Folding of soluble proteins in vitro has been studied extensively in the past, and much insight into the details of folding has been gained. Folding proceeds along a pathway with a number of more or less stable intermediates (for a review, see (1) ). By contrast, the study of folding of membrane proteins is still in its infancy. Only a few integral membrane proteins have been unfolded and refolded in vitro: the -helical proteins bacteriorhodopsin (2) and the photosynthetic antenna complex(3) , as well as the -barrel proteins OmpA (4) and the porin OmpF(5) . The kinetics of refolding were studied only for bacteriorhodopsin (6) and OmpF(5) . In all these cases, refolding took place into mixed micelles of detergent and lipid, which are considered as a good model system for a lipid membrane. This is certainly true as far as protein structure is concerned, but does not necessarily hold for the kinetics of folding which may be much slower for insertion into a membrane than into a micelle. This problem may be solved by studying refolding into lipid membranes(7) .

The main problem with membrane proteins lies in their insolubility in water which makes it difficult to keep them solubilized and unfolded before starting the refolding process. If water is replaced by another solvent such as trifluoroacetic acid, they may become soluble in the unfolded form as shown for bacteriorhodopsin(2) , but such a model system bears less resemblance to the in vivo situation given by a membrane in water. In another approach, the proteins bacteriorhodopsin and OmpF were made soluble by attaching polyethylene glycol to them(8) , but again this represents a drastic change compared to the in vivo situation. An exceptionally favorable case is provided by members of the -barrel class of membrane proteins which can be kept soluble in water in the unfolded form(4, 5) . This is a consequence of their low hydrophobicity which itself is a characteristic feature of their -barrel structure and distinguishes them from the -helical membrane proteins(9) . We have shown previously that OmpA, a protein of the outer membrane of Escherichia coli, refolds from its unfolded state and inserts directly into lipid vesicle membranes(7) .

In the present paper, the kinetics of this process are investigated. This is the first kinetic study of refolding and direct membrane insertion of an integral membrane protein. As experimental techniques, we used CD spectroscopy, Trp fluorescence, ANS ()fluorescence, and protease digestion with ensuing SDS-gel electrophoresis.

MATERIALS AND METHODS

Purification and Unfolding

Refolding and Membrane Insertion

Folding and insertion of OmpA into vesicle membranes was started by adding a small volume of the stock solution of unfolded OmpA to preformed vesicles at 30 °C (only in the experiment of Fig. 3A, unfolded OmpA was diluted into water prior to addition of vesicles). The experiments were routinely performed with highly sonicated lipid vesicles consisting of DMPC or the mixture DMPC/DMPG at a molar ratio of 95/5. The vesicles were prepared essentially as described previously(7) . After sonication, the lipid concentrations and the pH values of the vesicle dispersions were adjusted to their final values, followed by storage of the vesicles overnight at 30 °C. The size of the vesicles was characterized by quasielastic light scattering (Coulter N4/SD) and electron microscopy (Philips 201), the diameter was 30-40 nm. For one experiment, unsonicated vesicles were prepared by extrusion of lipid dispersions through filters (Nuclepore). The diameter of these vesicles (again after 1 night at 30 °C) was about 200 nm. The end concentrations of the folding and insertion experiments were 2.4 µM OmpA, 1.5 mM (or 0.75 mM) DMPC/DMPG, 100 mM NaCl, 16 mM urea, 20 mM buffer when parallel measurements of CD, Trp fluorescence, and protease digestion were performed, and 0.6 µM OmpA, 1.5 mM (or 4.5 mM) DMPC, 4 mM urea, 20 mM buffer when only Trp fluorescence was measured. Spectra of inserted OmpA were recorded 3 h after the start of insertion. Since the kinetics of folding and insertion depend on the vesicle size, it was important to use the same preparation of vesicles in performing the parallel kinetic measurements.

