Minireview
`K-Lactalbumin: structure and function
`
`Eugene A. Permyakova;*, Lawrence J. Berlinerb
`
`aInstitute for Biological Instrumentation of the Russian Academy of Sciences, 142292 Pushchino, Moscow region, Russia
`bDepartment of Chemistry, The Ohio State University, Columbus, OH 43210, USA
`
`Received 6 March 2000; received in revised form 14 April 2000
`
`Edited by Vladimir Skulachev
`
`Abstract Small milk protein KK-lactalbumin (KK-LA), a compo-
`nent of lactose synthase, is a simple model Ca2+ binding protein,
`which does not belong to the EF-hand proteins, and a classical
`example of molten globule state. It has a strong Ca2+ binding
`site, which binds Mg2+, Mn2+, Na+, and K+, and several distinct
`Zn2+ binding sites. The binding of cations to the Ca2+ site
`increases protein stability against action of heat and various
`denaturing agents, while the binding of Zn2+ to the Ca2+-loaded
`protein decreases its stability. Functioning of KK-LA requires its
`interactions with membranes, proteins, peptides and low
`molecular weight substrates and products. It was shown that
`these interactions are modulated by the binding of metal cations.
`Recently it was found that some folding variants of KK-LA
`demonstrate bactericidal activity and some of
`them cause
`apoptosis of tumor cells.
`z 2000 Federation of European Biochemical Societies.
`
`[2,3] and for this reason it is frequently used as a simple,
`model Ca2(cid:135) binding protein. It is very convenient for studies
`of calcium binding e¡ects on interactions of the protein with
`proteins, peptides, membranes and low molecular weight or-
`ganic compounds, which frequently have physiological signi¢-
`cance.
`Third, K-LA has several partially folded intermediate states,
`which are being studied by many researchers interested in
`protein folding problems. It is very attractive for studies of
`the properties and structure of intermediate molten globule-
`like states since at acidic pH and in the apo-state at elevated
`temperatures K-LA is the classic ‘molten globule’ [3^5].
`Fourth, it has been found recently that some forms of K-LA
`can induce apoptosis in tumor cells [6,7] which suggests that
`this protein can ful¢ll many important biological functions.
`
`Key words: K-Lactalbumin; Structure; Function;
`Metal cation binding
`
`2. Primary, secondary and tertiary structure
`
`1. Introduction
`
`K-Lactalbumin (K-LA) is a small (Mr 14 200), acidic (pI 4^
`5), Ca2(cid:135) binding milk protein, which is very important from
`several points of view. First of all, K-LA performs an impor-
`tant function in mammary secretory cells: it is one of the two
`components of lactose synthase, which catalyzes the ¢nal step
`in lactose biosynthesis in the lactating mammary gland [1].
`The other component of this system is galactosyltransferase
`(GT), which is involved in the processing of proteins in var-
`ious secretory cells by transferring galactosyl groups from
`UDP-galactose to glycoproteins containing N-acetylglucos-
`amine. In the lactating mammary gland, the speci¢city of
`GT is modulated by interaction with K-LA, which increases
`its a⁄nity and speci¢city for glucose:
`UDP-Gal (cid:135) glucose (cid:255)!GT=K-LA lactose (cid:135) UDP
`
`The reaction takes place in the Golgi
`Mn2(cid:135) ions.
`Second, K-LA possesses a single strong Ca2(cid:135) binding site
`
`lumen and requires
`
`Most of K-lactalbumins, including human, guinea pig, bo-
`vine, goat, camel, equine and rabbit proteins, consist of 123
`amino acid residues (see for example [8]). Only rat K-LA con-
`tains 17 additional C-terminal residues. K-LA is homologous
`in sequence to the lysozyme family, but it exhibits cell lytic
`activity about 1036 of the speci¢c activity of hen egg white
`lysozyme [9]. X-ray crystallography has shown that the three-
`dimensional structure of K-LA is very similar to that of lyso-
`zyme [10,11]. Native K-LA consists of two domains: a large
`K-helical domain and a small L-sheet domain, which are con-
`nected by a calcium binding loop (Fig. 1). The K-helical do-
`main is composed of three major K-helices (residues 5^11, 23^
`24, and 86^98) and two short 310 helices (residues 18^20, and
`115^118). The small domain is composed of a series of loops,
`a small three-stranded antiparallel L-pleated sheet (residues
`41^44, 47^50, and 55^56) and a short 310 helix (three residues
`per turn and an intrachain hydrogen bond loop containing 10
`atoms; residues 77^80). The two domains are divided by a
`deep cleft between them. At the same time, the two domains
`are held together by the cysteine bridge between residues 73
`and 91, forming the Ca2(cid:135) binding loop. A second important
`disul¢de bridge 61^77 connects the two domains as well.
