–123 comprises almost all of the 123 residue chain of a-LA (Fig. 1), and adopted a significantly folded structure in solution, as shown by the characteristics of its far-UV CD spec- trum (Fig. 2C). In this case, the addition of OA induced a conformational change, but not as dramatic as seen with the other two fragments. Here, we used fragment solutions devoid of Ca 2+ , because we have previously shown that the conformational features of fragments 53–103 and 1–40 ⁄ 53–123, containing the Ca 2+ -binding loop of the intact protein (Fig. 1), are affected by Ca 2+ [31,32]. Determination of the aggregation state of OA The phase behavior of OA is strongly dependent on pH and fatty acid concentration [39]. In NaCl ⁄ P i (pH 7.4), OA forms oil droplets and vesicles of variable size [40–42]. To understand the effect of protein fragments on OA aggregation state, we measured the OA CAC. We use this term because a complete morphological characterization of the aggregate state of OA is not yet available. The CAC of OA at pH 7.4 was estimated by using the fluorescent dye 6-(p-toluidino)-2-napthalene- sulfonate (TNS) [43]. In NaCl ⁄ P i (pH 7.4), the CAC of OA was calculated as 19.8 ± 0.3 lm (Fig. 3A, insert). This value is similar to that previously determined for OA [40]. The same measurements using TNS were con- ducted in the presence of a-LA (Fig. 3A) or its frag- ments (Fig. 3B). All protein species were able to reduce by  20-fold the CAC value of OA. Estimated values of the CAC of OA were 0.94 ± 0.24 lm in the presence of a-LA, and 0.86 ± 0.29, 1.22 ± 0.24 and 0.93 ± 0.56 lm, respectively, in the presence of frag- ment species 1–40 ⁄ 53–123, 1–40 ⁄ 104–123 and 53–103. We also conducted turbidity analysis of OA solu- tions and mixtures, as this method is often used for measuring the critical vesicular concentration of lipids [44]. Figure 3C shows the variation of absorbance (A) at 400 nm of samples containing increasing amounts of OA in the absence or presence of a-LA. The strik- ing observation deriving from these measurements is that the added protein was able to completely inhibit the formation of large aggregates that caused light scattering at 400 nm (Fig. 3C, open circles). Fragment species 1–40 ⁄ 53–123 and 1–40 ⁄ 104–123 were also able to similarly depress the OA aggregation. Fragment 53–103 also caused a reduction in the aggregation of OA, but to a minor extent (Fig. 3D). –15 –10 –5 0 5 [θ] x 10 –3 (deg·cm 2 ·dmol –1 ) –15 –10 –5 0 5 1 : 1 1 : 3 1 : 5 1 : 7 1 : 10 1 : 15 by column 1−40/104−123 190 210 230 250 190 210 230 250 190 210 230 250 –15 –10 –5 0 5 (53−103)/OA (by mix 1 : 15) (53−103)/OA (by column) 53−103 B A C (1−40/53−123)/OA (by column) 1−40/53−123 Wavelength (nm) (1−40/53−123)/OA (by mix 1 : 15) Fig. 2. Far-UV CD spectra of a-LA fragments in NaCl ⁄ P i (pH 7.4). (A) Far-UV CD spectra of fragment 1–40 ⁄ 104–123 in the absence (dotted line) or presence (continuous line) of increasing amounts of OA. Numbers near the CD spectra refer to fragment ⁄ OA molar ratios of 1 : 1, 1 : 3, 1 : 5, 1 : 7, 1 : 10, and 1 : 15. The CD spectrum of the OA complex of the fragment obtained by chromatography is also shown (dashed line). (B) CD spectra of the OA complex of fragment 53–103 obtained by chromatography (dashed line) or by mixing in solution (continuous line). The spectrum of the OA free fragment is reported as a reference (dotted line). (C) CD spectra of fragment 1–40 ⁄ 53–123 (dotted line) and its OA complex obtained by chromatography (dashed line) and by mixing the fragment and OA in solution at a fragment ⁄ OA molar ratio of 1 : 15 (continuous line). Oleic acid complexes of a-lactalbumin fragments S. Tolin et al. 166 FEBS Journal 277 (2010) 163–173 ª 2009 The Authors Journal compilation ª 2009 FEBS Cellular toxicity The ability of the OA complexes of the a-LA frag- ments to induce apoptosis-like cell death was exam- ined. Assays were conducted on Jurkat cells, using the OA complexes prepared by direct mixing or by chro- matography using an OA-conditioned column (see Experimental procedures). Cells treated with the fragment–OA complexes suffered considerable loss of viability through apoptosis-like death, whereas the OA-free fragments displayed negligible toxicity (Fig. 4). The OA complexes of the various fragment species were also tested at a fragment ⁄ OA molar ratio of 1 : 3, a condition that caused only a slight confor- mational change in the fragment’s secondary structure, as deduced from far-UV CD spectra. In this case, the OA complexes did not display cellular toxicity (not shown). The extent of apoptotic activity of the OA complexes of the fragments was comparable