Photoswitchable phospholipid FRET acceptor: Detergent free intermembrane transfer assay of fluorescent lipid analogs

We have developed and characterized a novel photoswitchable phospholipid analog termed N-nitroBIPS-DPPG. The fluorescence can be switched on and off repeatedly with minimal photobleaching by UV or visible light exposure, respectively. The rather large photochromic head group is inserted deeply into the interfacial membrane region conferring a conical overall lipid shape, preference for a positive curvature and only minimal intermembrane transfer. Utilizing the switchable NBD fluorescence quenching ability of N-nitroBIPS-DPPG, a detergent free intermembrane transfer assay system for NBD modified lipids was demonstrated and validated. As NBD quenching can be turned off, total NBD associated sample fluorescence can be determined without the need of detergents. This not only reduces detergent associated systematic errors, but also simplifies assay handling and allows assay extension to detergent insoluble lipid species.

at the depth of the FA ester carbonyl groups. At the same relative membrane depth, the fractional presence of water is ~0.2, indicative of an 80% reduction of H-bonding donor presence in this environment. Interestingly, the H-bond enthalpy in pure ethanol is also ~80% reduced compared to pure water 19 . Together, this suggests that the negatively charged MC moiety is well inserted into the interfacial region of the bilayer, slightly above the carbonyl moieties of the FA esters. It has been previously reported 17,20 that free MC preferentially associates with bulk water in the presence of zwitterionic 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) liposomes, but is able to interact with anionic 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-(1′-rac-glycerol) (POPG) liposomes. Under the experimental conditions employed herein, the covalently bound MC moiety of N-nitroBIPS-DPPG is rather deeply inserted into the interfacial region of DOPC bilayers. This likely resulted in a relatively large head group size compared to the lipid tail, conferring a more conical overall shape to N-nitroBIPS-DPPG.
Photoswitching characteristics of N-nitroBIPS-DPPG were probed by irradiating DOPC: N-nitroBIPS-DPPG liposomes with UV (340 nm) and green light (543), concomitantly, while monitoring fluorescence emission at 600 nm (Fig. 4). The increase in fluorescence during UV exposure is associated with an increased population of the fluorescence active MC species of N-nitroBIPS-DPPG. Exposure to visible light (543 nm) only, resulted in a drastic drop of fluorescence emission as N-nitroBIPS-DPPG reverted to SP species. Repeated photoswitching cycles did not significantly alter the initially observed switching characteristics. Additionally, photoswitching was also not associated with significant photobleaching, consistent with previous reports on other spiropyran compounds 21,22 .
The membrane distribution of N-nitroBIPS-DPPG in DOPC liposomes was studied utilizing the membrane-impermeable reducing agent, sodium dithionite. Sodium dithionite has been previously employed to irreversibly reduce nitro groups to amino groups in the outer leaflet of the phospholipid bilayer, effectively inactivating NBD fluorescence 23 . We adopted this method for N-nitroBIPS-DPPG quenching, incorporating N-NBD-DOPE as control into the liposomes. The time course of NBD and N-nitroBIPS fluorescence quenching by sodium dithionite was recorded in independent experiments (Fig. 5). In both cases, the experiment was terminated by addition of Triton X-100 detergent to allow determination of the quenching end point. Nevertheless, addition of detergents reduces solution turbidity by about 10% (see below), complicating quantification of the total range of quenched fluorescence.  During the initial phase, dithionite rapidly quenched NBD as well as N-nitroBIPS fluorescence, followed by a slow but steady quenching. The slow fluorescence decrease during extended exposure to the reducing agent is more likely associated with dithionite leakage to the inner leaflet, than to spontaneous flip-flop of the fluorescence emitting probe 24 . NBD fluorescence was reduced to 63% during the initial phase. Taking the decrease in turbidity into account during the final quenching, about 70% of N-NBD-DOPE is estimated to be present in the outer leaflet of the liposomes. The diameter of SUVs prepared by ethanol injection has been reported to be ~30 nm 25 . Assuming a bilayer thickness of ~5 nm, the ratio of the surface area between the outer and inner leaflet is 2.25:1. Consequently, ~69% of phospholipids should theoretically be present in the outer SUV leaflet, which is consistent with the corrected experimental results. In contrast to N-NBD-DOPE, N-nitroBIPS-DPPG fluorescence was reduced by ~80% during the initial quenching phase. Considering the error induced by Triton X-100 reduction of turbidity, nearly 90% of N-nitroBIPS-DPPG fluorescence was quenched during the initial phase. Increased leakage of dithionite due to the presence of N-nitroBIPS-DPPG can be excluded as N-nitroBIPS-DPPG and N-NBD-DOPE were incorporated simultaneously in the same liposomes. Similarly, liposome size and thus ratio between inner and outer leaflets were identical between the N-nitroBIPS-DPPG and N-NBD-DOPE experiments as liposomes from the same preparation were used. Also, the slow second phase quenching rate of N-nitroBIPS-DPPG is comparable to the rate exhibited by N-NBD-DOPE, suggesting that the spontaneous  flip-flop speed of the remaining 10% of N-nitroBIPS-DPPG in DOPC membranes might be similarly slow as for N-NBD-DOPE. On the one hand, the remaining 10% of fluorescence intensity may be caused by a small population of oligo-or multilamellar liposomes. On the other hand, the conical shape of N-nitroBIPS-DPPG and the small radius of the employed SUV preparation could be a driving factor for lipid asymmetry between the inner and outer leaflet. Indeed, computer simulations 26 suggest that in liposomes with a radius below 60 nm conical shaped lipids favoring positive curvature (large head group with small tail section) tend to be enriched in the outer leaflet, if present at low concentration in a curvature neutral matrix. It was proposed that the degree of enrichment is directly proportional to the ratio of head to tail size in conical lipids. Consequently, quenching of ~90% of N-nitroBIPS-DPPG fluorescence would suggest that the head group (including hydration shell) is about 1.5 times larger compared to its lipid tail section. Taken together, the high degree of N-nitroBIPS-DPPG quenching due to shape driven lipid asymmetry seems attractive at this point, while rapid flip-flop of N-nitroBIPS-DPPG during the initial phase cannot be ruled out.
