| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH |
|
|
|
|||||||
|
|||||||||
Electronic Letters to:
|
|
Electronic letters published:
![[Read eLetter]](/icons/misc/sectionDOWN.gif){kind=icon}
Tear Lipocalin vs Phospholipid Transfer Protein as the Major Lipid Carrier in Tears
Author Response: Tear Lipocalin vs Phospholipid Transfer Protein as the Major Lipid Carrier in Tears
|
|
|||
|
Ben J. Glasgow
Send letter to journal:
bglasgow{at}mednet.ucla.edu Ben J. Glasgow
|
We appreciate the article by Saaren-Seppala et al.1 the subject of which is interaction of tear lipocalin (TL) with liposomes and lipid monolayers composed of variably charged lipids. The authors conclude that while TL interacts with lipid in tear film interfaces it does not possess a role in phospholipid transfer from cell membranes and that "the notion of the role of TLc as the major lipid-transferring protein in human tears should be revised." Further they propose that phospholipid transfer protein (PLTP) performs the function of tear lipid trafficking in the tear film. Because the authors cite our laboratory's pre-submission discussions and critical comments as well as our publications in their paper, we feel it is important to publicly state our viewpoint on these issues. It is important for the readership to know that neither our laboratory nor others in this field have ever proposed that TL transfers lipids from cell membranes. We have provided a body of information that TL is the major lipid binding protein in human tears,2 identified the endogenous lipids which include phospholipids,2 characterized the scavenging of lipid from a variety of surfaces and the surface of the normal cornea,3,4 demonstrated that the acyl chains of bound lipids are immersed in the hydrophobic cavity,5,6 and studied structural elements and mechanistic features that lead to its promiscuous binding.7-11 Nowhere in our previous work is any notion or implication that TL transfers lipids from those that compose the membrane of cells. The authors also state that TL is unlikely to have the capacity to carry phospholipid in the tear film. They advance the idea that PLTP is the principal phospholipid binder because its binding affinity and capacity exceed those for TL. However, the authors must come to terms with important data. First, phospholipid transfer protein is present in small amounts in tears, 0.137 mM, and has not yet been found to carry any endogenous lipid in tears.1 Second, the measured Kd of TL and PLTP for phospholipid binding are about the same, 151nM and 131nM, respectively.4,12 Third, in the article in the same IOVS issue as the article by Saaren-Seppala et al., we performed gel filtration of tears after application to the corneal surface on which fluorescent lipid had been added.4 We expect that these lipids were distributed on the external glycocalyx of epithelial cell or in various environments on the surface of the cell but not within the cell membrane. The corneal cell membranes were not disrupted. The fluorescent phospholipid was associated with the one fraction and only one gel filtration fraction, that of TL. The data clearly demonstrate that native TL, despite the presence of endogenous ligands, is not saturated in tears and functions to pick up lipids that lie on the surface of the cornea rather than from the lipids that compose the cell membrane. No fluorescence was detected with fractions eluted for other proteins and particularly not at 80kDa, the molecular weight of the glycosylated PLTP in tears or at 160kDa, the apparent molecular weight on a size exclusion column.13 Even in tears without tear lipocalin no detectable fluorescent peak was associated with the elution of PLTP.4 PLTP does not appear to contribute significantly to a scavenging function in tears. The PLTP concentration in tears is about 11 mg/mL or 0.137mM or about 1/540 the concentration of TL in tears (~74mM). Even with a theoretical binding capacity of 43 phospholipid molecules/ molecule of PLTP12 the total binding capacity is still only 1/12 of lipocalin. Also, it does not appear that PLTP rapidly transfers the phospholipid to the tear surface or to another tear protein because no other peak of fluorescence is seen in gel filtration.4 Furthermore, there does not appear to be any "crosstalk" between PLTP and TL because isolated purified tear lipocalin scavenges as well as whole tears.4 Therefore, both theoretical and experimental data refute the authors' assertion that PLTP somehow supplants TL as a major carrier of phospholipid. The authors present experimental data that PLTP, not TL, transfers lipid from liposomes to HDL3, a protein-specific interaction.1,13 The authors recognize the lack of any functional significance in tears because HDL3 is not found in tears; one cannot extrapolate from these data to a phospholipid transfer function in the tear film. Our presubmission critical comments exposed some serious experimental problems with this paper that the authors may wish to address. Fluorescence anisotropy can be used to assess rotational diffusion if the timescale of rotational diffusion of the molecule or side chain of the chromophore in question is comparable to the lifetime of the fluorophore. The overall rotational correlation time of TL is on the order of 13 ns at 20 °C while the emission lifetime of tryptophan 17 is about (1.18 ns).5,11 The fluorescence anisotropy of tryptophan 17 reflects the restricted motion of the tryptophan side chain rather than the slow rotational motion of the TL molecule. Hence the authors interpretation of r=0.34 in the absence of liposomes as reflecting slow Brownian rotational diffusion of TL in solution is erroneous. The published value is closer to 0.18 at the excitation wavelength of 295 nm.11 Furthermore, the correlation time of TL bound to the much larger liposome (up to 100 nm in size) is much longer so that tryptophan emission would be completed long before rotational diffusion could effect depolarization. Therefore, all