Figure 3: Yield of folding and membrane insertion of 2.4 µM OmpA. A, after different incubation times of partially folded OmpA in water at pH 7.3 and pH 10 prior to addition of 7.5 mM DMPC; B, at different concentrations of DMPC at pH 7.3 and zero incubation time; and C, at different pH and zero incubation time in 7.5 mM DMPC. The yield of insertion has been determined by quantitating the membrane-protected fragment on SDS/gels after insertion and subsequent trypsin digestion (lane 3 of the inset). Inset, SDS-gel of OmpA unfolded in urea (lane 1), inserted into DMPC vesicles at pH10 without digestion (lane 2), and after digestion by 0.1 µM trypsin (lane 3) (Samples of lane 1 and 2 were not boiled before being loaded on the gel.)

Peptidyl-prolyl-isomerase from pig kidney was a gift of F.-X. Schmid (10) and was added in one experiment at an equimolar ratio of isomerase/OmpA.

Spectroscopy

Trypsin Digestion and SDS-Gel Electrophoresis

RESULTS

Conformational States

Figure 1: Circular dichroism (A) of OmpA (2.4 µM), Trp fluorescence (B) of OmpA (0.6 µM), and ANS fluorescence (C) in the absence and presence of OmpA (2.4 and 4.8 µM) under different conditions. - -, unfolded OmpA in 8 M urea at pH 7.3; - - -, partially folded OmpA in water at pH 7.3; --, partially folded OmpA in water at pH 10.0; -, folded OmpA in 0.75 mM DMPC/DMPG (A) or 1.5 mM DMPC (B) at pH 7.3. In C, the ANS fluorescence before the addition of OmpA is included in water at pH 7.3 (- -) and in the presence of 8 M urea ().

Unfolded OmpA in 8 M urea is characterized by a CD spectrum indicating a completely random structure. Its Trp fluorescence is relatively weak, with a maximum at 350 nm indicating a polar environment. The ANS fluorescence remains unchanged by the addition of OmpA, i.e. OmpA does not bind the hydrophobic probe ANS.

When the urea concentration is decreased to 20 mM by dilution of unfolded OmpA into water, the CD spectrum indicates that OmpA adopts secondary structure, which has been described previously as a mixture of , , and random structure(7) . The Trp fluorescence is still weak, but the emission maximum is already shifted to 343 nm. The hydrophobic probe ANS (12) binds to this partially folded OmpA as indicated by an increase of the intensity of the ANS fluorescence and a shift of the emission maximum from 520 nm to 480 nm. This means that in contrast to unfolded OmpA, partially folded OmpA has hydrophobic regions which are accessible to ANS. Increase of the pH from 7.3 to 10.0 leads to a loss of secondary structure, as indicated by CD, and a reduced Trp fluorescence. The increase of ANS fluorescence at pH 10 is less pronounced than at pH 7.3 (data not shown).

After dilution of urea-unfolded OmpA into a dispersion of small DMPC vesicles, the CD spectra indicate an increased content of -structure. The intensity of the Trp fluorescence is high and the emission maximum is at 325 nm. The CD and fluorescence spectra agree with those of conventionally reconstituted OmpA(7) . ANS binding is not suitable to test accessible hydrophobic regions of membrane-inserted OmpA, because in this case ANS fluorescence is always high due to its partitioning into the lipid membranes.

Kinetics of the Transition from Urea to Water

Figure 2: Kinetics of partial folding of 2.4 µM OmpA into water at pH 7.3 as observed by CD at 220 nm, Trp fluorescence at 330 nm, and ANS fluorescence at 480 nm.

In contrast to ellipticity, the intensity of the Trp fluorescence at 330 nm was slowly increasing within 30 min, but did not reach the value of membrane-inserted OmpA. This rise in Trp fluorescence is indicative of aggregation of OmpA, as will be demonstrated below.

Both processes, i.e. fast folding and slow aggregation in water, can be visualized by ANS fluorescence. Within the mixing time, the ANS fluorescence increased rapidly due to binding of ANS to partially folded OmpA, followed by a slow increase of ANS fluorescence due to aggregation of OmpA. The latter increase parallels the slow increase of the Trp fluorescence.

Yield of Folding and Membrane Insertion

The effect of incubation time on the yield is shown in Fig. 3A. At pH 7.3, the yield decreases from 68% for zero incubation time to 10% for 27-h incubation time. This indicates that aggregation of the partially folded protein in water competed with membrane insertion. By centrifugation, aggregated OmpA could indeed be precipitated. ()At pH 10, refolding and membrane insertion was almost quantitative, and the yield decreased only slightly with increasing incubation time. This implies a reduced tendency to aggregate which presumably is caused by an increased charge of the protein at this pH.

