Water-in-Oil 2-Hr Protocol

[OneScientista]

Build-A-Cell POPCCholesterol 11.pdf

[zjuradoq]

Is there a way to determine if the vesicles are the intended internal solution? Also if the vesicles are bi-layered or residues of oil.

Are you also proposing to change the composition to 1:1 and skip making lipid film, step 0, altogether?

[OneScientista]

Is there a way to determine if the vesicles are the intended internal solution?

• You could use a fluorescent label in the inner solution. Vesicles contain the inner solution that was used to make droplets in the suspension. Also, note in Build-A-Cell-POPC.pdf (below) the repeat of the microscopy of these vesicles. This was done about 24 hours after formation. This is another piece of evidence that these are vesicles. It is my understanding that water-in-oil droplets are not stable over long periods of time.

Also if the vesicles are bi-layered or residues of oil.

• Multilamellar vesicles are sometimes observed. They are typically not round, but appear as vesicle clusters/aggregates with noticeable irregularities. Electron microscopy could be used to confirm the thickness of the vesicle membrane.

Are you also proposing to change the composition to 1:1 and skip making lipid film, step 0, altogether?

• Yes! One of the considerations for this protocol was the time it takes to prepare the vesicles. Omitting formation of the lipid film greatly reduces the time requirement.

[OneScientista]

Build-A-Cell-POPC.pdf

[murrayrm]

@gregor0607 Can you remind us of the status of the "unit test" for testing whether or not we have functional extract inside of a vesicle?

@BuildACell/liposome-curators Note that there is a proposal to change the standard protocol based on experience in implementation. If the curators agree this is something that should be done, it should be discussed (either here or in a new issue) and decided upon.

[zjuradoq]

@OneScientista Thank you for sharing the florescence images. I can see what you mean but I also wonder if you have used the internal solution check of 2 uM HPTS, 250 uM HEPES. This way the vesicles will also be seen in the GFP channel. Also, was wondering if your outer solutions differ from the posted protocol?

I look forward to trying this method and will let you you know what I get.

[OneScientista]

@zjuradoq These are all good questions. Below are the composition and concentration of inner and outer solution I've used so far. I will amend my report to include this information. I may try 2 uM HPTS in the inner solution in the future. Did you use HPTS or calcein?

Inner solution:
200 mM sucrose
100 mM HEPES pH 8

Outer solution:
200 mM glucose
100 mM HEPES pH 8

[PaolaTorre]

@OneScientista
I doubt that bright field or DIC images could truly lead you to the conclusion that you have vesicle.
I think that you have very nice water in oil emulsion droplets, stabilized at the interface with lipid!
My comments are the following:

1. I think we included the presence of the 18:1 Liss PE to have a clear distinction among the kind of spherical objects you may have at the end of the protocol. If you exclude this lipid for your formulation, you are basically excluding your only way to understand if you have vesicle or not.
2. I tried in the past to evaporate the chloroform solely by heating at 80C deg. I don't think is a good way to really get rid of the chloroform . I mean it is if we think that the melting point is around 60C deg.
   But again you are adding too many variables to the protocol. Think about it, you add the lipid in chloroform to the oil, already after this step you may have formation of phases. Than you evaporate at 80C deg and you are melting the chloroform, who is going to tell you that really you eliminate all the residual?

[OneScientista]

@PaolaTorre Hi Paola, please take a look at Build-A-Cell-POPC.pdf (above). There are images of fluorescently-labeled vesicles. Comments?

[PaolaTorre]

@OneScientista
Errata Corrige, that it is -60C deg, but as you said that the T of boiling is 60Cdeg.
If you want to be sure that you really have vesicle, as we said, you shoul probably od some confocal microscopy.
My suggestions are the following:

1. Combine the two protocols.
   2)Try to be more patient and careful with the evaporation of the chloroform.This means don't aim to prepare an oil in 5 mL. My preference is to prepare many oils of 1 mL. I always prefer to evaporate the chloroform overnight.

[p-stano]

Hi, to reply the initial question:

Is there a way to determine if the vesicles are the intended internal solution? Also if the vesicles are bi-layered or residues of oil.

as mentioned the easiest way is to include a fluorescent marker in the I-solution.