experiments dealing with the interaction of liposomes and TL are fundamentally flawed and the correlation time of the liposome-protein complex would need to be measured and a fluorophore chosen with a comparable fluorescence lifetime. It is inappropriate to use Trp 17 as the fluorophore for these experiments unless the authors expect increased side chain mobility upon liposome binding, which would yield lower r values. Given this fundamental flaw, the best explanation for the anisotropy data presented by the authors is scattered light leading to artifactual values for tryptophan anisotropy. The authors' data corroborate the presence of scattered light because the anisotropy values in Figures 3 and 7 were often greater than 0.4. If the measured anisotropy for a randomly oriented sample is greater than 0.4, one can confidently infer the presence of scattered light.14 Furthermore the r value of 0.34 obtained for TL alone is far greater than the published values of 0.18 at an excitation wavelength of 295 nm and 0.25 at 305 nm, respectively.11 Although it is not clear what excitation wavelength was used in these experiments, a presumed excitation wavelength of 295 nm and a band path of 10 nm would likely yield an even lower r value than 0.18. In addition the authors used a rotating emission polarizer in combination with a monochromator to measure anisotropy. It is important that the authors account for differences in sensitivity between vertically and horizontally polarized components via the instrument's G factor. This factor was not given in the authors' methods and may be a source of additional error in the data. Unfortunately, the combination of these problems leaves the anisotropy data completely uninterpretable as well as the quenching experiments that were combined with or utilized the anisotropy data. The interaction studies of variably charged lipids and lipocalin at the air-fluid interface (Figs. 1, 2) seem to contradict the KI and pyrene quenching experiments. Clearly, TL has an effect at concentrations of phospholipid that yield surface pressure below 30 mN/m. But once the surface pressure of 35mN/m is reached (equivalent to the threshold surface pressure of a stable monolayer) lipocalin no longer affects the surface pressure (does not insert itself into the lipid monolayer). The authors conclude that this shows an equal interaction of the TL with variously charged lipids, contradicting their own assertion that there are separate TL domains interacting with oppositely charged lipids. A hint to the resolution of this apparent contradiction is perhaps evident from their KI quenching experiments (Fig. 6). As a preface, the use of KI to measure accessibility by quenching is not appropriate because the negatively charged ion is known not to readily penetrate the hydrophobic cavity in which Trp 17 resides.7,14 A neutral quencher, acrylamide, would be more appropriate because the collisional quenching constant for Trp 17 in the holoprotein is 1.0M-1,10 compared to the values published by the authors for KI of 0.0002M-1. We calculated a quenching constant of 0.2M-1 from their data, still 1/5 the value of acrylamide quenching. While the authors conclude that electrostatic interaction between the charged liposomes and TL result in varying accessibility to I-, there is another more likely explanation. Unlike neutral or negatively charged phospholipids, the positively charged phospholipids in the liposomes attract negatively charged I-. The diffusion constant of I- and therefore the collisional frequency will be greater in the positively charged environment. The greater quenching constant does not necessarily reflect the increased exposure of the relatively buried tryptophan 17 with liposomes as the authors suggest, but perhaps the quenching constant of I- affected by the charged environment of the liposomes. As expected the change noted is very small overall (less than 10%) and lower with negatively charged lipids (repulsion of I-), all readily explained by this mechanism and consistent with their air fluid interface studies. Neutral acrylamide as a quencher would have eliminated the electrostatic effect induced by the charged lysosomes. With regard to the experiments with pyrene-containing liposomes (Figs. 4, 5), the authors conclude from the pyrene quenching experiments that Trp is positioned so the resonance energy transfer can occur immediately in neutral and anionic liposomes but is somehow slower for cationic liposomes, contradicting their time dependent experiments that show no difference between the lipids. The pI given for the major isoform of human TL is erroneous; theoretical and experimental values have shown it to be 5.3-5.4,15 respectively not 4.8 as the authors cite for porcine VEG protein.1 In summary the data that are interpretable in these experiments demonstrate that TL is surface active and does not transfer lipids from liposomes to HDL3. The concept that TL is the principle scavenger of lipid from the corneal surface stands; PLTP does not assume this role in tears. The role of TL at the tear film-air interface has yet to be explored. Likewise, the role of PLTP in the tear film remains to be elucidated, bearing in mind that HDL3 is absent in tears. Ben J. Glasgow Jules Stein Eye Institute, University of California, Los Angeles, California References 1. Saaren-Seppälä H, Jauhiainen M, Tervo TMT, Redl B, Kinnunen PKJ, Holopainen JM. Interaction of purified tear lipocalin with lipid membranes. Invest Ophthalmol Vis Sci. 2005;46:3649-3656. |
|||
|
|
|||
|
Juha Holopainen
Send letter to journal:
juha.holopainen{at}hus.fi Juha Holopainen
|