`Overall, the structure of K-LA is stabilized by four disul¢de
`bridges (6^120, 61^77, 73^91, and 28^111).
`
`3. Location of cation binding sites
`
`One of the most interesting features of K-LA is its ability to
`bind metal cations. It does not belong to the EF-hand protein
`family. The protein has a single strong calcium binding site,
`which is formed by oxygen ligands from carboxylic groups of
`
`0014-5793 / 00 / $20.00 (cid:223) 2000 Federation of European Biochemical Societies. All rights reserved.
`PII: S 0 0 1 4 - 5 7 9 3 ( 0 0 ) 0 1 5 4 6 - 5
`
`

`

`270
`
`E.A. Permyakov, L.J. Berliner/FEBS Letters 473 (2000) 269^274
`
`Fig. 1. X-ray K-LA structure derived from native bu¡alo (the structure was kindly presented by Dr. Acharya) and recombinant bovine protein
`(the structure was taken from Brookhaven Protein Data Bank) prepared in conjunction with Serge E. Permyakov from the Institute for Biolog-
`ical Instrumentation, Pushchino, Russia and Dr. Charles Brooks, Department of Veterinary Biosciences, Ohio State University, Columbus, OH,
`USA. K-Domain is shown in blue while L-domain is shown in green. Trp residues are shown in blue and S^S bridges are shown in yellow. The
`residues which take part in coordination of Zn2(cid:135) ions are shown in red.
`
`three Asp residues (82, 87 and 88) and two carbonyl groups of
`the peptide backbone (79 and 84) in a loop between two
`helices. The loop contains two residues less than the typical
`EF-hand Ca2(cid:135) binding domain. In addition, one or two water
`molecules take part in direct coordinating Ca2(cid:135). Overall the
`oxygen ligands form a distorted pentagonal bipyramidal
`structure. Recently a secondary calcium binding site was
`found by X-ray crystallography revealed in human K-LA 7.9
`A(cid:238) away from the primary strong calcium binding site [12].
`Four residues are involved in Ca2(cid:135) coordination at this site
`in a tetrahedral arrangement (Thr-38, Gln-39, Asp-83 and the
`carbonyl oxygen of Leu-81). This secondary site is located
`near the surface of the K-LA molecule.
`K-LA also has several zinc binding sites [13], one of which is
`located in the ‘cleft’ region (i.e. the region which forms the
`active site of lysozyme) [14]. In the X-ray structure of human
`K-LA the zinc is sandwiched between Glu-49 and Glu-116
`(Asp in the bovine protein) of the symmetry-related subunit
`in the dimeric crystal unit cell (Fig. 1). This site was assigned
`to as the strong zinc site in human K-LA. The intramolecular
`distance between the strong zinc and calcium sites in bovine
`
`Table 1
`Apparent binding constants of metal ions for bovine K-LA
`Association constants (M31)
`Cation
`37‡C
`20‡C
`2U107
`3U108
`3U108
`2000 (cid:254) 100; 200 (cid:254) 20
`100 (cid:254) 10
`8 (cid:254) 3
`
`211 (cid:254) 20; 46 (cid:254) 10
`36 (cid:254) 10
`6 (cid:254) 3
`
`Ca2(cid:135)
`Mn2(cid:135)
`Mg2(cid:135)
`Na(cid:135)
`K(cid:135)
`
`K-LA, as measured by Fo«rster £uorescence energy transfer,
`was 14^18 A(cid:238) utilizing Co(II) as acceptor and Tb(III) as donor,
`respectively [15]. This distance agreed well with the distance
`(17.5 A(cid:238) ) found in the human K-LA X-ray structure. Never-
`theless, recent studies on K-LA mutants showed that the
`strong Zn2(cid:135) binding site in solution is not consistent with
`the site presumed from the human K-LA X-ray structure
`and, in fact, appears to be located near the N-terminus of
`the protein: site directed mutagenesis of Glu-1 to a Met res-
`idue results in the disappearance of the strong Zn2(cid:135) binding
`site in bovine K-LA [16]. A proposed site involves Glu-1, Glu-
`7, Glu-11 and Asp-37 (Fig. 1). The distance between the
`strong Ca2(cid:135) binding site and this putative N-terminus Zn2(cid:135)
`binding site is about 14 A(cid:238) , which is in good agreement with
`the £uorescence energy transfer data [15]. In addition, there is
`evidence that some of the weak secondary Zn2(cid:135) sites in K-LA
`contain His residues [17,18].
`
`4. Calcium-induced conformation changes
`
`The binding of Ca2(cid:135) to K-LA causes pronounced changes
`in structure and function, mostly in tertiary, but not second-
`ary, structure, which is clearly seen from both £uorescence
`[3,19], and CD [20] data. Calcium binding results in both a
`tryptophan £uorescence blue shift and a decrease in £uores-
`cence quantum yield. Time-resolved £uorescence measure-
`ments demonstrated that the Ca2(cid:135)-induced e¡ects are due to
`changes in the environment of all emitting tryptophan resi-
`dues (four in bovine K-LA and three in human K-LA, Fig. 1)
`[21]. The pronounced £uorescence changes can be used as
`accurate monitors of the Ca2(cid:135) association constant, which is
`
`FEBS 23678 11-5-00 Cyaan Magenta Geel Zwart
`
`Page 2
`
`

`

`E.A. Permyakov, L.J. Berliner/FEBS Letters 473 (2000) 269^274
`
`very large and usually cannot be evaluated from direct Ca2(cid:135)
`titrations, but can be more accurately calculated from titra-
`tions of Ca2(cid:135)-loaded K-LA with a strong Ca2(cid:135) chelator [3].
`
`5. Equilibrium and kinetic metal ions binding constants
`
`Besides calcium, K-LA binds other physiologically signi¢-
`cant cations such as Mg2(cid:135), Mn2(cid:135), Na(cid:135) and K(cid:135), which can
`compete with Ca2(cid:135) for the same binding site [19,22]. They
`induce similar, albeit smaller, structural changes in K-LA.
`Table 1 lists comparative binding constants for these cations
`[19,23]. All these cations seem to bind to the calcium binding
`site.
`In Table 1 we note that two binding constants are listed for
`magnesium since K-LA appears to possess secondary binding
`sites for this cation. The values of the binding constants for
`Mg2(cid:135), Na(cid:135) and K(cid:135) ions are rather low but, taking into con-
`sideration their high concentrations in the cell, one might
`speculate that they could successfully compete with Ca2(cid:135)
`ions in vivo.
`The synthetic peptide corresponding to the residues 72^100
`of native K-LA contains the Ca2(cid:135) binding loop with a £ank-
`ing helix comprising residues 86^98 of the K-domain and the
`310 helix of the L-domain. If Cys-73, Cys-77 and Cys-91 are
`replaced by alanines, it binds Ca2(cid:135) very weakly (102 M31)
`[24]. On the other hand, formation of the native disul¢de
`bond between Cys-73 and Cys-91 does not increase its a⁄nity
`to Ca2(cid:135) in water but increases it in 50% tri£uoroethanol.
`According to £uorescence stopped £ow measurements, the
`values of dissociation rate constants for complexes of K-LA
`with Ca2(cid:135) and Mg2(cid:135) are practically the same and within 0.006
`to 1 s31 in the temperature region from 10 to 40‡C [25]. The
`association rate constants for Ca2(cid:135) and Mg2(cid:135) in this temper-
`ature region are within 106^107and 101^102 M31 s31, respec-
`tively. It is of interest that the association rate constant for the
`Ca2(cid:135)^K-LA complex is almost 1 to 2 orders of magnitude
`lower than the di¡usion-controlled limit, which is rather un-
`usual for calcium binding proteins. One of the possible rea-
`sons for this may be the existence of the S^S bridge connect-
`ing the ends of the Ca2(cid:135) binding loop in K-LA.
`
`6. Protein stability
`
`Cation binding to the strong calcium site increases the
`stability of K-LA. From di¡erential scanning calorimetry
`(DSC) data, the binding of Ca2(cid:135) shifts the thermal transition
`to higher temperatures by more than 40‡C [26,27]. The bind-
`ing of Mg2(cid:135), Na(cid:135) and K(cid:135) increases protein stability as well.
`The stronger an ion binds to the protein, the more pro-
`nounced is thermal transition shift.
`Surprisingly the binding of Zn2(cid:135) ions to Ca2(cid:135)-loaded K-LA
`decreases thermal stability, causes aggregation and increases
`its susceptibility to protease digestion [13,28]. The thermal
`transition for calcium-loaded K-LA occurs at room temper-
`atures at high zinc concentrations (Zn:protein molar ratio
`about 100). Overall the results also showed that K-LA is in
`a partially unfolded and partially aggregated state in the pres-
`ence of high [Zn2(cid:135)].
`In the absence of calcium ions, but in the presence of phys-
`iological concentrations of magnesium, sodium and potassium
`ions, the thermal transition in K-LA occurs in the region from
`about 30 to 45‡C [29]. This might be related to some temper-
`
`271
`
`ature regulation of K-LA stability and function in the mam-
`mary gland.
`The binding of metal cations also increases the stability of
`K-LA against the action of denaturing agents such as urea or
`guanidine hydrochloride [19]. Here, important features of the
`denaturation curves are distinct, intermediate molten globule-
`like states arising at intermediate denaturant concentrations.
`Remarkably the binding of calcium stabilizes K-LA against
`pressure as monitored by a 200-Mpa increase in the pressure
`where denaturation occurs [30]. Interestingly, calcium binding
`increases the pressure stability of the calcium binding loop to
`a higher degree than the pressure stability of the overall K-LA
`secondary structure.
`It is very important to point out that any denaturation
`transition in K-LA (temperature, pressure, denaturant concen-
`tration) depends upon metal ion concentration (especially that
`of calcium ion). Thus values such as denaturation temperature
`or urea or guanidine hydrochloride denaturing concentration
`are relatively meaningless for K-LA without specifying the
`metal ion content(s) and their solution concentration(s).
`Kuwajima et al. [31,32] studied the kinetics of refolding of
`apo-K-LA by stopped £ow pH jump experiments. The free
`refolding kinetics of the protein has a simple single exponen-
`tial character. Chaperone GroEL was shown to retard the
`refolding of apo-K-LA by interacting with the molten globule
`state of the protein. The binding constant between GroEL
`and an early folding intermediate of K-LA is the order of
`106 M31.
`Lastly refolding of bovine K-LA from its 6 M GuHCl de-
`natured state [33] indicates that folding of the protein with its
`four disulphide bonds intact corresponds to one of the limit-
`ing cases of protein folding in which rapid collapse to a glob-
`ule with native-like fold is followed by a search for native-like
`side-chain contacts that enable e⁄cient conversion to the close
`packed native structure.
`
`7. E¡ects of N-terminus mutants on protein properties
`
`Since we began studies of various mutant forms of K-LA, it
`was immediately apparent that wild-type recombinant bovine
`K-LA, which di¡ers from the milk isolated native protein by
`the addition of an N-terminal Met, has more accessible tryp-
`tophan residues, lower thermostability and decreased calcium
`a⁄nity compared to the native protein [34]. Enzymatic remov-
`al of the N-terminal Met restores the native properties of
`K-LA. Taken together, £uorescence, circular dichroism, and
`DSC results showed that recombinant wild-type K-LA in the
`absence of calcium ion is in a ‘molten globule-like’ state. The
`delta-E1 (or E1M) mutant, where the Glu-1 residue of the
`native sequence is genetically substituted, leaving an N-termi-
`nal methionine in its place after bacterial expression, shows
`almost one order of magnitude higher a⁄nity for calcium and
`higher thermostability (both in the absence and presence of
`calcium) than the milk isolated native protein.
`The e¡ect of the single mutation in the N-terminus of K-LA
`is very interesting for many reasons. The charged Glu-1 in
`bovine K-LA is located on the protein surface and gives no
`evident contribution to the tertiary structure formation (e.g.
`salt bridges, hydrogen bonds with some proteins groups). At
`the same time, being charged group,
`it contributes to the
`balance of electrostatic interactions in the protein. Further-
`more, as a hydrophilic residue, it strongly interacts with water
`
`FEBS 23678 11-5-00 Cyaan Magenta Geel Zwart
`
`Page 3
`
`

`

`272
`
`E.A. Permyakov, L.J. Berliner/FEBS Letters 473 (2000) 269^274
`
`and therefore it probably tends to unfold the protein struc-
`ture. The addition of Met to the N-terminus makes the sit-
`uation even worse and it causes its transition to the molten
`globule-like conformation. The removal of Glu-1 abolishes
`such tendency, that is why, probably, the mutant protein be-
`comes more stable. Perhaps, the Glu-1 residue in bovine K-LA
`is located in some critical region, which allows switch the
`protein structure from the highly rigid to the molten glob-
`ule-like.
`
`8. Acid transition
`
`At low pH values protons can compete with Ca2(cid:135) for the
`same carboxylate oxygens and displace calcium ions at very
`acid pH values. One can describe the competition between
`calcium ions and protons for the same binding site by a simple
`scheme:
`P (cid:135) Ca2(cid:135)IPCa
`P (cid:135) nH(cid:135)HPHn
`
`The ¢tting of the experimental £uorescence pH titration data
`by theoretical curves computed according to this scheme
`shows that the Ca2(cid:135) binding site in bovine K-LA contains
`three carboxylic groups with pKs about 5 [19], which agrees
`well with the structural data.
`
`9. Molten globule state
`
`The acid state of K-LA at low pH values, which is the
`classical molten globule state, was studied extensively by
`many researchers beginning with Dolgikh et al. [4], who de-
`¢ned it as a compact state with £uctuating tertiary structure.
`The radius of gyration of native Ca2(cid:135)-loaded K-LA is 15.7 A(cid:238) ,
`but the acid molten globule has a radius of 17.2 A(cid:238) [35]. Mol-
`ten globule K-LA still retains a globular shape, but is simply
`‘swollen’ from the native state. It is a highly hydrated state,
`containing about 270 bound water molecules; the intrinsic
`mass density of the swollen interior of the protein molecule
`is 5% less and the intrinsic compressibility coe⁄cient 2 times
`higher than that of native molecule [36]. An analysis of a set
`of point mutations in the helical domain of K-LA allowed the
`identi¢cation of a stabilizing hydrophobic core, which likely
`contains some native-like packing interactions [37]. A subset
`of hydrophobic residues is most important for formation of
`the native-like topology [38,39]. The most persistent structure
`in the molten globule is localized in the helical domain and
`the helices most protected from hydrogen exchange in the
`molten globule are less protected in the native state than other
`regions of the protein [40]. It is interesting to note that an
`K-LA mutant in which all eight cysteines were mutated to
`alanine, was nearly as compact as wild-type K-LA at acidic
`pH [41]. Overall the architecture of the protein fold of K-LA is
`determined by the polypeptide sequence itself and not as a
`result of disul¢de bond cross-linking.
`
`10. Binding of low molecular weight organic compounds and
`peptides
`
`K-LA interacts with various low molecular weight organic
`compounds and these interactions are modulated by cation
`
`binding. For example, K-LA binds UDP-galactose, the sub-
`strate of lactose synthase reaction, as well as UDP, and UTP
`[13,29]. The binding parameters depend upon the state of the
`protein, but the strongest binding constant for UDP-galactose
`falls in the range of 103 to 104 M31.
`K-LA binds melittin, a short peptide from bee venom [42],
`which is frequently used as a model target protein for calmod-
`ulin. In contrast to the other calcium binding proteins, such as
`calmodulin or troponin C, K-LA binds melittin only in the
`absence of calcium ions. Apo-K-LA binds melittin with the
`binding constant 5U107 M31. Binding alters the melittin con-
`formation from a random coil in solution to a helical struc-
`ture in the binary complex with apo-K-LA.
`K-LA possesses several classes of fatty acid binding sites. It
`binds 5-doxylstearic acid (DSA, spin-labeled fatty acid analog,
`C22H42NO4), stearic acid and palmitic acid. The binding pa-
`rameters depend upon the protein state. Apparent binding
`constants for DSA are in the range from 104 to 106 M31 [43].
`
`11. Interactions with membrane systems
`
`K-LA also interacts with lipid membranes [44^49]. Sepha-
`dex G-200 chromatography of a mixture of K-LA and
`DMPC, DPPC or lecithin vesicles reveals that a signi¢cant
`portion of the protein binds to the vesicles where it is possible
`to subsequently study some physical properties of the protein
`in this state. The intrinsic £uorescence of vesicle bound K-LA
`is sensitive to two thermal transitions: The ¢rst is the gel-
`liquid crystal transition of the lipid vesicles; the second arises
`from the denaturation of the protein. The £uorescence max-
`imum position suggests that, at low temperatures, tryptophan
`accessibility increases upon protein-vesicle association. Above
`the protein transition tryptophans appear to interact signi¢-
`cantly with the apolar phase of the vesicles. Quenching experi-
`ments also suggest
`that
`tryptophan accessibility increases
`upon protein-vesicle association. Thermal denaturation of
`the liposome bound K-LA depends on the metal bound state
`of the protein [44,45].
`At pH 2, where the protein rapidly inserts into the bilayer,
`the isolated vesicle^K-LA complex shows a distinct £uores-
`cence thermal transition, consistent with a partially inserted
`protein possessing some degree of tertiary structure that un-
`folds cooperatively [44]. This is in contrast with the behavior
`of acid state K-LA in solution.