Determining the vesicle lamellarity is not straightforward, and needs more work. See the following two articles:

• Akashi, K.; Miyata, H.; Itoh, H.; Kinosita, K. Preparation of Giant Liposomes in Physiological Conditions and Their Characterization under an Optical Microscope. Biophys. J. 1996, 71 (6), 3242–3250.
• Chiba, M.; Miyazaki, M.; Ishiwata, S. Quantitative Analysis of the Lamellarity of Giant Liposomes Prepared by the Inverted Emulsion Method. Biophys J 2014, 107 (2), 346–354.

Determining the residual oil (meaning: "traces") solubilized inside the lipid bilayer is also difficult. In one work the C. Sykes group has looked at the membrane mechanical properties, and compare them with those of "pure" membranes. But this cannot be done in a normal preparative lab.

If the presence of residual oil hamper the scope of the experiment, a suggestion, already present in the Weitz 2003 papers is to shift from "mineral oil" to squalene, which is almost insoluble in the lipid bilayer.

[OneScientista]

@murrayrm @zjuradoq @PaolaTorre @p-stano
Hello All. Based on the comments and communication with the members of the containers group, the main issue appears to be the result variability between users. I've heard a lot of comments to the effect that: " this didn't work in our hands". This needs to be addressed.

To address the issue of variability, my intern Theresa Chu and I undertook preparation of vesicles side by side from the same lipid/oil stock (made fresh most days) for three days. We both observed vesicles, but also observed differences. The only part of the protocol that was different was our handling of the emulsification process. This was not intended, just minor differences in how we each handle tubes, how we use the vortex, how long we keep them at room temperature, how much they warm up etc... The next three reports include both Theresa's and my results. Note that we only used POPC for this exercise. Since then, I've used different POPC:cholesterol combinations as well, more reports to come...

Here is what we learned:

Emulsification process is critical to variability in the number and size of droplets produced. Lipid composition, concentration, temperature, and type and intensity of shear forces make the most impact on the outcomes.

What is clear is that the process of emulsification needs to be standardized. We have purchased a vortex head adapter that, if successful, should remove the operator factor from the emulsification process.

It also appears that we will need slight variations to this protocol for different lipid compositions.

Looking forward to devising a robust vesicle-making protocol, and unleashing it onto the world in the near future!!!

Build-A-Cell_082218_Theresa started w Lipids.pdf

Build-A-Cell_08-23-18_POPC.pdf

Build-A-Cell_08-27-18_POPC.pdf

[OneScientista]

@zjuradoq Hi Zoila, were you able to try the short protocol?

[p-stano]

Hi Milena, the emulsification seems to be indeed a key moment in this preparation method. As you said it depends on chemical factors (type of oil, type and conc. of lipids, type of I-solution, and ratio between these 3 things) and mechanical ones.

Your experiments shows that between-operator reproducibility is not so good as expected, but at the same time you have GVs consistently, every time you want. I would take this part of the story as the positive one.

The differences between preparations are visible in the size distribution, and maybe the subtler lamellarity. Your idea about standardization of emulsification sounds good. But to extend this to everyone probably the same vortex machine should be used too.

One can be happy to have a reproducible, yet wide, size distribution.
But surely the dream is to have monodispersed w/o droplets (and thus monodispersed GVs). In my experience this never happens with POPC in manual emulsification.
But w/o droplets could be made by microfluidics for instance, and centrifuged manually later on. See here.

My comments to images and text in the pdf files:

• there is no size bar, but I have the impression that vesicles are smaller than what one would like to have. Hint to make larger GVs: reduce lipids and/or reduce shearing force.
• for homogeneity in size: if we think to reverse micelles (5-20 nm), they are rather homogeneous is size because they are at the thermodynamic equilibrium. All states are accessible and easily convert into each other. This does not happen so well in POPC w/o droplets, and not at all in POPC vesicles. Following this hypothesis, perhaps doping the lipid with a small amount of a non-lipid surfactant can be something to try. Which surfactant? Several studies on w/o emulsion have been done by using pure Tween/Span-stabilized droplets, after Tawfik and Griffiths. They are compatible with gene expression, see also here. Would a POPC/Tween/Span mix give a more homogeneous w/o droplet size distribution? I never tried - but I would be curious to know the result.