The effect of lipid concentration on the yield is shown in Fig. 3B. The yield increased with lipid concentration and saturated above about 4 mM DMPC, corresponding to a molar lipid/protein ratio of roughly 1000 or a molar vesicle/protein ratio of about 1/10.

In Fig. 3C, the pH dependence of the yield is shown for the range from pH 2.5 to 11.5. At pH 10, the yield was optimal with 98%. At higher pH, the yield decreased, presumably due to the drastically increased charge of the protein which interferes with folding. Decreasing the pH from 10 to 5.7, the theoretical isoelectric point of OmpA, led to a reduced yield, probably caused by a stronger tendency to aggregate. Below the isoelectric point, virtually no membrane insertion was observed.

Kinetics of the Transition from Water to Membranes

Fig. 4A shows the mean residue ellipticity at 206 nm over a time range of 3 h after the start of folding and insertion. At least three phases can be distinguished. The initial value (-7.5 10 degrees cm dmol) is already characteristic for partially folded OmpA in water (Fig. 2); hence, the fast transition from the unfolded state in urea to the partially folded state in water occurred within the mixing time of the experiment (about 1 s). The subsequent process of folding from the partially folded state in water to the completely folded and membrane-inserted state is slow and at least biphasic, the half-times being about 5 min for the moderately slow step and about 40 min for the slow step, with amplitudes of 75% and 25%, respectively.

Figure 4: Kinetics of folding and membrane insertion of 2.4 µM OmpA into vesicles of 0.75 mM (A, B) or 1.5 mM (C) DMPC/DMPG at pH 7.3 as observed by CD at 206 nm (A), Trp fluorescence at 330 nm (B), and protease digestion (C). In B, a logarithmic plot of the Trp fluorescence is shown as an inset. In C, Trp fluorescence at the higher lipid concentration is included for comparison and the SDS-gel is shown as an inset.

The time course of the Trp fluorescence shown in Fig. 4B is also at least biphasic with half-times of 4 min and 35 min and amplitudes of 73% and 27%, respectively. This demonstrates that CD at 206 nm and Trp fluorescence detect the same transitions.

The time course of the yield of insertion as determined by protease digestion is shown in Fig. 4C. Because the lipid concentration in this experiment was twice as high as in the CD and fluorescence experiments (Fig. 4, A and B), the time course of fluorescence at this lipid concentration is included. Both time courses are biphasic with roughly equal half-times of 2 min and 30 min, but different amplitudes. The amplitude of the fast phase is only 25% for the yield of insertion, compared to 80% for the fluorescence intensity. This would indicate that upon contact with membranes, in a first step all OmpA molecules adopt secondary structure and a small fraction inserts into the membrane, while in the second step the major part inserts and all fold into the final structure. However, it is also possible that in the first step actually no molecules insert, but a fraction of them inserts during cooling through the lipid phase transition (which is done to stop insertion).

The slowest step observed can probably not be attributed to proline isomerization of part of the OmpA molecules, because it was not accelerated by a peptidyl-prolyl-isomerase added in equimolar amounts (data not shown). Likewise, any effect of the periplasmic part of OmpA on folding and membrane insertion is negligible, because the purified tryptic fragment exhibited the same folding behavior (data not shown). Finally, the possibility that insertion is slow due to a coupling of protein insertion to vesicle fusion must be rejected, because fusion was shown to proceed still slower (data not shown)(14) .

Variation of Parameters Affecting the Kinetics

The velocity of insertion increased with lipid concentration as shown in Fig. 5A, the half-times of the fluorescence increase being inversely proportional to the lipid concentration. Increasing the pH from 7.3 to 10.0 decreased the velocity of insertion, as shown also in Fig. 5A. At pH 10, the half-times are 5 times larger than at pH 7.3. The effect of membrane curvature is demonstrated in Fig. 5B. Insertion into unsonicated vesicles of low curvature was extremely slow compared to insertion into sonicated lipid vesicles of high curvature. The fastest kinetics were observed for folding into detergent micelles of dodecyl maltoside. Essentially the same behavior was obtained with micelles of octyl glucoside (data not shown). The moderately slow step of insertion is accelerated compared to insertion into vesicle membranes, its half-time decreasing from 5 min to 20 s, but the slow step proceeds with roughly the same kinetics as for insertion into vesicle membranes. The observed increase in fluorescence is not caused simply by binding of detergent molecules to the partially folded state of OmpA, but reflects structural changes of the protein as indicated by CD measurements(7) . The faster folding and insertion of OmpA into micelles results either from the higher curvature of the micelles or from the formation of a new micelle around the folding protein.