`These results suggest a model where a limited expansion of
`conformation occurs upon membrane association at neutral
`pH, physiological temperature, with a concomitant increase in
`tryptophan exposure to solvent and external quenchers. The
`data may shed light on the in vivo function and mechanism of
`K-LA since it interacts with galactosyltransferase on mem-
`brane surfaces in the Golgi lumen.
`
`12. Functions of KK-LA
`
`It has long been known that K-LA is a component of lac-
`tose synthase [1]: it complexes with galactosyl transferase only
`in the presence of substrates and modi¢es its speci¢city.
`Nevertheless, the role of metal cations in lactose synthase
`function is still far from clear. One of the substrates, which
`binds to galactosyltransferase, UDP-galactose, also binds to
`K-LA, but with rather low a⁄nity [13,29]. It is unknown
`whether this binding is of any physiological signi¢cance or
`
`FEBS 23678 11-5-00 Cyaan Magenta Geel Zwart
`
`Page 4
`
`

`

`E.A. Permyakov, L.J. Berliner/FEBS Letters 473 (2000) 269^274
`
`not. Lactose synthase requires Mn2(cid:135) ions for optimal function
`and both galactosyl transferase and K-LA bind Mn2(cid:135) rather
`tightly.
`Surprisingly, the role of calcium binding to K-LA in lactose
`synthesis is still unclear. On the other hand, we have learned
`that Zn2(cid:135) binding to K-LA can modulate lactose synthase
`function and this may be physiologically signi¢cant
`[48].
`Zinc binding to K-LA changes both the apparent Michaelis
`constant Km(app) and Vmax of lactose synthase. These e¡ects
`depend upon manganese concentration as well [48]: Zn2(cid:135) in-
`duces a decrease in both Km(app) and Vmax for Mn2(cid:135), which
`results in an apparent increase, followed by a decrease, in
`lactose synthase activity at Mn2(cid:135) concentrations below satu-
`ration of the ¢rst Mn2(cid:135) binding site in GT. At high Mn2(cid:135)
`concentrations Zn2(cid:135) decreases lactose synthase activity.
`Pelligrini et al. [49] found that proteolytic digestion of K-LA
`by trypsin and chymotrypsin yields three peptides with bac-
`tericidal properties. The polypeptides are mostly active against
`Gram-positive bacteria suggesting a possible antimicrobial
`function of K-LA after its partial digestion by endopeptidases.
`Hakansson et al. [50] revealed an K-LA folding variant with
`bactericidal activity against antibiotic-resistant and antibiotic-
`susceptible strains of Streptococcus pneumoniae. The active
`form of K-LA was puri¢ed from casein by a combination of
`anion exchange and gel chromatography. It is of interest that
`native K-LA could be converted to the active bactericidal
`form by ion exchange chromatography in the presence of a
`cofactor from human milk casein, characterized as a C18:1
`fatty acid. As it was shown above, K-LA possesses several
`classes of fatty acid binding sites [43].
`Recent data of Hakansson et al. [6] and Svensson et al. [7]
`address one more most intriguing possible function of K-LA.
`They found that some multimeric, yet not thoroughly charac-
`terized human K-LA derivative is a potent Ca2(cid:135)-elevating and
`apoptosis-inducing agent with broad, yet selective, cytotoxic
`activity, killing all transformed, embryonic, and lymphoid
`cells tested. The multimeric K-LA forms were isolated from
`the casein fraction of milk [7]. It was found that apoptosis-
`inducing fraction of K-LA contains oligomers of K-LA that
`have undergone a conformational change toward a molten
`globule-like state. Oligomerization appears to conserve K-LA
`in a state with molten globule-like properties in physiological
`conditions. As outlined above,
`it is now well known that
`aggregated forms of K-LA can be obtained in the presence
`of zinc ions [13] and it is interesting whether or not they
`possess cytotoxic activity as well. Multimeric K-LA was
`shown to bind to the cell surface, to enter the cytoplasm
`and accumulate in cell nuclei [7]. It is in line with our ¢nding
`that apo- and Zn2(cid:135)-loaded K-LA (molten globule-like confor-
`mations) interact with model phospholipid membranes better
`than the Ca2(cid:135)-loaded protein [44]. Data of Ko«hler et al. [51]
`demonstrate that caspases (cysteine-containing aspartate-spe-
`ci¢c proteases) are activated and involved in apoptosis in-
`duced by aggregated K-LA and that direct interaction of
`K-LA with mitochondria leads to the release of cytochrome
`c, which may be an important step in the initiation of the
`caspase cascade in these cells.