[OneScientista]

@p-stano Thanks Pasquale! Always good suggestions, I'll take a closer look.

[OneScientista]

@zjuradoq @murrayrm @PaolaTorre @p-stano
Hello all, here is the first attempt at using HTPS.
Build-A-Cell_10_05_18_HPTS_vesicles.pdf

[PaolaTorre]

@zjuradoq
At first glance I will say that 2mM HPTS it is too much.
Ideally the result should be vesicles with a nice red ring and HPTS inside.
I will focus later in the details of the protocol you followed.

[zjuradoq]

@PaolaTorre
I control calls for 2 uM of HPTS, which is not visible in visible light but well images under GFP channel. I think the idea of image you are referring to is more like below.

@OneScientista
I ran through the 2 hour protocol, fairly quickly (need to redo it without rushing), but my major concern is that there is chloroform in the lipid and oil mixture. I found that after allowing the chloroform to evaporate I still had >2 mL of solution, which I take to mean that the solution was not only the 2 mL oil I added. Though I thought the lipid mixture looked well I fear that any additional chloroform, though may not hugely affect the production of vesicles, will effect the TXTL reaction.

[OneScientista]

@zjuradoq That is great news Zoila! Do you use a confocal microscope?

How much chloroform did you use? And how long did you evaporate? What was the surface area (what size beaker) and temperature?

P.S. I only used 2mM HPTS because our light source was very weak, so I reasoned that I would only be able to see a very concentrated solution. We have since had our microscope serviced and the light source is as bright as could be now with a new light bulb! We will redo HPTS at 2 uM and Rho-DPPE at 0.1 mol%.

BTW, I've expressed sfGFP using PUREfrex 2.0 inside of 1:1 POPC:chloroform (total lipid 10 mg/ml) vesicles several times already and I can see green vesicles with the new light source!

[OneScientista]

@zjuradoq Can you upload your protocol? I have a new suggestion for emulsification, but will test it a few times before posting a protocol. We find it much better for thorough emulsification of 20 uL inner solution in 300 uL lipid/oil.

[zjuradoq]

The total chloroform used was ~650 uL and was evaporated for 1 hour at 60C, mostly this long because I got held up with other things. The beaker was a 25 mL test tube of ~1 cm in diameter.

I do use a use a confocal microscope.

I will try again this week and have a clean protocol I will upload.

[OneScientista]

@zjuradoq Hi Zoila, Hmm, it sounds like you are not be removing bulk chloroform fully.

I would recommend using a standard 25-mL beaker, diameter is 4 cm and 80C. The 16 times larger surface area and 80C helps to remove most chloroform. I find that I can easily remove the 200 uL in 15 min (to the limit of my detection). What also worked for me was a 250-mL beaker (6.2 cm diameter) to remove 800 uL chloroform in 30 min at 80C. But, since time is of essence here, I would go with the recommended temp and a larger surface area but set a constant time.

[OneScientista]

Here are some vesicles with PUREfrex expressing sfGFP while being imaged on a slide. It took about 1 Hr to start seeing green fluorescence (the time it took me to take a bunch of photos).

[PaolaTorre]

@OneScientista
These are really nice vesicles! Even though you still have some residual of chloroform, as you can notice in the first image on the top. By leaving the tubes under vacuum O/N., after drying the chloroform you can avoid the formation of "rafts"at the interface of the vesicles.
@zjuradoq Yes, the confocal image is closer to what I meant, but I think you also should work on removing carefully the chloroform.

[OneScientista]

@PaolaTorre Can you share the literature about the effects of residual chloroform on raft formation?

[zjuradoq]

@PaolaTorre When you say leaving the tubes under vacuum, do you mean in the process of making a dry lipid film or evaporating the chloroform in the shorten procedure? I would also just point for clarity, the image I posted are vesicles produced starting from the dry lipid film. Does the comment about carefully removing the chloroform then imply for this image?

@OneScientista Thank you for the suggestions about how to better remove the chloroform. I will run thought the procedure again tomorrow and let you know. Also as for the expression time, I also find that in vesicles that expression is delayed.