The close proximity of the NBD fluorescence emission maxima (λem = 535 nm) and the N-nitroBIPS-DPPG MC species absorption maxima (λex = 543 nm) suggests that FRET can occur. Indeed, NBD associated fluorescence emission is quenched in DOPC/N-nitroBIPS-DPPG/N-NBD-DOPE (89:10:1) liposomes concomitantly irradiated with UV and blue green light (480 nm, for excitation of NBD), eventually reaching a plateau ( Fig. 6, red line). The remaining low level of NBD fluorescence could be associated with the turbidity of the liposome solution or a distribution differences of the FRET partners between the leaflets. Subsequent switching of N-nitroBIPS-DPPG to SP form by visual light excitation abolished NBD quenching ( Fig. 6 blue line). This allows repeated on/off cycling of NBD fluorescence quenching without significant photobleaching. Finally, Triton X-100 was added to the solution (Fig. 6, red arrow) while NBD quenching was turned off. Apparently, NBD fluorescence did not increase, confirming complete deactivation of NBD quenching by visual light exposure of N-nitroBIPS-DPPG. The reduced NBD fluorescence of about 10-15% is most likely caused by turbidity alteration due to detergent addition, exposing the pitfalls of detergent associated determination of maximal fluorescence.
After characterization of the photochemical properties of N-nitroBIPS-DPPG, we established our envisaged detergent free lipid transfer assay. In short, the NBD tagged target lipid is incorporated at an NBD-lipid to N-nitroBIPS-DPPG ratio of 1:10 into DOPC (89%) liposomes. Acceptor liposomes composed entirely of DOPC were added to the reaction mixture upon reaching maximum NBD quenching by the UV light induced MC species of N-nitroBIPS-DPPG (Fig. 7, green arrow). The increase in NBD fluorescence associated with spontaneous transfer of NBD-lipids to the quencher free liposomes was monitored while maintaining UV light exposure to ensure MC species persistence. To determine total NBD fluorescence, UV light irradiation was switched to visible light irradiation (Fig. 7, blue line), converting N-nitroBIPS-DPPG to SP form and abolishing NBD quenching. Only for validation purposes to proof complete abolishment of quenching and to demonstrate the pitfalls of detergent addition Triton X-100 was added at the end of each experiment (Fig. 7, red arrow).
In total, three NBD-labeled phospholipids were selected due to their previously reported specific characteristics 7,[27][28][29][30] . N-NBD-DOPE was utilized as a negative control due to its known lack of spontaneous transfer between liposomes 27-29 . Short chain FA labeled C 6 -NBD-PC acted as positive control as the combination of short FA and hydrophilic NBD label provides sufficient hydrophilicity to facilitate rapid intermembrane transfer 7, 29, 30 , while the medium chain NBD conjugated FA featuring C 12 -NBD-PC has been reported to exhibit slow intermembrane transfer 7,29,30 . Only a marginal fluorescence increase of ~0.1%/min was observed in the presence of N-NBD-DOPE during UV irradiation (Fig. 7C). Although this value is nearly 10 times higher than previously reported (0.01%/min) 29 , it is sufficiently slow to allow measurement of the intermembrane transfer rate of other NBD lipids. Additionally, it suggests that N-nitroBIPS-DPPG does not exhibits a significant intermembrane transfer speed. The initial transfer rate of C 12 -NBD-PC (Fig. 7B) and C 6 -NBD-PC (Fig. 7A) were determined as 0.77%/min and 80.6%/min, respectively. These values are in good agreement with the 0.6%/min for C 12 -NBD-PC and 94.4%/min for C 6 -NBD-PC reported previously 29 .
Interestingly, in case of C 6 -NBD-PC, after rapid transfer of ~48% of the total NBD fluorescence, a drastic reduction of the fluorescence increase rate to 0.27%/min can be observed. This change could be associated with the establishment of a C 6 -NBD-PC equilibrium between the outer leaflets of the donor and acceptor liposomes, rendering the slow flip-flop of C 6 -NBD-PC from the inner to the outer leaflet of the donor liposome as the rate limiting step. Equilibrium of intermembrane transfer with about 48% of total NBD fluorescence is achieved after 3.8 min. At this point, one can assume that all C 6 -NBD-PC formerly present in the outer leaflet of the donor liposomes is equally spread over the outer leaflet of both donor and acceptor liposomes. Additionally, a maximum of 1% of C 6 -NBD-PCs has undergone spontaneous flip-flop from the inner to the outer leaflet of the donor liposomes. Furthermore, the donor and acceptor liposome preparations are likely to exhibit nearly identical size distributions as both are DOPC-based SUVs. As the double amount of acceptor liposomes compared to donor liposomes were added, the recorded C 6 -NBD-PC signal of 48% corresponds to 2/3 of the C 6 -NBD-PC population present in the outer leaflet of the donor liposomes at the start of the experiment. This suggests that 72% of C 6 -NBD-PC was initially incorporated into the outer leaflet of the donor liposomes. This is in good agreement with the above mentioned presence of N-NBD-DOPE and theoretical distribution calculations, indicating that neither N-NBD-DOPE nor C 6 -NBD-PC experience significant asymmetric enrichment.