to that observed with the OA complex of the intact protein prepared by chromatography, i.e. BAMLET, or by mixing the intact protein with 15 equivalents of OA in solution. In the absence of added protein species, OA alone and at the concentration used for formation of OA complexes displayed negligible toxicity, similarly to NaCl ⁄ P i or the control sample (culture medium). Conversely, as previously shown, OA can be toxic to Jurkat cells via an apoptosis mechanism at higher con- centrations and with a much longer duration of incu- bation [33]. Fragment 1–40 ⁄ 104–123 is a two-chain species cross- linked by the two disulfide bridges 6–120 and 28–123 of intact a-LA. Reduction of this fragment with tris(2-carboxyethyl)phosphine, followed by S-alkylation with iodoacetamide and RP-HPLC chromatography, allowed us to prepare the single-chain, S-carboxami- domethylated fragments 1–40 and 104–123. The inter- action of OA with these fragments was monitored by far-UV CD measurements. As shown in Fig. S1, OA induced a-helical secondary structure in both frag- ments, which were otherwise largely unfolded in the absence of the fatty acid. It is of note that fragment 1–40 encompasses helix H1 (5–11) and helix H2 (23–34), and fragment 104–123 encompasses helix H4 (105–110), in native a-LA [4]. The OA complexes of the two fragments, as obtained by mixing them with 15 equivalents of the fatty acid, displayed significant apoptotic activity on Jurkat cells (Fig. S1, bottom). It is of note that the OA–fragment 104–123 complex was even more active than BAMLET, i.e. the OA–a-LA complex prepared by column chromatography (see Experimental procedures). Therefore, OA complexes of Relative fluorescence 1.0 1.2 1.4 1.6 1.8 [OA] μM 0 20 40 60 80 100 Rel. fluorescence 0.8 1.0 1.2 1.4 1.6 1.8 [OA] (μM) AB D 02468100 2 46 810 0 100 200 300 400 500 0 100 200 300 400 500 A 400 nm 0.0 0.1 0.2 0.3 C Fig. 3. Characterization of the physical state of OA solutions by TNS fluorescence emis- sion (A, B) and turbidity (C, D). All measure- ments were conducted in NaCl ⁄ P i (pH 7.4), in the absence (d) or presence of a-LA (s), fragment 1–40 ⁄ 53–123 ( ), fragment 1– 40 ⁄ 104–123 (n), or fragment 53–103 ()). (A, B) Aliquots of OA (from 0 to 10 l M) were added to a solution of TNS (20 l M), and the intensity of fluorescence emission at 460 nm was recorded, after excitation at 360 nm. The CAC is defined as the lipid concentration at which the two linear por- tions of the lines of fluorescence intensities intersect [61]. The TNS fluorescence of OA (up to 100 l M) solutions was also measured in the absence of intact a-LA (A, insert). (C, D) Turbidimetric analysis of OA solutions in the absence (filled circles) or presence of 10 l M protein (open circles) or fragment species (symbols as above). Measurements of absorbance were conducted at 400 nm on samples containing OA up to 500 l M. S. Tolin et al. Oleic acid complexes of a-lactalbumin fragments FEBS Journal 277 (2010) 163–173 ª 2009 The Authors Journal compilation ª 2009 FEBS 167 peptide fragments even shorter than the three frag- ments shown in Fig. 1 can display cellular toxicity. Discussion The data presented here indicate a strong mutual inter- action between OA and a-LA fragments. Indeed, the addition of OA induces enhancement of secondary structure of the fragment species, and these signifi- cantly modify the physical state of the fatty acid in solution. The increase in the a-helical structure in the fragment species upon addition of OA (Fig. 2) derives from the fact that fatty acids, lipids and detergents can provide a hydrophobic environment that is able to induce and stabilize secondary structure of polypep- tides [45–47]. The interaction of protein species with OA is also simply shown by the fact that a turbid OA suspension in water at neutral pH becomes clear after the addition of protein ⁄ fragment species. The OA complexes of the a-LA fragments have been prepared by direct mixing in solution and by the chro- matographic method of Svensson et al. [18], utilizing an OA-conditioned anion exchange column, followed by extensive dialysis of the OA complex eluted from the column at high salt concentration, and then lyoph- ilization. As the conformational and biological proper- ties of the OA complexes prepared here by the two methods are very similar, we consider the mixing procedure in solution to be suitable, being easier, reproducible and less cumbersome than the chromato- graphic one. Furthermore, we have previously shown that the mixing procedure can be effectively utilized for the preparation of an a-LA–OA complex with comparable structural features to BAMLET [35], and other reports have more recently documented its successful use [26,36,37]. The phase behavior of OA is critically dependent on the ionization degree of the fatty acid, and is thus affected by the pH and ionic strength of the aqueous solution [40–42]. In NaCl ⁄ P i (pH 7.4), OA forms aggregates of different sizes (diameter 25–250 nm), as deduced by transmission electron microscopy (not shown). The a-LA fragment species, as well as the intact protein, strongly affect these structures. Both turbidimetric and transmission electron microscopy analyses show the disappearance of the large, aggre- gated structures, and by means of fluorescence emis- sion measurements, a  20-fold decrease in CAC was found, indicating the formation of smaller aggregates at a lower OA concentration in the presence of protein species. As the formation of OA micelles requires the complete ionization of the fatty acid molecules, and micelles are formed at pH > 9 [39], it is likely that, under the experimental conditions described here, small vesicles or oil droplets are induced in OA in the presence of the protein or its fragments. A depression of the CAC of anionic detergents simi- lar to that observed here with OA aggregates was reported to occur also with other proteins, and this effect was explained by considering both electrostatic 1−4 0 /53−123 (1 − 4 0 /104−123/OA) by column (1− 4 0 /53 −1 2 3)/OA by mix (1 − 4 0 /53−123)/OA by column (1−4 0 /104−123)/OA b y mix 1−4 0 /104− 1 2 3 (53−1 0 3)/OA b y mix α-L A /OA by mix 5 3 − 1 0 3 (53 −1 0 3)/OA b y column BAMLET α-L A OA NaCl/P i CT 0 20 40 60 80 100 Late apoptosis Early apoptosis Cell death (%) Fig. 4. Cytotoxicity of OA complexes. Jur- kat cells (10 6 cells per mL) were incubated at 37 °C with the OA complexes of a-LA or its fragments prepared by column chroma- tography or direct mixing in solution (see Experimental procedures). All protein ⁄ frag- ment samples were tested at 7 l M.Asa control, OA was tested at 100 l M. After incubation for 6 h, cell death by apoptosis was evaluated by appropriate changes of nuclei stained with Hoechst-33258 (early apoptotic cells, gray bars) and propidium iodide (late apoptotic ⁄ necrotic cells, open bars). The test was also conducted on a-LA, OA, NaCl ⁄ P i , and the medium (CT) as a con- trol. Data are shown as percentage of BAM- LET activity. Values are means ± standard deviation of at least three experiments. Oleic acid complexes of a-lactalbumin fragments S. Tolin et al. 168 FEBS Journal 277 (2010) 163–173 ª 2009 The Authors Journal compilation ª 2009 FEBS and hydrophobic interactions [48–50]. A reduction of the electrostatic repulsion between negative charges at the surface of the detergent aggregates and positively charged amino acid side chains of a protein allows the formation of aggregates at a lower concentration. However, considering that the apo form of a-LA is negatively charged, it may well be that the interaction with OA aggregates of this protein in its Ca 2+ -free form is mostly mediated by hydrophobic interactions, as the apo-a-LA is more hydrophobic than the holo form [51,52]. Nonetheless, even the negatively charged protein molecule may possess positively charged clus- ters or areas that mediate the interaction with the neg- atively charged head groups of OA aggregates, as, for example, indicated by the fact that the negatively charged a-synuclein contributes to CAC depression of anionic surfactants [48]. In the 123 residue chain of a-LA at neutral pH, Lys and Arg residues are clus- tered at the level of helical segments A, C and D of the native protein, whereas the central region of the protein contains many negatively charged carboxylates of Asp and Glu residues. Therefore, it could be that fragment 53–103 interacts with the negatively charged OA aggregates less effectively than the other fragments investigated here, as shown by the results of turbidi- metric analyses (Fig. 3D). The mechanisms of biological activity of the OA complexes of a-LA are not yet understood and, in fact, a variety of diverse biological effects have been described for HAMLET [20,25,27]. For this reason, HAMLET was metaphorically named ‘Hydra’ [25]. Besides the cytotoxicity via an apoptotic mechanism, an OA complex of human a-LA was shown to also possess bactericidal activity against Streptococcus pneu- moniae and Haemophilus influenzae [53]. It is of interest that digestion of a-LA with trypsin and chymotrypsin yields three peptides displaying bactericidal activity against Gram-positive bacteria. These bactericidal spe- cies are peptide 1–5 and the two-chain peptides linked by a disulfide bridge, 17–31 ⁄ 