Review: PE and PS Lipids Synergistically Enhance Membrane Poration by a Peptide with Anticancer Properties

In 2008, an exciting lead for a new cancer chemotherapeutic was discovered in the venom of the Brazilian social wasp, Polybia Paulista. [1] Isolated in a petri dish, the protein Polybia MP1 showed a remarkable degree of specificity in cytotoxicity against leukemia, bladder and prostate cancers. Early publications on the protein showed that its mechanism of action was membrane perturbation; it ripped holes in the membranes of cancer cells. As far as chemotherapeutics go, this is new. Most other known chemotherapeutics target cancer cells by attacking peptides or DNA within the cell leaving the drug vulnerable to getting ejected by Pgp pumps embedded in the cell membrane; the cancer cell’s drug resistance mechanism of choice. MP1 bypasses this process by targeting the cell membrane itself, doing its job just as well outside the cell as it does inside.

The specificity however, is what makes MP1 such a fascinating molecule. It acts against cancer membranes leaving normal cells alone. It became clear in 2012 [2] that the reason for this was because the peptide required a certain set of anionic lipids to be present on the cell membrane surface to allow for it to bind and open significantly large pores. These were the lipids phosphatidylserine (PS) and phosphatidylethanolamine (PE) that were normally found in the inner leaflet of the cell membrane in healthy cells. In some cancer cells however, among the other massive changes in trans-membrane protein architecture, certain overexpressed pumps vital for the cancer cell’s increased metabolism carry PE and PS to the outer leaflet along with their usual transports thus enabling MP1.

The question remained however, how exactly do these two lipids influence MP1’s action? A journal article published recently by a UK/Brazilian research group tests this relationship using model membranes loaded with PE and PS lipids to isolate their interactions. The model membranes were large unilamellar vesicles (LUVs), or large bubbles of lipid formed by the hydrophobic interactions of phospholipids in an aqueous environment. In some cases when necessary to test certain hypotheses, even larger vesicles were used, known as giant unilamellar vesicles (GUVs). The research group formed 4 kinds of GUVs/LUVs containing a combination of PE, PS and a neutral lipid with no significant interactions, PC. These vesicles contained either PC/PE/PS, PC/PE, PC/PS or PC by itself.

This article, published in the Biophysical Journal, is a critical piece of the puzzle that explains how this peptide works and how it might be used to treat human disease. The experimental data gathered was grouped behind 5 individual claims about the mechanics of MP1 on PE and PS in a membrane. While the article has its flaws in particular areas of note, it is well written and its arguments are well supported by the evidence. The 5 claims are summarized below with criticism where necessary on the supporting evidence.

1) PS lipids enhance MP1 binding to the membrane

Figure 1

PS containing lipids showed higher K_p values during CD spectroscopy. This laboratory technique outputs a K_p value, the partitioning coefficient_, which increases when more peptides are bound onto the membrane of a LUV. One critique of this section of the paper is that the data interpretation was made quite difficult for the reader. The data plots showed no particular trends by themselves, as seen in figure 1, and the paper is sparse in explaining the mathematical reasoning from which the K_p values were calculated. While CD spectroscopy data interpretation is admittedly complex, a reader not familiar with the technique would have a hard time reaching the same conclusions as the authors with the graphs alone, so a brief explanation of the concepts and more detail on the math would have been beneficial for understanding.

2) Membranes containing both PE and PS are more susceptible to pore formation

Figure 2

Here, LUVs were incubated in a solution containing three (.37, 3 and 10kDa) different sizes of fluorescent dye, the exchange of which could be measured across the membrane using fluorescence microscopy. The amount of dye leakage across the 4 model membranes was plotted against MP1 concentration, in figure 2, to show that PE/PS/PC lipids showed the greatest leakage for the lowest concentration. It is noteworthy that in graph c and d of figure 2, the PC/PE and PC lines are seen overtaking the PC/PE/PS. This is not a contradiction; it shows that at equivalent concentrations of bound MP1 (PS enhances MP1 binding so these graphs normalize the binding of MP1), PE containing membranes have higher leakage rates, as in larger pores are forming. The interesting bit is that PC overtook PC/PE/PS; the authors claim this must be because PS has a minor inhibitory effect on MP1 pore-forming activity, so that at equal bound MP1 concentrations, having PS in the membrane is a disadvantage. The PS inhibition claim was, however, poorly developed and backed only by one set of data that could have easily been a fluke. Further investigation on this matter is needed before drawing any conclusions. The main conclusion of the argument however is solid and well supported by the evidence.

3) GUVs leakage mechanism is shown to follow the all-or-none model

Figure 3

Vesicle leakage can be described in one of two possible models; either the vesicle leaks gradually through a set of small pores, known as gradual release, or gets lysed completely leaking everything at once, known as the all-or none model. A fluorescence-requenching experiment set to measure immediate dye flow from individual GUVs shows that the model is very strongly aligned with all-or-none, as seen in figure 3. The data shows a strong trend but is contradictory to earlier LUV experiments where the vesicles were unquestionably shown to be gradually leaky. This section of the experiment would have benefitted greatly from trying different concentrations of MP1 to induce leakage, since the chosen concentration may have been so high that it caused the all-or-none leakage event. Furthermore, the data was measured from individual GUVs as opposed to populations of LUVs so the data is weaker for drawing conclusions. The explanation offered by the paper, that the pores must have been large and long-lived enough to make the data seem to be all-or-none, is a wishy-washy explanation that works against the strong trend observed in the data.

4) PE and PS synergize to cause MP1 induced leakage faster then either alone.

In the GUV experiment described in the third argument, time was measured between when the MP1 was added to when the first leakage event of the smallest dye occurred. The data table clearly shows the PE/PS/PC GUV having a significantly smaller time to leakage then the other 3 and the data largely makes sense. This conclusion is well supported by the evidence presented.

5) PE containing membranes have significantly higher permeability and pore size.

The time to leakage data discussed in argument 4 was plotted log-linearly against leakage intensity to generate a value for permeability between the 4 membranes. That data was then fed through the Einstein-Stokes equation relating it to fractional permeable area to show that PE containing lipids had greater permeable area sizes (pore sizes) by fraction than membranes that did not contain PE. The resulting data is quite strong. Further supporting this conclusion, atomic force microscopy images show a size difference in the pores of PE containing lipids. The images were the strongest set of evidence for the argument showing an easily discernable difference in size and frequency of the pores, and leave little room for doubt.

Main Citation

All images/figures and content are from and in reference to:

Leite,Natália Bueno, et al. “PE and PS Lipids Synergistically Enhance Membrane Poration by a Peptide with Anticancer Properties.” Biophysical journal 109.5 : 936-47. Web.

Supporting Citations

[1] Wang, Kai-Rong, Bang-Zhi Zhang, Wei Zhang, Jie-Xi Yan, Jia Li, and Rui Wang. “Antitumor Effects, Cell Selectivity and Structure–activity Relationship of a Novel Antimicrobial Peptide Polybia-MPI.” Peptides29.6 (2008): 963-68. Web.

[2] Cabrera, Marcia Perez Dos Santos, Manoel Arcisio-Miranda, Renata Gorjão, Natália Bueno Leite, Bibiana Monson De Souza, Rui Curi, Joaquim Procopio, João Ruggiero Neto, and Mario Sérgio Palma. “Influence of the Bilayer Composition on the Binding and Membrane Disrupting Effect of Polybia-MP1, an Antimicrobial Mastoparan Peptide with Leukemic T-Lymphocyte Cell Selectivity.” Biochemistry51.24 (2012): 4898-908. Web.