[OneScientista]

@zjuradoq Interesting about expression time. I didn't yet compare with bulk Pfrex. Some publications show that expression continues for a longer time in vesicles as compared to bulk solution.

[OneScientista]

Here is a repeat of 2 mM HPTS vesicles. Theresa made these vesicles on 10/9. I reimaged them them today 10/16. These vesicles are 1 week old! She used a different emulsification protocol I mentioned above, but same evaporation conditions (200 uL chloroform at 80C, 4 cm-diameter beaker in 1 mL oil, 15 min).

[PaolaTorre]

@zjuradoq I meant the step of production of the lipid film. After removing the chloroform with nitrogen or argon, I generally move the tube in a desiccator over-night. I adopt this procedure to be really sure that any residual of chloroform is gone.
@OneScientista This is a paper that explain how chloroform behaves with lipid bilayer:
Chloroform alters interleaflet coupling in lipid bilayers: an entropic mechanism
Ramon Reigada, Francesc Sagués J R Soc Interface. 2015 May 6; 12(106): 20150197. doi: 10.1098/rsif.2015.0197.
I can confirm with my experience that the residual of chloroform can produce "rafts". This was also object of a conversation with @npkamat in the last meeting.

[OneScientista]

@PaolaTorre @p-stano @murrayrm @zjuradoq Is raft formation only associated with chloroform presence or are there other mechanisms for raft formation? How do we confirm that the observed increased red fluorescence indeed is a raft? Is formation of lipid rafts something we should worry about or is it an expected phase transition?
But more generally, maybe this is a good point to start talking about quantitative tests for the outcomes of these protocols. What benchmarks do we need to accept a change to a protocol? Some of the criteria were the speed of production, morphological characteristics, number, encapsulation efficiency. FACS seems like the only high throughput method for this characterization. Should we move toward this?

[zjuradoq]

@OneScientista I have usually seen rafts and have had different answers when I asked about them to others. I do not usually mix the oil and chloroform together and have started with the dry lipid film. To make the lipid film I have placed them in a vacuum with desiccator for 2 hours after evaporating over night. I always wondered if there could be additional chloroform or perhaps water is introduces during sonication. I would actually repeat your question to @PaolaTorre for some more insight.

FACS is a good way of characterization of, through I am not sure if you remember a method proposed by Aaron. It was a way to calculate encapsulation efficiency that did not require FACS accessibility.

[PaolaTorre]

@OneScientista
Lipids rafts can be reconstituted by selecting "ad hoc" lipid compositions. By looking at the chemical structure of the lipids chosen in a composition it is possible to predict if they can give rise to domains or not. There are phase diagrams, published in the literature, that can even explain which percentage of a certain lipid you will need to have rafts. But this is a whole biophysics chapter.
@zjuradoq
Correct, the oil should never be in contact with the oil. I am glad to hear that you dried the lipids and then that you left the sample in the desiccator and additionally overnight. It is possible that the sample can be contaminated by water during the step of sonication, which I suggest to perform at 50C deg for 30'. But if this is the case, you will see that the oil become cloudy. What do you do after sonication?
How do you dry the lipids? Nitrogen, Argon or what else?Let me know, I can advice accordingly.

[p-stano]

@OneScientista and all

Is raft formation only associated with chloroform presence or are there other mechanisms for raft formation? How do we confirm that the observed increased red fluorescence indeed is a raft? Is formation of lipid rafts something we should worry about or is it an expected phase transition?

Hello, I do not know a lot about lipid rafts, but from the discussion here it seems that trying a chloroform-free preparation is a way to verify that the GVs do not have "rafts", accordingly.
Maybe I said in the Slack chat that direct mixing lipid powder + oil also works (we used to do in this manner).

Your question: shall we worry about the "rafts" or whatever they are? From your images, even if the "rafts" are due to residual chloroform, it seems that it does not stop the TX-TL reactions. One can try, at least once, a comparison with a chloroform-free preparation.

Other effects? I expect, but I am not sure, that at the boundary around a "raft" is a locus for problems/opportunities. Perhaps this has been already studied by people working on true lipid rafts (permeability change? binding of some species?).