109–114 and 61–68 ⁄ 75–80 [54]. However, the structural features responsible for their bactericidal action were not clarified. Probably, bactericidal action of the LA–OA complex requires a different molecular mechanism than that occurring in apoptosis. Hence, HAMLET-like complexes can be detrimental by various cellular pathways, and exert their actions by different molecular mechanisms. The proteolytic fragments of a-LA investigated here have widely differing chain lengths and amino acid sequences (Fig. 1), and it therefore does not seem possible to explain their cytotoxicity in terms of their specific structural features. The variability in struc- ture of the polypeptide chain in forming active OA complexes seems to indicate instead that a generic poly- peptide chain can eventually interact with OA, and thus that the toxic action of an OA complex resides in the fatty acid rather than in the protein moiety. The pres- ent results show that OA displays new physicochemical and aggregation properties in the presence of a-LA or its fragments. With a decrease in the CAC of the fatty acid in the presence of the protein or its fragments, soluble and smaller aggregates of protein–OA or frag- ment–OA complexes are easily formed and stabilized. In previous studies, the tumor-selective cytotoxicity of HAMLET or BAMLET was correlated with the conformational properties of a-LA upon formation of the OA complex [17,18]. In particular, it was proposed that the fatty acid acts as a stabilizer of a partially folded or MG conformation of the protein under phys- iological conditions [21]. In a very recent paper, it was reported that a recombinant mutant a-LA with all eight Cys residues replaced by Ala residues (named all- Ala mutant), and thus devoid of the four disulfide bridges of the native protein, formed a cytotoxic OA complex equivalent to HAMLET [55]. Even if the con- formation of the all-Ala mutant at neutral pH is simi- lar to the MG of a-LA at low pH [6–9], the addition of OA to the all-Ala protein is required in order to form a cytotoxic species, indicating that the fatty acid is needed for the development of cytotoxicity [55]. Here, we show that a variety of a-LA fragments can mimic the action of the entire 123 residue chain of the protein in forming OA complexes displaying cytotoxic- ity. A reasonable deduction from this and previous studies is that the protein ⁄ peptide moiety can act as a carrier of the inherently toxic fatty acid [33], and there- fore that OA itself is the active species of a cytotoxic protein ⁄ peptide complex. This view is in line with the fact that all variants of a-LA of human, bovine, equine, porcine and caprine origin, as well as recombi- nant mutants of a-LA devoid of Ca 2+ -binding proper- ties, were all able to form HAMLET-like complexes with little difference in biological activity [21,24]. Inter- estingly, it was recently reported that the OA complex of lysozyme displays cellular toxicity similar to that of HAMLET [56]. In our laboratory, we have performed initial experiments indicating that even the 153 residue chain of apomyoglobin can form cytotoxic complexes when combined with OA [57]. In summary, the results of this study indicate that, besides substantial variation in amino acid sequence of the polypeptide chain of a-LA, severe truncation of the polypeptide chain of the protein is also tolerated in the formation of biologically active OA complexes. Therefore, we are inclined to conclude that the poly- peptide moiety can serve mainly as a carrier of the S. Tolin et al. Oleic acid complexes of a-lactalbumin fragments FEBS Journal 277 (2010) 163–173 ª 2009 The Authors Journal compilation ª 2009 FEBS 169 fatty acid. We have shown here that the addition of a protein ⁄ fragment species strongly influences the aggre- gation behavior of OA, in particular making it more water-soluble and thus enhancing its intrinsic apoptotic effects [33]. Nevertheless, we cannot exclude the possi- bility that the protein itself can act in a synergic way in the observed cytotoxicity of the OA complexes. This could be particularly true in the case of OA complexes of a-LA, considering that: (a) a-LA itself can display an inherent apoptotic activity [58,59]; (b) a-LA alone can interact with histones at the cellular level and thus display cytotoxic effects [29,30]; and (c) even various fragments of a-LA have been shown to have bacterici- dal activity, and a-LA fragments can therefore be toxic [54]. Despite these caveats regarding the specific role of the protein moiety in HAMLET-like or BAMLET-like species, it should be emphasized that the beneficial effects of these OA complexes