A Comparative Study to Visualize Ptdins (4,5) P2 and Ptdins (3,4,5) P3 in MDA-MB-231 Breast Cancer Cell Line

Background: Phosphatidylinositol 3,4,5-trisphosphate [PtdIns(3,4,5) P3) and Phosphatidylinositol 4,5trisphosphate (PtdIns(4,5) P2] form an insignificant amount of phospholipids but play important roles in controlling membrane-bound signalling. Little attention has been given to visualize and monitor changes or differences in the local generation of PtdIns(4,5) P2 and PtdIns(3,4,5) P3 in the cell membranes of MDAMB-231 breast cancer cell lines. Methods: PLCδ1-PH-GFP and Btk-PH-GFP were used as biosensors to detected PtdIns(4,5) P2 and PtdIns(3,4,5)P3 respectively. These biosensors and antibodies were transfected, immuostained and then visualized by confocal microscopy on different cell surfaces. Results: Our results showed that PLCδ1-PH-GFP/mCherry was localized at the cell membrane, while BtkPH-GFP/mCherry was sometimes localized at the cell membrane but there was also a large amount of fluorescence present in the cytosol and nucleus. Our results also showed that the cells that expressed low levels of Btk-PH-GFP the fluorescence was predominantly localised to the cell membrane. While the cells that expressed high levels of Btk-PH-GFP the fluorescence was localization in the cytosol and cell membrane. Our results demonstrated that both anti-PtdIns(4,5)P2 and anti-PtdIns(3,4,5)P3 antibodies were localized everywhere in cell. Conclusions: Our results suggest that PLCδ1-PH-GFP and Btk-PH-GFP/mCherry have more specificity, reliability, suitability and accuracy than antibodies in binding with and detecting PtdIns(4,5)P2 and PtdIns(3,4,5)P3 and in studying the molecular dynamics of phospholipids in live and fixed cells.


Introduction
Phospholipids are the main constituent of all cell membranes and form lipid bilayers due to their amphiphilic nature. The phospholipid molecule contains two hydrophobic fatty acid "tails" and a hydrophilic "head" containing a phosphate group. The two constituents are combined together with a glycerol molecule. The phosphate groups can be amended with simple organic molecules such as choline or inositol (1).
Pleckstrin Homology (PH) domains consist of around 120 amino acids and have a seven-strand β-barrel which forms two anti-parallel β sheets and a C-terminal α helix. The β1, β2, β3 and β4 loops form specific sites to bind with the inositol ring of phosphoinositide (2). Approximately 10% of all PH domains have high affinity and specificity to bind with phosphoinositide (3). PtdIns (4,5) P2 has high affinity and specificity to bind with phospholipase C delta (PLCδ) PH domain and PtdIns (3,4,5) P3 also has high affinity and specificity to bind with Bruton's tyrosine kinase (Btk) PH domain due to neighbouring phosphates in their inositol ring (4). However, other types of phosphoinositide, such as PtdIns3P, PtdIns5P and PtdIns (3,5)P2 bind with another type of domains, such as FYVE domains, PHD finger and PX (3). This diversity in the specificity of phosphoinositide to associate with different types of PH domain containing proteins is important due to the recruitment of a variety of proteins to the cell membrane to perform their function and interact with other signalling pathways (3). Some of these proteins phosphorylate the inositol ring to generate a new phosphoinositide or produce new second messengers, such PI3K phosphorylate PtdIns(4,5)P2 to PtdIns (3,4,5)P3 and PLC cleaves PtdIns(4,5)P2 to Inositol trisphosphate (IP3) and diacylglycerol (DAG). Other proteins dephosphorylate sites on the inositol ring, such as PTEN dephosphorylates PtdIns (3,4,5)P3 to PtdIns(4,5)P2 (2,3). Phospholipid signalling plays a key role in the regulation of a substantial number of cellular processes, such as migration, cell proliferation, apoptosis, metabolism, and focal adhesions (5,6,7). Phosphatidylinositols (PtdIns) are a type of phospholipids that exist in the cytoplasmic layer of cell (8). They consist of a glycerol group which is esterified to two fatty acid chains and a phosphate group that binds to a polar inositol head group and extends into the cytoplasm.
Phosphorylated forms of Phosphatidylinositols are phosphoinositides and constitute a small percentage of the lipid in the plasma membrane. The amount of PtdIns has been evaluated and found that PI forms ∼10-20% (mol%) of total cellular phospholipids, (PtdIns(4,5) P2) form ∼0.2-1%, and PtdIns (3,4,5) P3 constitutes approximately 2-5% of the PtdIns(4,5) P2 (8). Until recently the direct detection of (PtdIns(4,5) P2) and PtdIns (3,4,5) P3 at cell membrane has been difficult. Therefore, in this study we aimed to use biosensors and antibodies to visualize them and examine which one is better to use. Experiments were performed using MDA-MB-231(ATCC® HTB26™),  and  MCF7 (ATCC® HTB22™) human adenocarcinoma cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA). MDA-MB-231 cells were tested by Mycoplasma kit (Cat No. K1-0210, EZ-PCR Mycoplasma Test Kit, 20 assays, Project. A3088606MDA-MB-231 cells were transferred into 10ml of DMEM (Gibco). The flask was incubated at 37 °C, 5% CO2 and 95% humidity until the cells become confluent. The MDA-MB-231 cells were grown in T75 flasks and cell passaging was used to maintain growth of cells. To count cells, a haemocytometer was used to determine the number of cells for each experiment and a formula was used:

Cell lines and cell culture
No. of Cells/ml= (average No. of cells counted/No. of cells wanted) x 10 4 x dilution factor.