Figure 5: Influence of different parameters on the kinetics of folding and membrane insertion of 0.6 µM OmpA as observed by Trp fluorescence. A, insertion into vesicles of DMPC at a lipid concentration of 1.5 or 4.5 mM and pH 7.3 (-- and ) or pH 10.0 (-); B, insertion at pH 10.0 into sonicated or not sonicated vesicles of DMPC at 4.5 mM and into micelles of dodecyl maltoside at 1 mM. As control, the constant fluorescence of partially folded OmpA in water at pH 10.0 is included, indicating that OmpA does not aggregate at this pH.

DISCUSSION

Folding and membrane insertion of OmpA was found to proceed in at least three steps. In water, OmpA undergoes a transition from the unfolded state U into an intermediate state I with a characteristic time below a second. Upon association with membranes, two slower transitions take place with characteristic times of about 5 min and h leading to the folded and inserted state F. These results will now be interpreted within the framework of a kinetic model for folding and membrane insertion of OmpA (Fig. 6).

Figure 6: Scheme of the proposed folding pathway of OmpA. U, unfolded in water/urea; I, misfolded in water; A, aggregated in water; I, partially folded and inserted in membrane; F, fully folded and inserted in membrane.

The process of partial folding in water is reminiscent of the first step in folding of soluble proteins. They often adopt a partially folded state within ms. The state is called molten globule and believed to arise from a hydrophobic collapse in which the hydrophobic core of soluble proteins forms, but not their detailed structure(15) . Postulating such a hydrophobic collapse to take place in the case of OmpA, it should be fast as indeed observed, but the state I would not be a partially correctly folded state, but a completely misfolded state, something like an inside-out version of native OmpA.

Guided again by the analogy to soluble proteins(16) , from state I two pathways seem possible. The protein may associate with membranes or it may aggregate, and both processes are indeed observed. Both are driven by the hydrophobic effect. Aggregation may be suppressed by increasing the repulsion between OmpA molecules (e.g. due to charges and/or loss of structure with exposed hydrophobic patches) as well as by increasing the interaction with membranes (e.g. by increasing the lipid concentration or the curvature of the vesicle membranes). Under optimal conditions, aggregation is negligible, and the yield of folding and insertion is virtually quantitative.

The two transitions which occur upon association with membranes are relatively slow. Which structural changes might they reflect or, in other words, what is the nature of the intermediate I? Folding of soluble proteins is sometimes rate-limited by proline isomerization (1) or by pairing of domains (17) . Both possibilities could be excluded for OmpA, as was the case with fusion of vesicles. We therefore propose a model for folding of OmpA which closely resembles the standard model for folding of soluble proteins. Upon contact with a membrane, OmpA starts to insert and to fold into a state I whose structure is globally correct, but in detail incorrect, i.e. a barrel-like structure is formed, but the details remain wrong. By this structural characterization, I would be the analogue of the molten globule state of soluble proteins (and not the misfolded state I in water). Furthermore, I would resemble the intermediate state in folding of -helical membrane proteins proposed within the framework of a two-state model(18) . A vaguely related molten-globule state has been postulated for the pore-forming domain of colicin(19) . Because the membrane can be considered as a reactant, the rate of the transition from I to I should be proportional to the lipid concentration, as observed for an amphipathic helix(20) . The final transition from I to F would require fine-tuning of the protein structure and should be the slowest step.