We wish to thank Drs. Glasgow and Gasymov for a critical reading of our recent paper published in IOVS.1 The main criticism they raise is the interpretation of our spectroscopic results. In this study1 we used fluorescence spectroscopy and monolayer technique to gain insight into the lipid binding properties of human tear lipocalin (Tlc) and also asked the question whether it may transport lipids in a lipid transfer assay. Essentially our results indicate that Tlc readily interacts with lipid membranes, but in the lipid transfer assay (which is based on a transfer of radiolabeled phospholipid between a liposome and lipoprotein, HDL3), no lipid transfer was observed. The criticism raised by Drs. Glasgow and Gasymov is well taken and points to the fact that the methods we have used are indirect and thus prone to possible misinterpretations. We would like, however, to emphasize that at least some of the possible misinterpretations were already discussed in the original publication, especially those concerning the anisotropy data.1 In contrast to the statement by Glasgow and Gasymov we are not proposing that Tlc would transfer lipids from any cellular or more biological membrane. We state in the text1 that Tlc may transfer lipids and other lipophilic substances from the corneal surface, as is actually elegantly described in their very recent paper in IOVS.2 The only notion of lipid uptake from cell membranes is the discussion part concerning the origin of the endogenous lipids carried by Tlc, and these lipids seem to originate from the E. coli membranes where Tlc has been expressed.3 In contrast to the statement by Glasgow and Gasymov we have not argued about the capacity of Tlc to carry lipids. We believe, however, that the function of Tlc in the tear film is not related to its capacity to carry lipids. Instead, as we have suggested,1 we believe that the main function of Tlc in tear fluid is to maintain and stabilize the oil-water interface in a more or less similar manner as surfactant proteins in the alveolar lavage fluid of the lungs.4 This suggestion has also been previously presented by Nagyova and Tiffany.5 At the time we were preparing our publication1 we did not have any knowledge of the data published by the Glasgow lab very recently.2 We use commonly a lipid transfer assay based on the use of radiolabeled phospholipid transfer between liposomes and specific lipoprotein classes. More specifically, this assay is based on the transfer of radiolabeled phosphatidylcholine from liposomes to HDL3, and further assaying the radioactivity from the lipoproteins. This assay is largely utilized and thus suitable to study lipid transfer. We are currently analyzing the transfer mechanism of phospholipid transfer protein (PLTP) utilizing in part this system. As we mentioned in our previous papers1,6 we realize that lipoproteins are not present in the tear fluid, but we believe that this assay can be used to measure lipid transfer capacity in many other systems, as well. Concerning the role of Tlc in tear fluid, it might well be that Tlc actually does not need an acceptor such as lipoproteins, if it remains attached to the external lipid layer and stabilizes this. On the other hand, if Tlc does scavenge lipid or lipophilic substances from the corneal surface it does not necessarily need to have lipid transfer activity, i.e., it does not have to release its bound lipids as these could be carried through the naso-lacrimal system from the tear fluid. We agree that the new data presented by Glasgow and colleagues2 is compelling evidence that Tlc can carry lipids from the corneal surface, as would be expected if the role of Tlc would be to scavenge lipid contaminants from the corneal surface. It should be noted, however, that several other common proteins also have this capacity to exchange lipids attached to the protein, and perhaps the most common of these is albumin. We agree with the criticism raised by Glasgow and Gasymov on the problems with fluorescence anisotropy. We were, however, well aware of this as we mention in our publication,1 and accordingly other assays were used in addition to these. We strongly oppose the statement that monolayer studies are in contradiction with some of the fluorescence spectroscopy measurements, as Drs. Glasgow and Gasymov suggest. Monolayer studies carried out in our publication1 do not provide any information of the orientation of Tlc on the surface of the lipid layer. The critical surface pressure of the lipid monolayer, that is, the critical surface density of lipid molecules, determines the penetration efficiency of Tlc. In a simplified way one can say that when the surface pressure is low (and the surface density as well is low) there are small 'holes' on the lipid surface which can be filled with any surface active molecule, given that there is enough space for this molecule to be inserted. Accordingly, in this case Tlc will insert itself into the lipid monolayer despite its orientation and fill the free space (and maximize the gain in free energy). According to our data the interaction of Tlc with lipid membranes is mainly driven by hydrophobic interaction rather than electrostatic interaction. The recent paper by Glasgow and collaborators2 is very interesting. The methods they use are sound, and certainly their finding that they cannot observe any fluorescence at the fractions where PLTP should be puts our assumption of the role of PLTP in human tears in a new light. We now know that PLTP is secreted actively into the tear fluid and it is certainly not passively diffused from the plasma (Setälä et al. unpublished). Likewise, we know some details of the mechanism by which PLTP is able to transfer lipids between liposomes and lipoproteins (Setälä et al. unpublished). We are currently searching for the candidate interacting proteins in human tear fluid to understand the role of this protein in maintenance of structure of this complex multicomponent assembly. Juha M. Holopainen Department of Ophthalmology, University of Helsinki, Finland References 1. Saaren-Seppälä H, Jauhiainen M, Tervo TM, Redl B, Kinnunen PKJ, Holopainen JM. Interaction of purified tear lipocalin with lipid membranes. Invest Ophthalmol Vis Sci. 2005;46:3649-3656. |
|||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH |
eLetters: