Figure 1.
Structures of the five main SBD peptide variants used in this study.
The linker structure between the peptide and the N-terminal TMR fluorescent tag is shown, as well as the C-terminal ending.
Figure 2.
Surface Plasmon Resonance (Biacore) assay of SBD variants' association with immobilized liposomes.
Surfaces with and without sphingolipid and ganglioside consisted of the mixtures listed below the X-axis. (A) Fraction binding on the Y-axis refers to the peak height of the association curve in response units (RU), normalized to the average immobilization level over all flow cells for that experiment, indicated by the numbers (in RU) under the graph. The value for fraction binding was obtained by averaging this normalized peak value for at least three flow cells (see Materials and Methods). (B) Fraction bound on the Y-axis refers to the height of the dissociation curve in RU, normalized to the average over all flow cells for that experiment. The error bars represent the variation in the magnitude of response between flow cells.
Figure 3.
Secondary structure probability distributions for the different SBD variants.
Secondary structure propensity (DSSP) analysis from 200 ns REMD simulations, showing the percentage of the time that each structure occurs. Amino-PEG (AEEAc2)-linked and PEG4-linked forms of the E16 peptide with a carboxy terminus were compared in order to assess possible effects of the linker on structure. The different linkers appeared to have no significant effect on the occurrence of structural features. The last four groups show the distribution of structural features that occur in the four variants, either E16 or K16, and with carboxy (COO-) or amide (NH2) termini, in the absence of linker. The inclusion of E vs. K at position 16 affects the frequency of random coil, but does not appear to otherwise change the relative amounts of structural features. Ramachandran plots showing the distribution of peptide bond angles, and exact frequencies of each structural feature are given in fig. S2.
Figure 4.
Histogram and average diffusion times derived from FCS measurements on SBD variants in neuroblastoma cells.
(A) Histogram distribution of τD values with the confocal volume centered on the plasma membrane, over the indicated times ranges in milliseconds for all five SBD variants listed in the key. All variants show a bimodal distribution of τD values except for weakest binding PEG-K16-NH2. The >30 ms group in (A) is redistributed in (B) over the indicated time ranges. (C) Table of average τD values on the membrane and in solution ± standard deviation. Average values were derived from n>60 measurements for each probe either at the membrane or in solution.
Figure 5.
FCS autocorrelation curves and intensity traces on SBD variants labelling neuroblastoma cells.
Normalized ACFs (left; traces) and 2D,2p models (left; line curves), with lag time of the autocorrelation function G(τ) indicated in seconds on the X-axis. Corresponding intensity traces are given on the cell membrane (A, B, E, F) and in the solution (C, D, G, H) for SBD with either the amine-PEG AEEAc linker, E16, and COO- terminus (A–D), or for SBD with a PEG linker, K16, and an amidated C-terminus (E–H). Deviations in the ACF from the fitting model occur between ∼1–10 sec in the measurement of PEG-K16-NH2 on the membrane (E), and spikes in the corresponding intensity fluctuation trace for the membrane measurement in (F) are evident, but not in the ACFs and intensity traces for AEEAc-E16-COO- (A, B) or for either peptide in solution (C, D, G, H). τD values indicated correspond to the particular graphs shown in the figure and not the average values.
Figure 6.
Frequency of aggregation and non-binding of SBD variants.
Spiked intensity traces and complex ACFs by FCS indicated aggregation of SBDs. A complex ACF is defined as deviating from the fitting model in the time-lag range of 1–10 sec. The last two columns show the fraction of peptide assumed to be bound to the membrane, based on %τD1 (millisecond range) vs. %τD2 (microsecond range). Spikes were counted in each measurement over 20 seconds.
Figure 7.
Correlation of aggregation-induced intensity spikes with complex autocorrelation curves in the different SBD variants.
(A) Correspondence of spikes in intensity traces with incidence of non-fitting of the ACF with the 2D,2p model (curves in fig. 4) between 1–10 sec (“complex” ACFs). The number of ACFs and intensity traces observed in each case is given in brackets. For the AEEAc-E16-COO- peptide, the number of spiked intensity traces and complex ACFs was very low compared to the others (n = 9 and 7, respectively). (B) Venn diagrams illustrating the degree of correspondence between readings showing spikes in intensity traces (green circles), vs. those that do not (blue circles) and readings having complex ACFs (beige circles), for each SBD variant. The actual relative proportions of readings in each category, as well as the proportion of readings that were shared between categories, are represented by the relative sizes of the circles and overlaps.
Figure 8.
Labelling of neuroblastoma SH-SY5Y with SBD variants.
SBD was incubated with neuroblastoma SH-SY5Y for 30 mins at room temperature, washed, and imaged by widefield fluorescence microscopy after 1 hour. Five mid-cell optical sections were projected after deconvolution. All peptides can be seen to have been endocytosed into vesicular compartments. Peptides were: A) PEG-K16-NH3; B) PEG-K16-COO-; C) PEG-E16-NH2; D) PEG-E16-COO-; E) AEEAc-E16-COO-. Scale bar is 20 µm.