in selective killing a variety of tumor cells appear to be remarkable and will prompt additional studies on OA–protein ⁄ peptide complexes as possible new anticancer agents. Experimental procedures Materials Bovine a-LA and DEAE-Trisacryl M resin were purchased from Sigma (St Louis, MO, USA); OA and the fluorescent dye TNS were from Fluka (Buchs, Switzerland). All other chemicals were of analytical reagent grade and were Sigma or Fluka products. Preparation of the OA complexes of a-LA fragments The a-LA fragments investigated, 1–40 ⁄ 53–123, 1–40 ⁄ 104– 123 and 53–103 (Fig. 1A), were produced by limited prote- olysis of the protein with pepsin at pH 2.0 [9,31]. The OA complexes of a-LA and its fragments were prepared follow- ing two procedures, column chromatography and mixing in solution. Column chromatography The protein material was loaded onto an OA-conditioned anion exchange chromatographic column (1.0 · 7.0 cm), following the procedure reported by Svensson et al. [18]. A DEAE-Trisacryl M resin was employed, equilibrated with 10 mm Tris ⁄ HCl and 0.1 m NaCl (pH 8.5). An aliquot of the protein or fragment material ( 2mgÆmL )1 ) was dis- solved in 10 mm Tris ⁄ HCl (pH 8.5), containing 1 mm EDTA, and then loaded onto the anion exchange column, which was eluted with a gradient of 10 mm Tris ⁄ HCl and 1 m NaCl (pH 8.5). The absorbance of the effluent from the column was monitored at 214 nm. The high-salt eluates from the column containing the OA complexes were desalt- ed by dialysis against water, using a membrane of 3.5 kDa cut-off, and then lyophilized. Mixing in solution The OA complexes were prepared by direct mixing of pro- tein species with 10 or 15 equivalents of OA dissolved (20 mgÆmL )1 ), in ethanol and then diluted with NaCl ⁄ P i (8 mm Na 2 HPO 4 , 137 mm NaCl, 2 mm KH 2 PO 4 , 2.7 mm KCl, pH 7.4) [35]. The fatty acid was added to the protein solution, and the mixture was analyzed after 1 h of incuba- tion in the dark. CD spectroscopy CD spectra were recorded on a Jasco J-710 spectropolarim- eter (Tokyo, Japan). The spectra were recorded in NaCl ⁄ P i (pH 7.4), in the absence or presence of OA, at a pro- tein ⁄ fragment concentration of 0.05–0.1 mgÆmL )1 , using 1 mm quartz cells. The interaction of OA with protein spe- cies was followed by far-UV CD measurements by adding aliquots of an OA solution to the protein fragment samples (10 lm) in NaCl ⁄ P i (pH 7.4). Mean residue ellipticity [h]is reported as degÆcm 2 Ædmol )1 . Protein fragment concentra- tions were determined by absorption measurements at 280 nm on a double-beam Lambda-25 spectrophotometer (Perkin-Elmer, Norwalk, CT, USA). The molar extinction coefficients at 280 nm for a-LA fragments were 1.22 mg )1 Æcm )1 for fragment 53–103, 2.89 mg )1 Æcm )1 for fragment 1–40 ⁄ 104–123, and 2.23 mg )1 Æcm )1 for fragment 1–40 ⁄ 53–123, as calculated according to Gill and von Hippel [60]. Determination of the aggregation state of OA The CAC of OA was determined by using the fluorescent dye TNS [43]. The TNS (20 lm) fluorescence emission at 460 nm, after excitation at 360 nm, was measured at 25 °C in the presence of increasing concentrations of OA in NaCl ⁄ P i (pH 7.4). The analyses were conducted in the absence or presence of a-LA or its fragments at 10 lm. Three readings were taken, and the average fluorescence intensities relative to blanks were plotted. The first and the last five data points were joined separately by statistically fitted straight lines. The CAC is defined as the lipid concen- tration at which the two linear portions of the fluorescence emission intensity lines intersect [61]. The aggregation state of OA, in the absence or presence of a-LA or its fragments (10 lm), was also analyzed by turbidity measurements at 400 nm of different samples containing increasing amounts of OA (from 0 to 500 lm) in NaCl ⁄ P i (pH 7.4) [44]. Oleic acid complexes of a-lactalbumin fragments S. Tolin et al. 170 FEBS Journal 277 (2010) 163–173 ª 2009 The Authors Journal compilation ª 2009 FEBS Apoptosis assays Cell culture T-lymphoblastoid Jurkat cells were cultured in RPMI-1640 medium supplemented with 10% heat-inacti- vated fetal bovine serum, 2 mm glutamine, 100 IUÆmL )1 penicillin and 100 lgÆmL )1 streptomycin in 5% CO 2 ⁄ 95% air at 37 °C. The Jurkat cells (10 6 cells per mL) were incu- bated with the protein ⁄ fragment samples in serum-free medium for 6 h at 37 °C. These samples were tested at 7 lm, and the OA complexes were prepared by mixing 10 molar equivalents of OA for fragment 1–40 ⁄ 104–123 and 15 equivalents for fragments 1–40 ⁄ 53–123 and 53–103, as well as intact a-LA. In order to assess cell viability, Jurkat cells were stained with 10 lm Hoechst-33258 and 1 lm pro- pidium iodide for 5 min, in order to allow visualization of early and late apoptotic ⁄ necrotic cells, respectively. Cells were then washed with Hanks’ balanced salt solution, and visualized with an Olympus IMT-2 inverted microscope equipped with a xenon lamp and a 12-bit digital, cooled, charge-coupled device camera (Princeton Instruments, Monmouth Junction, NJ, USA). Excitation ⁄ emission cubes of 340 ⁄ 440 ± 25 nm and a 568 ⁄ 585 ± 25 nm long-pass fil- ter were used for Hoechst-33258 and propidium iodide, respectively. Three randomly selected fields were acquired for each treatment. The corresponding bright field images were also acquired, and the three channels were overlaid using the appropriate function of metamorph software (Universal Imaging, West Chester, PA, USA). The percent- age of cell death and the standard deviation were calculated from three acquisitions of each treatment. The data are reported as percentage of BAMLET activity. Acknowledgements We gratefully acknowledge the financial support of the Italian Ministry of University and Research (MIUR) through PRIN-2004, PRIN-2006 and the FIRB Project on Protein Misfolding and Aggregation (Project No. RBNEOPX83). We thank M. Zambonin for his excel- lent technical assistance. This work was presented at the Symposium on HAMLET (12–14 May 2009, Lund, Sweden) and at the XXI Symposium of the Protein Society (21–25 July, 2007, Boston, MA, USA) [Protein Sci 16 (Suppl. 1), Commun. 257]. References 1 Hill RL & Brew K (1975) Lactose synthetase. Adv Enzymol 43, 411–490. 2 Kuwajima K (1996) The molten globule state of a-lactalbumin. FASEB J 10, 102–109. 3 Fink AL (1995) Compact intermediate states in protein folding. Annu Rev Biophys Biomol Struct 24, 495–522. 4 Pike AC, Brew K & Acharya KR (1996) Crystal struc- tures of guinea-pig, goat and bovine alpha-lactalbumin highlight the enhanced conformational flexibility of regions that are significant for its action in lactose syn- thase. Structure 4, 691–703. 5 Permyakov EA & Berliner LJ (2000) Alpha-lactalbu- min: structure and function. FEBS Lett 473, 269–274. 6 Dolgikh DA, Gilmanshin RI, Brazhnikov EV, Bychk- ova VE, Semisotnov GV, Venyaminov SY & Ptitsyn OB (1981) a-Lactalbumin: compact state with fluctuat- ing tertiary structure?. FEBS Lett 136, 311–315. 7 Alexandrescu AT, Evans PA, Pitkeathly M, Baum J & Dobson CM (1993) Structure and dynamics of the acid- denatured molten globule state of alpha-lactalbumin: a two-dimensional NMR study. Biochemistry 32, 1707– 1718. 8 Ptitsyn OB (1995) Molten globule and protein folding. Adv Protein Chem 47, 83–229. 9 Polverino de Laureto P, De Filippis V, Di Bello M, Zambonin M & Fontana A (1995) Probing the molten globule state of alpha-lactalbumin by limited proteoly- sis. Biochemistry 34, 12596–12604. 10 Ban ˜ uelos S & Muga A (1995) Binding of molten glob- ule-like conformations to lipid bilayers. Structure of native and partially folded alpha-lactalbumin bound to model membranes. J Biol Chem 270, 29910–29915. 11 Cawthern KM, Permyakov E & Berliner LJ (1996) Membrane-bound states of alpha-lactalbumin: implica- tions for the protein stability and conformation. Protein Sci 5, 1394–1405. 12 Grishchenko VM, Kalinichenko LP, Deikus GY, Veprintsev DB, Cawthern KM, Berliner LJ & Permya- kov EA (1996) Interactions of alpha-lactalbumins with lipid vesicles studied by tryptophan fluorescence. Biochem Mol Biol Int 38, 453–466. 13 Halskau O, Froystein NA, Muga A & Martinez A (2002) The membrane-bound conformation of a-lactal- bumin studied by NMR-monitored 1 H exchange. J Mol Biol 321, 99–110. 14 Agasøster AV, Halskau O, Fuglebakk E, Frøystein NA, Muga A, Holmsen H & Martınez A (2003) The interac- tion of peripheral proteins and membranes studied with a-lactalbumin and phospholipid bilayers of various compositions. J Biol Chem 278, 21790–21797. 15 Cawthern KM, Narayan M, Chaudhuri D, Permyakov EA & Berliner LJ (1997) Interactions of alpha-lactalbu- min with fatty acids and spin label analogs. J Biol Chem 272, 30812–30816. 16 Ha ¨ kansson A, Zhivotovsky B, Orrenius S, Sabharwal HR & Svanborg C (1995) Apoptosis induced by a human milk protein. Proc Natl Acad Sci USA 92, 8064– 8068. 17 Svensson M, Sabharwal H, Ha ¨ kansson A, Mossberg AK, Lipniunas P, Leffler H, Svanborg C & Linse S (1999) Molecular characterization of a-lactalbumin fold- ing variants that induce apoptosis in tumor cells. J Biol Chem 274, 6388–6396. S. Tolin et al. Oleic acid complexes of a-lactalbumin fragments FEBS Journal 277 (2010) 163–173 ª 2009 The Authors Journal compilation ª 2009 FEBS 171 18 Svensson M, Ha ¨ kansson A, Mossberg AK, Linse S & Svanborg C (2000) Conversion of a-lactalbumin to a protein inducing apoptosis. Proc Natl Acad Sci USA 97, 4221–4226. 