Transfection
The ibidi dish was coated with one ml of 10 µɡ/ml v/v collagen (BD Bioscience), then seeded with 1 x 10 5 MDA-MB-231 cells for 24 hours at 70% confluence before transfection. Next day, 3 µɡ of PLCδ1-PH-GFP (Addgene; Plasmid #21179) or Btk-PH-GFP (Addgene; Plasmid #51463) was added in 100 µl of serum-free media in labelled Eppendorf tubes. 6 µl Polyethyleneimine (PEI) (Sigma Aldrich) transfection reagent was added to each labelled tube and then immediately mixed by vortex for 10 seconds. The labelled tubes were incubated for 15 minutes in the culture hood. The Plasmid/PEI mixture was added into each well which contains 2 ml of complete media. The plate was gently swirled and incubated overnight.

Confocal microscopy and live cell imaging
The ibidi dish was put in the confocal chamber and the chamber was connected to CO2 supply needle and warmed at 37 °C, one hour prior to use. The 488 nm and 568 nm channels of confocal microscopy were used for GFP-C1-PLCdelta-PHand Btk-PH-GFP or mCherry. Laser intensity and exposure time were adjusted to get rid of noise background.

Effect of Btk-PH-GFP overexpression
MDA-MB-231 cells were seeded and transfected. Our results also showed that the cells that expressed low levels of Btk-PH-GFP the fluorescence was predominantly localized to the cell membrane (Fig. 2). While the cells that expressed high levels of Btk-PH-GFP there was localization in the cytosol and cell membrane (Fig. 2). The fluorescence intensity in the cell membrane and cytosol were measured to compare the fluorescence intensity between these. In cells that expressed low levels of Btk-PH-GFP, the peak of fluorescence intensity in the cell membrane was around 150 levels of pixel intensity, while in the cytosol it was around 10 levels of pixel intensity. This referred that localization of Btk-PH-GFP in cell membrane was higher than in the cytosol (Fig. 2). In cells that expressed high levels of Btk-PH-GFP, the peak of intensity in the cell membrane was around 100 levels of pixel intensity, while in the cytosol it was around 125 levels of pixel intensity. This also referred that localization of Btk-PH-GFP in the cytosol was higher than in the cell membrane (Fig. 2).

Comparison between antibodies and PHdomains to detect PtdIns(3,4,5) P3and PtdIns(4,5)P2
MDA-MB-231 cells were seeded on fibronectin 20 µg/ml, then fixed with paraformaldehyde and stained with anti-PtdIns(4,5) P2 and anti-PtdIns(3,4,5) P3 antibodies. The results showed that both anti-PtdIns(4,5) P2 and anti-PtdIns(3,4,5) P3 antibodies were localized in the cytosol and cell membrane (Fig. 3). The fluorescence intensity in the cell membrane and cytosol of cell were measured to compare between these. The data showed that the peak of fluorescence intensity of PLCδ1-PH-GFP in the cell membrane was around 300 levels of pixel intensity and in the cytosol, it was 0 levels of pixel intensity (Fig. 3). While the peak of fluorescence intensity of anti-PtdIns(4,5) P2 antibody in both cell membrane and cytosol was around 250 levels of pixel intensity (Fig. 3). This referred that PLCδ1-PH-GFP had more specificity to bind with and detect PtdIns(4,5) P2. The peak of intensity of Btk-PH-GFP in the cell membrane was around 170 levels of pixel intensity and in the cytosol, it was around 20 levels of pixel intensity (Fig. 4). While the peak of intensity of anti-PtdIns(3,4,5) P3 antibody in both cell membrane and cytosol was around 60 levels of pixel intensity (Fig. 4). This also referred that Btk-PH-GFP had more specificity to bind with and detect PtdIns(3,4,5) P3.