Comparing the predictions of the model with the experimental data, the rate of the moderately slow transition from I to I was indeed proportional to the lipid concentration and was different for insertion into vesicle membranes and into micelles. As expected, insertion into micelles was faster. By contrast, the slow transition from I to F proceeded with the same kinetics for membranes and micelles. This again would be expected, because fine-tuning of the structure should be roughly independent of the environment. To which extent OmpA is inserted in the membrane in state I may be estimated from the data on protease digestion. Unfortunately, these data are not completely conclusive, but indicate that at least the major part of OmpA molecules became protected with the half-time of the slow transition from I to F. Hence, in I, the major part if not all of the OmpA molecules are associated with the membrane but not inserted. They insert upon folding into the final state F.

Concerning folding in vivo, OmpA faces two kinetic problems after being translocated across the inner bacterial membrane. It is prone to aggregation(13) , and its interaction with the outer bacterial membrane, which has negligible curvature, is extremely slow. Periplasmic chaperones and catalysts for sorting to and insertion into the outer membrane may help to overcome these problems. Our approach should be useful to elucidate in vitro the effect of periplasmic chaperones and catalysts on folding and sorting of OmpA.

FOOTNOTES

*

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore by hereby marked ``advertisement'' in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

§

Present address: Dept. of Physics and Molecular Biology, Princeton University, Princeton, NJ 08544.

¶

To whom correspondence should be addressed. Tel.: 49-7071-601-238; Fax: 49-7071-629-71.

()

The abbreviations used are: ANS, 1-anilino-8-naphthalenesulfonate; DMPC, dimyristoylphosphatidylcholine; DMPG, dimyristoylphosphatidylglycerol.

()

T. Surrey and F. Jähnig, submitted for publication.

ACKNOWLEDGEMENTS

REFERENCES

1. Schmid, F.-X. (1992) in Protein Folding (Creighton, T. E., ed) pp. 197-241, W. H. Freeman and Co., New York and other articles therein
2. Huang, K.-S., Bayley, H., Liao, M.-J., London, E. & Khorana, H. G. (1981) J. Biol. Chem. 256, 3802-3809 [Abstract/Free Full Text]
3. Paulsen, H., Rümler, U. & Rüdiger, W. (1990) Planta 181, 204-211 [CrossRef]
4. Dornmair, K., Kiefer, H. & Jähnig, F. (1990) J. Biol. Chem. 265, 18907-18911 [Abstract/Free Full Text]
5. Eisele, J.-L. & Rosenbusch, J. P. (1990) J. Biol. Chem. 265, 10217-10220 [Abstract/Free Full Text]
6. Booth, P. J., Flitsch, S. L., Stern, L. J., Greenhalgh, D. A., Kim, P. S. & Khorana, H. G. (1995) Nature Struct. Biol. 2, 139-143 [CrossRef][Medline] [Order article via Infotrieve]
7. Surrey, T. & Jähnig, F. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 7457-7461 [Abstract/Free Full Text]
8. Wei, J. & Fasman, G. D. (1995) Biochemistry 34, 6408-6415 [CrossRef][Medline] [Order article via Infotrieve]
9. Vogel, H. & Jähnig, F. (1986) J. Mol. Biol. 190, 191-199 [CrossRef][Medline] [Order article via Infotrieve]
10. Fischer, G., Wittmann-Liebold, B., Lang, K., Kiefhaber, T. & Schmid, F.-X. (1989) Nature 337, 476-478 [CrossRef][Medline] [Order article via Infotrieve]
11. Schweizer, M., Hindenach, I., Garten, W. & Henning, U. (1978) Eur. J. Biochem. 82, 211-217 [Medline] [Order article via Infotrieve]
12. Semisotnov, G. V., Rodionova, N. A., Razgulyaev, O. I., Uversky, V. N., Gripas, A. F. & Gilmanshin, R. I. (1991) Biopolymers 31, 119-128 [CrossRef][Medline] [Order article via Infotrieve]
13. Freudl, R., Schwarz, H., Stierhof, Y.-D., Gamon, K., Hindennach, I. & Henning, U. (1986) J. Biol. Chem. 261, 11355-11361 [Abstract/Free Full Text]
14. Scotto, A. W. & Zakim, D. (1985) Biochemistry 24, 4066-4075 [CrossRef][Medline] [Order article via Infotrieve]
15. Christensen, H. & Pain, R. H. (1991) Eur. Biophys. J. 19, 221-229 [Medline] [Order article via Infotrieve]
16. Goldberg, M. E., Rudolph, R. & Jaenicke, R. (1991) Biochemistry 30, 2790-2797 [CrossRef][Medline] [Order article via Infotrieve]
17. Garel, J. R. (1992) in Protein Folding (Creighton, T. E., ed) pp. 405-454, W. H. Freeman and Co., New York
18. Engelman, D. M. & Steitz, T. A. (1984) in The Protein Folding Problem (Wetlaufer, D. B., ed) pp. 87-113, Westview Press, Boulder, CO
19. van der Goot, F. G., Gonzalez-Manas, J. M., Lakey, J. H. & Pattus, F. (1991) Nature 354, 408-410 [CrossRef][Medline] [Order article via Infotrieve]
20. Schwarz, G., Gerke, H., Rizzo, V. & Stankowski, S. (1987) Biophys. J. 52, 685-692 [Abstract/Free Full Text]