19 Svensson M, Mossberg AK, Pettersson J, Linse S & Svanborg C (2003) Lipids as cofactors in protein fold- ing: stereo-specific lipid–protein interactions are required to form HAMLET (human alpha-lactalbumin made lethal to tumor cells). Protein Sci 12, 2805–2814. 20 Svanborg C, Agerstam H, Aronson A, Bjerkvig R, Du ¨ ringer C, Fischer W, Gustafsson L, Hallgren O, Leijonhuvud I, Linse S et al. (2003) HAMLET kills tumor cells by an apoptosis-like mechanism: cellular, molecular and therapeutic aspects. Adv Cancer Res 88, 1–29. 21 Svensson M, Fast J, Mossberg AK, Duringer C, Gustafsson L, Hallgren O, Brooks CL, Berliner L, Linse S & Svanborg C (2003) a-Lactalbumin unfolding is not sufficient to cause apoptosis, but is required for the conversion to HAMLET (human alpha-lactalbumin made lethal to tumor cells). Protein Sci 12, 2794–2804. 22 Bernback S, Blackberg L & Hernell O (1990) The com- plete digestion of human milk triacylglycerol in vitro requires gastric lipase, pancreatic colipase-dependent lipase and bile salt-stimulated lipase. J Clin Invest 85, 1221–1226. 23 Sarles J, Moreau H & Verger R (1992) Human gastric lipase: ontogeny and variations in children. Acta Paedi- atr 81, 511–513. 24 Pettersson J, Mossberg AK & Svanborg C (2006) a-Lactalbumin species variation, HAMLET formation and tumor cell death. Biochem Biophys Res Commun 345, 260–270. 25 Mok KH, Pettersson J, Orrenius S & Svanborg C (2007) HAMLET, protein folding, and tumor cell death. Biochem Biophys Res Commun 354, 1–7. 26 Knyazeva EL, Grishchenko VM, Fadeev RS, Akatov VS, Permyakov SE & Permyakov EA (2008) Who is Mr HAMLET? Interaction of human a-lactalbumin with monomeric oleic acid. Biochemistry 47, 13127– 13137. 27 Pettersson–Kastberg J, Aits S, Gustafsson L, Mossberg A, Storm P, Trulsson M, Persson F, Mok HK & Svan- borg C (2008) Can misfolded proteins be beneficial? The HAMLET case. Ann Med 41, 162–176. 28 Duringer C, Hamiche A, Gustafsson L, Kimura H & Svanborg C (2003) HAMLET interacts with histones and chromatin in tumor cell nuclei. J Biol Chem 278, 42131–42135. 29 Permyakov SE, Pershikova IV, Khokhlova TI, Uversky VN & Permyakov EA. (2004) No need to be HAMLET or BAMLET to interact with histones: binding of monomeric alpha-lactalbumin to histones and basic poly-amino acids. Biochemistry 43, 5575–5582. 30 Permyakov SE, Pershikova IV, Zhadan AP, Goers J, Bakunts AG, Uversky VN, Berliner LJ & Permyakov EA (2005) Conversion of human alpha-lactalbumin to an apo-like state in the complexes with basic poly- amino acids: toward understanding of the molecular mechanism of antitumor action of HAMLET. J Prote- ome Res 4, 564–569. 31 Polverino de Laureto P, Scaramella E, Frigo M, Gefter- Wondrich F, De Filippis V, Zambonin M & Fontana A (1999) Limited proteolysis of bovine a-lactalbumin: iso- lation and characterization of protein domains. Protein Sci 8, 2290–2303. 32 Polverino de Laureto P, Vinante D, Scaramella E, Frare E & Fontana A (2001) Stepwise proteolytic removal of the b-subdomain in a-lactalbumin: the protein remains folded and can form the molten globule in acid solu- tion. Eur J Biochem 286, 4324–4333. 33 Cury-Boaventura MF, Pompe ´ ia C & Curi R (2004) Comparative toxicity of oleic acid and linoleic acid on Jurkat cells. Clin Nutr 23, 721–732. 34 Fast J, Mossberg AK, Svanborg C & Linse S (2005) Stability of HAMLET – a kinetically trapped alpha- lactalbumin oleic acid complex. Protein Sci 14, 329–340. 35 Polverino de Laureto P, Frare E, Gottardo R & Fon- tana A (2002) Molten globule of bovine a -lactalbumin at neutral pH induced by heat, trifluoroethanol and oleic acid: a comparative analysis by circular dichroism spec- troscopy and limited proteolysis. Proteins 49, 385–397. 36 Yang F Jr, Zhang M, Chen J & Liang Y (2006) Struc- tural changes of a-lactalbumin induced by low pH and oleic acid. Biochim Biophys Acta 1764 , 1389–1396. 37 Zhang M, Yang F Jr, Yang F, Chen J, Zheng CY & Liang Y (2009) Cytotoxic aggregates of alpha-lactalbumin induced by unsaturated fatty acid induce apoptosis in tumor cells. Chem Biol Interact 180, 131–142. 38 Woody RW (1995) Circular dichroism. Methods Enzymol 246, 34–71. 39 Cistola DP, Hamilton J, Jackson D & Small D (1988) Ion- ization and phase behavior of fatty acids in water: applica- tion of Gibbs phase rule. Biochemistry 27, 1881–1888. 40 Edwards K, Silvander M & Karlsson G (1995) Aggre- gate structure in dilute aqueous dispersions of oleic acid ⁄ sodium oleate and oleic acid ⁄ sodium oleate ⁄ egg phosphatidylcholine. Langmuir 11, 2429–2434. 