Discussion
The aim of this study was to visualize and monitor the changes in levels of PtdIns(4,5)P2 and PtdIns (3,4,5)P3 directly in MDA-MB-231 cells. Our data showed that PLCδ1-PH-GFP which binds PtdIns(4,5)P2 (Fig. 1) was almost entirely recruited to the cell membrane in both the fixed and the live cells (Fig. 1), while an anti-PtdIns(4,5)P2antibody was localized throughout the cells, not just in the membrane (Fig. 2B). The probe Btk-PH-GFP for PtdIns (3,4,5) P3 was sometimes localized at the cell membrane, but in other cells there was a large amount of fluorescence present in the cytosol and nucleus in both the fixed and the live cells. The anti-PtdIns(3,4,5) P3 antibody was localized throughout the cell.
Many biosensors have been designed to detect PtdIns (3,4,5)P3, such as the cytosolic regulator of adenyl cyclase (CRAC) protein (8,9) and Akt-PH domain. Although Akt-PH domain is considered to be a good reporter, it also binds with PtdIns(4,5) P2 (10). Btk-PH domain is considered to be the best biosensor to visualize PtdIns (3,4,5)P3 as it has more affinity with, and specificity for PtdIns (3,4,5)P3 (11)(12)(13). Our results also showed in transfected cells that expressed only a small amount of Btk-PH-GFP, the majority of the GFP was recruited to cell membrane. MDA-MB-231 cells were transfected with Btk-PH-GFP for 24 hours, then the fluorescence intensity was measured in both transfected cells that expressed a small amount and more of Btk-PH-GFP.
Our results showed that transfected cells that expressed a small amount of the peak of fluorescence intensity in the cell membrane was higher than cytosol in the live and the fixed cells. While transfected cells that expressed more Btk-PH-GFP showed localization of the probe throughout the cell, and the peak of fluorescence intensity in the cell membrane and cytosol was slightly different in the live and the fixed cells. This may mean that Btk-PH-GFP interacts with PtdIns(3,4,5) P3 but it depends on the availability of PtdIns (3,4,5) P3 in the cell membrane to recruit Btk-PH-GFP to the cell membrane. Therefore, in transfected cells that expressed more Btk-PH, it is likely to saturate the available amount of PtdIns (3,4,5) P3 pools in the cell membrane, and the rest either binds with other types of inositides or remains in the cytoplasm (13).
It can be concluded that PLCδ1-PH is likely to be entirely recruited by PtdIns(4,5) P2 in the cell membrane. In transfected cells that expressed only a small amount of Btk-PH is likely to be recruited by PtdIns (3,4,5)P3, while in the transfected cells that expressed more Btk-PH, it is likely to saturate the available amount of PtdIns (3,4,5)P3 pools in the cell membrane and the rest either binds with other types of inositides or remains in the cytoplasm. It has been shown that PH-domains are more suitable, reliable and accurate than antibodies for visualizing and studying the molecular dynamics of phospholipids in the live cells (14,19). Indeed, our results showed that PH domains have more specificity and affinity than antibodies. However, there are some sideeffects, the overexpression of the PH domain fused with Green Fluorescent Protein (GFP) sometimes affects the physiological nature of the cell. It also can affect the binding of PHdomain with inositols. The overexpression of PH domain might block other kinases or proteins by interfering or interacting with their inositols which were assigned to interact with these proteins, this could affect their downstream signalling (12,17). Even though PH domain is the best option to visualize PtdIns(4,5)P2, not all of the lipids could be visualized using these methods. For instance, PtdIns(4,5)P2 and PtdIns (3,4,5)P3 might bind with other proteins, thus rendering them incapable of binding with the PH probes or antibodies, which could lead to inaccurate detection (20).
It is necessary to use PH domain in both the fixed and the live cells as both approaches have advantages and disadvantages. For example, in the fixed cell, it can be used with high power lasers and kept in a convenient location. However, the materials that are used for the fixation and permeabilization may affect the physiological and morphological nature of the cell might also affect the dynamic nature of PH domain-proteins interaction. Therefore, to get more accurate results, both the fixed and the live cells should be used (14). Confocal microscopy is the most commonly used method to detect the interaction between PH-domain and PtdIns(3,4,5)P3 and PtdIns(4,5)P2. Although confocal microscopy provides a very high quality image and has high resolution to monitor intracellular molecules, there are some limitations, such as inability to visualize the fine details of PtdIns (3,4,5) P3 and PtdIns(4,5)P2 due to diffraction limit (8). These resolution limitations are also known as the diffraction barrier, which limits the ability of optical resolution to visulize or distinguish between the smallest details of around 200nm by a lateral distance around half of the wavelength of light to image the specimen (21).
Recent studies have investigated PtdIns (3,4,5) P3 and PtdIns(4,5) P2 in more details due to the availability of modern techniques that can surpass the diffraction limit, which is around 200 nm (15,22). For example, the single-molecule super resolution microscopies have been designed to overcome these problems and measure the spatial colocalization more precisely (21,23). These methods have very high sensitivity to visualize the fine details of PtdIns(3,4,5) P3 and PtdIns(4,5) P2, and can even reach 10 nm in resolution. However, these methods have limitations, such as Stochastic Optical Reconstruction Microscopy (STORM) can only be used in the fixed cells and needs very efficient antibodies to detect PtdIns(3,4,5) P3 and PtdIns(4,5) P2 clearly (14,19,22,23). Due to difficulties in direct visualization of PtdIns(3,4,5) P3 and PtdIns(4,5) P2 at cell surface by antibodies, PH domains were used as the effective tools that bind selectively to PtdIns (3,4,5) P3 and PtdIns(4,5) P2 (24)(25)(26). In a nutshell, our results suggest that PLCδ1-PH and Btk-PH are more suitable, reliable and accurate than antibodies for visualizing and studying the molecular dynamics of phospholipids in the live and the fixed cells.