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				D. C. Bay, J. D. O'Neil, and D. A. Court Two-Step Folding of Recombinant Mitochondrial Porin in Detergent Biophys. J., January 15, 2008; 94(2): 457 - 468. [Abstract] [Full Text] [PDF]

				D. Marsh, B. Shanmugavadivu, and J. H. Kleinschmidt Membrane Elastic Fluctuations and the Insertion and Tilt of {beta}-Barrel Proteins Biophys. J., July 1, 2006; 91(1): 227 - 232. [Abstract] [Full Text] [PDF]

				J. E. Mogensen, J. H. Kleinschmidt, M. A. Schmidt, and D. E. Otzen Misfolding of a bacterial autotransporter Protein Sci., November 1, 2005; 14(11): 2814 - 2827. [Abstract] [Full Text] [PDF]

				C. Danelon, T. Brando, and M. Winterhalter Probing the Orientation of Reconstituted Maltoporin Channels at the Single-protein Level J. Biol. Chem., September 12, 2003; 278(37): 35542 - 35551. [Abstract] [Full Text] [PDF]

				P. V. Bulieris, S. Behrens, O. Holst, and J. H. Kleinschmidt Folding and Insertion of the Outer Membrane Protein OmpA Is Assisted by the Chaperone Skp and by Lipopolysaccharide J. Biol. Chem., March 7, 2003; 278(11): 9092 - 9099. [Abstract] [Full Text] [PDF]

				M. F. M. Engel, C. P. M. van Mierlo, and A. J. W. G. Visser Kinetic and Structural Characterization of Adsorption-induced Unfolding of Bovine alpha -Lactalbumin J. Biol. Chem., March 22, 2002; 277(13): 10922 - 10930. [Abstract] [Full Text] [PDF]

				M. Bogdanov and W. Dowhan Lipid-assisted Protein Folding J. Biol. Chem., December 24, 1999; 274(52): 36827 - 36830. [Full Text] [PDF]

				R. Koebnik Structural and Functional Roles of the Surface-Exposed Loops of the beta -Barrel Membrane Protein OmpA from Escherichia coli J. Bacteriol., June 15, 1999; 181(12): 3688 - 3694. [Abstract] [Full Text]

				H. de Cock, K. Brandenburg, A. Wiese, O. Holst, and U. Seydel Non-lamellar Structure and Negative Charges of Lipopolysaccharides Required for Efficient Folding of Outer Membrane Protein PhoE of Escherichia coli J. Biol. Chem., February 19, 1999; 274(8): 5114 - 5119. [Abstract] [Full Text] [PDF]

				I. Shin, D. Kreimer, I. Silman, and L. Weiner Membrane-promoted unfolding of acetylcholinesterase: A possible mechanism for insertion into the lipid bilayer PNAS, April 1, 1997; 94(7): 2848 - 2852. [Abstract] [Full Text] [PDF]

				L. K. Tamm, A. Arora, and J. H. Kleinschmidt Structure and Assembly of beta -Barrel Membrane Proteins J. Biol. Chem., August 24, 2001; 276(35): 32399 - 32402. [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Surrey, T. Articles by Jähnig, F. Search for Related Content PubMed PubMed Citation Articles by Surrey, T. Articles by Jähnig, F. Social Bookmarking                      What's this?