41 Rogerson ML, Robinson BH, Bucak S & Walde P (2006) Kinetic studies of the interaction of fatty acids with phosphatidylcholine vesicles (liposomes). Colloids Surf B Biointerfaces 48, 24–34. 42 Morigaki K & Walde P (2007) Fatty acid vesicles. Curr Opin Colloid Interface Sci 12, 75–80. 43 Das AK & Hajra AK (1992) Critical micellar concen- trations of palmitoyl dihydroxyacetone phosphate and 1-palmitoyl-rac-glycerol 3-phosphate. J Biol Chem 267, 9731–9737. 44 Namani T & Walde P (2005) From decanoate micelles to decanoic acid ⁄ dodecylbenzenesulfonate vesicles. Langmuir 21, 6210–6219. Oleic acid complexes of a-lactalbumin fragments S. Tolin et al. 172 FEBS Journal 277 (2010) 163–173 ª 2009 The Authors Journal compilation ª 2009 FEBS [...]... (2009) Protein oligomerization induced by oleic acid at the solid–liquid interface – equine lysozyme cytotoxic complexes FEBS J 276, 3975–3989 Pinato O, Spolaore B, Canton M, Polverino de Laureto P & Fontana A (2008) The interaction of apomyoglobin with oleic acid leads to a protein complex that displays cellular toxicity 53rd National Meeting of the Italian Society of Biochemistry and Molecular Biology,... far-UV CD spectra of fragments 1–40 and 104–123 in the presence of increasing amounts of OA Bottom: cytotoxicity of the OA complexes of peptides 1–40 and 104–123 This supplementary material can be found in the online version of this article Please note: As a service to our authors and readers, this journal provides supporting information supplied by the authors Such materials are peer-reviewed and may... effect on oleic acid ⁄ oleate vesicles J Liposome Res 16, 143–154 51 Wijesinha-Bettoni R, Dobson CM & Redfield C (2001) Comparison of the structural and dynamical properties of holo and apo bovine alpha-lactalbumin by NMR spectroscopy J Mol Biol 307, 885–898 52 Chrysina ED, Brew K & Acharya KR (2000) Crystal structures of apo- and holo-bovine alpha-lactalbumin at ˚ 2.2-A resolution reveal an effect of calcium... Svanborg C & Mok KH (2009) Alpha-Lactalbumin, engineered to be nonnative and inactive, kills tumor cells when in complex with oleic acid: a new biological function resulting from partial unfolding J Mol Biol doi:10.1016/ j.jmb.2009.09.026 56 Wilhelm K, Darinskas A, Noppe W, Duchardt E, Mok ´ KH, Vukojevic V, Schleucher J & Morozova-Roche LA Oleic acid complexes of a-lactalbumin fragments 57 58 59 60 61 (2009)... amino acid sequence data Anal Biochem 182, 319–326 Horowitz P (1977) A comparison between 8-anilinonaphthalene-1-sulfonate and 2-p-toluidinylnaphthalene6-sulfonate as fluorescent indicators of the critical micelle concentration of sodium dodecyl sulfate J Colloid Interface Sci 61, 197–198 Supporting information The following supporting information is available: Fig S1 Top: far-UV CD spectra of fragments. .. (http://www.fupress.com/Archivio /pdf% 5C2923 .pdf) Sternhagen LG & Allen JC (2001) Growth rates of a human colon adenocarcinoma cell line are regulated by the milk protein alpha-lactalbumin Adv Exp Med Biol 501, 115–120 Xu M, Sugiura Y, Nagaoka S & Kanamaru Y (2005) IEC-6 intestinal cell death induced by bovine milk alpha-lactalbumin Biosci Biotechnol Biochem 69, 1082– 1089 Gill SG & von Hippel PH (1989) Calculation of protein... H, Linse S, Lazou I, Lonnerdal B & Svanborg C (2000) ¨ A folding variant of alpha-lactalbumin with bactericidal activity against Streptococcus pneumoniae Mol Microbiol 35, 589–600 54 Pellegrini A, Thomas U, Bramaz N, Hunziker P & von Fellenberg R (1999) Isolation and identification of three bactericidal domains in the bovine alpha-lactalbumin molecule Biochim Biophys Acta 1426, 439–448 55 Pettersson-Kastberg... (1999) Amphitropic proteins: regulation by reversible membrane interactions Mol Membr Biol 16, 217–235 46 Cornell RB & Taneva SG (2006) Amphipathic helices as mediators of the membrane interaction of amphitropic proteins, and as modulators of bilayer physical properties Curr Protein Pept Sci 7, 539–552 47 Lee GA (2005) How lipids and proteins interact in a membrane: a molecular approach Mol Biosyst 1,... peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset Technical support issues arising from supporting information (other than missing files) should be addressed to the authors FEBS Journal 277 (2010) 163–173 ª 2009 The Authors Journal compilation ª 2009 FEBS 173 . The oleic acid complexes of proteolytic fragments of a-lactalbumin display apoptotic activity Serena Tolin 1 , Giorgia. (dashed lines). The solid bars indicate the fractions of the effluent from the column that were collected for further studies. Oleic acid complexes of a-lactalbumin fragments