`Kit et al. [52] showed that oligonucleotides from human
`milk block both the cytostatic and cytotoxic e¡ects of
`K-LA. They also found that both monomeric and multimeric
`K-LA binds oligonucleotides of various lengths [53]. They
`suggested that oligonucleotides are secreted from mammary
`
`273
`
`cells and that K-LA and endogenic oligonucleotides can serve
`as factors of regulation of physiological state of mammary
`gland cells. Moreover, oligonucleotides could control the cy-
`totoxic potential of milk.
`
`13. Concluding remarks
`
`It is intriguing that K-LA may potentially have a myriad of
`functions beyond its role in lactose biosynthesis. Yet practi-
`cally the only cation which exists in vivo at high enough
`concentrations so that is likely to be bound to the (intact)
`protein is calcium (some population of the protein can be
`loaded by sodium and potassium). While we are at a disad-
`vantage to date in being able to measure free unliganded
`concentrations of ions and metabolites in cellular compart-
`ments, most of the other cations are likely to be at too low
`level, including zinc. However, this does not address the bind-
`ing of e.g. zinc to aggregated forms of K-LA since the binding
`is certainly cooperative at higher zinc:protein levels. Hence
`the relationship between zinc binding and protein aggregation,
`as well as the susceptibility to aggregation of many molten
`globule state proteins, points to a mechanism for zinc binding
`to promote immobilization and transport of the protein for
`nutritional purposes. Recall also that membrane/lipid associ-
`ation with K-LA compromises the protein as well, placing it in
`a molten globule-like state. Consequently consider the fact
`that zinc or other cation binding might induce K-LA aggre-
`gation to forms that have anticancer activity, perform various
`transport functions with apolar, lipophilic vitamins and me-
`tabolites as well as serving as a detergent to apoptotic events.
`The data presented show that some proteins, even such small
`proteins as K-LA, can perform several physiological functions
`depending on its location.
`
`References
`
`[1] Hill, R.L. and Brew, K. (1975) Adv. Enzymol. 43, 411^490.
`[2] Hiraoka, Y., Segawa, T., Kuwajima, K., Sugai, S. and Murai, N.
`(1980) Biochem. Biophys. Res. Commun. 95, 1098^1104.
`[3] Permyakov, E.A., Yarmolenko, V.V., Kalinichenko, L.P., Moro-
`zova, L.A. and Burstein, E.A. (1981) Biochem. Biophys. Res.
`Commun. 100, 191^197.
`[4] Dolgikh, D.A., Gilmanshin, R.I., Brazhnikov, E.V., Bychkova,
`V.E., Semisotnov, G.V., Venyaminov, S.Y. and Ptitsyn, O.B.
`(1981) FEBS Lett. 136, 311^315.
`[5] Kuwajima, K. (1996) FASEB J. 10, 102^109.
`[6] Hakansson, A., Zhivotovsky, B., Orrenius, S., Sabharwal, H. and
`Svanborg, C. (1995) Proc. Natl. Acad. Sci. USA 92, 8064^8068.
`[7] Svensson, M., Sabharwal, H., Hakansson, A., Mossberg, A.K.,
`Lipniunas, P., Le¥er, H., Svanborg, C. and Linse, S. (1999)
`J. Biol. Chem. 274, 6388^6396.
`[8] Nitta, K. and Sugai, S. (1989) Eur. J. Biochem. 182, 111^118.
`[9] McKenzie, H.A. and White, F.H. (1987) Biochem. Int. 14, 347^
`356.
`[10] Acharya, K.R., Ren, J., Stuart, D.I., Phillips, D.C. and Fenna,
`R.E. (1991) J. Mol. Biol. 221, 571^581.
`[11] Acharya, K.R., Stuart, D.I., Phillips, D.C., McKenzie, H.A. and
`Teahan, C.G. (1994) J. Protein Chem. 13, 569^584.
`[12] Chandra, N., Brew, K. and Acharya, K.R. (1998) Biochemistry
`37, 4767^4772.
`[13] Permyakov, E.A., Shnyrov, V.L., Kalinichenko, L.P., Kuchar,
`A., Reyzer, I.L. and Berliner, L.J. (1991) J. Protein Chem. 10,
`577^584.
`[14] Ren, J., Stuart, D.I. and Acharya, K.R. (1993) J. Biol. Chem.
`268, 19292^19298.
`[15] Permyakov, E.A. and Berliner, L.A. (1994) J. Protein Chem. 13,
`277^281.