Conclusion
Taken together, N-nitroBIPS-DPPG is a novel UV-light dependent photoswitchable lipid analogue that can readily be incorporated into liposomes. The MC moiety is rather deep inserted into the interfacial region, close to the FA carbonyl oxygens, resulting in a rather large head group size compared to its saturated FA tail. Its suitability as a switchable FRET acceptor and quencher of NBD fluorescence has been demonstrated. Utilizing N-nitroBIPS-DPPG switchable quenching, a simplified and improved intermembrane lipid transfer assay has been validated. Importantly, as quenching can be switched off, the need to add detergent, such as Triton X-100 at the end of the experiment to determine maximal fluorescence can be omitted. This renders corrections for turbidity changes caused by the detergent obsolete and significantly reduces systematic errors.

Measurement of N-nitroBIPS-DPPG fluorescence.
The fluorescence characteristics (λex = 543 nm, λem = 600 nm) of a 50 μM (total lipid) DOPC/N-nitroBIPS-DPPG (9:1) liposome solution in PBS was monitored on a spectrofluorometer (FP-6500, JASCO) under constant mixing at room temperature. First, UV (340 nm) and green light (543 nm) were alternately irradiated every 5 sec. During green light irradiation, fluorescence emission at λem = 600 nm was monitored. After maximal fluorescence was attained, the irradiation scheme was switched to continuous irradiation with green light and simultaneous fluorescence detection at λem = 600 nm. Upon return to minimal fluorescence, the irradiation scheme was switched back to the initial scheme and the above detailed cycle was repeated.

Förster resonance energy transfer (FRET) between N-NBD-DOPE and N-nitroBIPS-DPPG.
50 μM (total lipids) DOPC/N-nitroBIPS-DPPG/N-NBD-DOPE/(89:10:1) liposomes were alternately irradiated with blue green (480 nm) and green (543 nm) light every 5 sec during constant mixing at room temperature. During the whole experiment the NBD fluorescence signals were monitored at λem = 535 nm during blue green light irradiation. Upon attaining saturation of the fluorescence signal, the irradiation scheme was switched and the sample was alternately irradiated with UV (340 nm) and blue green light every 5 sec. After the fluorescence signal stabilized at minimal levels, the irradiation scheme was switched back to the initial scheme and above outlined irradiation cycle was repeated. As indicated, 20 μL of 10% Triton-X 100 (final concentration 0.1%) solution was added to the liposomes mixture during alternating blue green and green light irradiation.
Intermembrane transport of NBD lipids. 50 μM (total lipids) DOPC/N-nitroBIPS-DPPG/NBD lipid (N-NBD-DOPE or C 6 -NBD-PC or C 12 -NBD-PC) (89:10:1) liposomes were alternately irradiated with UV (340 nm) and blue green light (480 nm) every 5 sec during constant mixing at room temperature. During the whole experiment fluorescence emission at λem = 535 nm was recorded during blue green light irradiation. After the fluorescence signal stabilized at minimal levels, DOPC liposomes (100 μM total lipids) were added. As indicated, the sample was alternately irradiated with blue green (480 nm) and green (543 nm) light every 5 sec while maintaining constant mixing at room temperature. As marked, 20 μL of a 10% Triton-X 100 (final concentration 0.1%) solution were added to the liposomes.