`
`FEBS 23678 11-5-00 Cyaan Magenta Geel Zwart
`
`Page 5
`
`

`

`274
`
`E.A. Permyakov, L.J. Berliner/FEBS Letters 473 (2000) 269^274
`
`[16] Permyakov, S.E., Veprintsev, D.B., Brooks, C.L., Permyakov,
`E.A. and Berliner, L.J. (2000) Proteins Struct. Funct. Genet.,
`accepted for publication.
`[17] Permyakov, E.A., Morozova, L.A. and Kalinichenko, L.A.
`(1988) Biophys. Chem. 32, 37^42.
`[18] Veprintsev, D.B., Permyakov, E.A., Kalinichenko, L.A. and Ber-
`liner, L.J. (1996) Biochem. Mol. Biol. Int. 39, 1255^1265.
`[19] Permyakov, E.A., Morozova, L.A. and Burstein, E.A. (1985)
`Biophys. Chem. 21, 21^31.
`[20] Anderson, P.J., Brooks, C.L. and Berliner, L.J. (1997) Biochem-
`istry 36, 11648^11654.
`[21] Ostrovsky, A.V., Kalinichenko, L.P., Emelyanenko, V.I., Klima-
`nov, A.V. and Permyakov, E.A. (1988) Biophys. Chem. 30, 105^
`112.
`[22] Permyakov, E.A., Kalinichenko, L.P., Morozova, L.A., Yarmo-
`lenko, V.V. and Burstein, E.A. (1981) Biochem. Biophys. Res.
`Commun. 102, 1^7.
`[23] Berliner, L.J., Ellis, P.D. and Murakami, K. (1983) Biochemistry
`22, 5061^5063.
`[24] Kuhlman, B., Boice, J.A., Wu, W.J., Fairman, R. and Releigh,
`D.P. (1997) Biochemistry 36, 4607^4615.
`[25] Permyakov, E.A., Ostrovsky, A.V. and Kalinichenko, L.P. (1987)
`Biophys. Chem. 28, 225^233.
`[26] Veprintsev, D.B., Permyakov, S.E., Permyakov, E.A., Rogov,
`V.V., Cawthern, K.M. and Berliner, L.J. (1997) FEBS Lett.
`412, 625^628.
`[27] Griko, Y.V., Freire, E. and Privalov, P.L. (1994) Biochemistry
`33, 1889^1899.
`[28] Hirai, Y., Permyakov, E.A. and Berliner, L.J. (1992) J. Protein
`Chem. 11, 51^57.
`[29] Permyakov, E.A. and Kreimer, D.I. (1986) Gen. Phys. Biophys.
`5, 377^390.
`[30] Dzwolak, W., Kato, M., Shimizu, A. and Taniguchi, Y. (1999)
`Biochim. Biophys. Acta 1433, 45^55.
`[31] Katsumata, K., Okazaki, A., Tsurupa, G.P. and Kuwajima, K.
`(1996) J. Mol. Biol. 264, 643^649.
`[32] Makio, T., Arai, M. and Kuwajima, K. (1999) J. Mol. Biol. 293,
`125^137.
`[33] Forge, V., Wijesinha, R.T., Balbach, J., Brew, K., Robinson,
`C.V., Red¢eld, C. and Dobson, C.M. (1999) J. Mol. Biol. 288,
`673^688.
`[34] Veprintsev, D.B., Narayan, M., Permyakov, E.A., Uversky,
`V.N., Brooks, C.L., Cherskaya, A.M., Permyakov, E.A. and
`Berliner, L.J. (1999) Proteins Struct. Funct. Genet. 37, 65^72.
`
`[35] Kataoka, M., Kuwajima, K., Tokunaga, F. and Goto, Y. (1997)
`Protein Sci. 6, 422^430.
`[36] Kharakoz, D.P. and Bychkova, V.E. (1997) Biochemistry 36,
`1882^1890.
`[37] Wu, L.C. and Kim, P. (1998) J. Mol. Biol. 280, 175^182.
`[38] Song, J., Bai, P., Luo, L. and Peng, Z. (1998) J. Mol. Biol. 280,
`167^174.
`[39] Uchiyama, H., Perez-Prat, E.M., Watanabe, K., Kumagai, I. and
`Kuwajima, K. (1995) Protein Eng. 8, 1153^1161.
`[40] Schulman, B.A., Red¢eld, C., Peng, Z., Dobson, C.M. and Kim,
`P.S. (1995) J. Mol. Biol. 253, 651^657.
`[41] Red¢eld, C., Schulman, B.A., Milhollen, M.A., Kim, P.S. and
`Dobson, C.M. (1999) Nat. Struct. Biol. 6, 948^952.
`[42] Permyakov, E.A., Grishchenko, V.M., Kalinichenko,