But more generally, maybe this is a good point to start talking about quantitative tests for the outcomes of these protocols. What benchmarks do we need to accept a change to a protocol? Some of the criteria were the speed of production, morphological characteristics, number, encapsulation efficiency. FACS seems like the only high throughput method for this characterization. Should we move toward this?

I can tell what we made some time ago (... a still unpublished study!): we wanted a rapid manner to estimate the "transfer efficiency". When a droplet reaches the flat oil/O-solution interface, two main events are possible: (1) the droplet is transformed into a vesicle, (2) or it breaks and release its content in the O-solution.

By including in the I-solution an easy-to-detect water soluble probe, one can calculate the

transfer efficiency = event 1 / (event 1 + event 2)

simply by measuring the concentration C of the probe in the O-solution and compare to the 0% efficiency (100% rupture). So, do not trow away the O-solution, but properly measure it. The efficiency is proportional to (C2-C1)/C2. We found typical values of 40-70%. This value also measures the captured volume (and therefore, indirectly, the vesicle size and number). The main advantage here is simplicity.
(Note: this is a "by difference" determination of entrapment. I would not apply this method when the expected entrapment is small).

Size distribution: yes, why not - with Image J (it has a standard tool for this). There are few clicks before applying the tool (preparatory operations on the digital images), but everyone can do it easily. Maybe the main problem is getting images of right quality (sharp difference between vesicles and background). Sample size: to have a mean +/- 0.5 um (more or less), the sample size N should be about 4*sigma^2. Maybe N > 200 vesicles sounds good.

[OneScientista]

@zjuradoq @murrayrm @p-stano @PaolaTorre A slightly updated protocol using microtube homogenizer to disperse inner solution. 2 uM HPTS, 0.02 mol% Rho-DPPE. Otherwise the same chloroform evaporation temperature and time.
The homogenizer info:
Wilmad-LabGlass_MicroTubeHomogenizerSystem_WLG Version.pdf

HPTS-102418 1138 AM-Build-A-Cell-uploaded.pdf

[p-stano]

@OneScientista
Thank you for the images and the protocol.
From the pictures, the samples look somehow similar with each other and with previous ones. I mean, it is difficult to detect differences from one picture. I suspect that these minor modifications are like small rounds near a region of the space parameters where the outcomes are more or less the same.

[OneScientista]

@p-stano One big difference is encapsulation efficiency. With vortexing a largish droplet forms at the bottom of the inverse emulsion. This droplet is very difficult or impossible to disperse (depends on the lipid composition). With this homogenizer, no droplet is visible. Having used it a few times, it looks like the fewer the pulses, the larger the vesicles produced. This seems intuitive.

[OneScientista]

@zjuradoq @murrayrm @PaolaTorre @p-stano sfGFP expressed inside vesicles. Protocol used mini homogenizer for emulsification.
Wilmad-LabGlass_MicroTubeHomogenizerSystem_WLG Version.pdf
Pfrex-101518 1223 PM_Build-A-Cell-uploaded.pdf

[OneScientista]

2nd Triplicate test-102918 359 PM .pdf
HTPS_20uM_Triplicate.pptx

[p-stano]

@OneScientista (referring to the message of 26 october)
Hi, at a first look, the protocol with the mini-homogenizer seems interesting, because I see (qualitatively) that the vesicles have a more homogeneous size, and that almost all vesicles appear green. After being sure that the negative control (+rhodamine -DNA) is not green, this homogeneity would be a quite nice result.

[OneScientista]

@p-stano Yes, this is GFP expression. I think we could get this standardized. So far it seems the way to disperse most efficiently. It is interesting that composition of internal solution (whether it's just buffered glucose or small components of Pfrex) also has a lot of influence on the quality of vesicles. After some thought, this should not be surprising.

[zjuradoq]

@PaolaTorre Sorry about long response, as for drying the lipid I leave them overnight then place it in a vacuum for 2 hours. I do believe that sonication could be where water is contaminating the oil and lipid solution so I have been more careful with this. Currently, have moved away from starting from dry lipid film and currently using @OneScientista proposal of mixing chloroform, lipid, and oil.

One difference that I still have is vortexing since we do not have a homogenizer.

[PaolaTorre]

@zjuradoq
I don't understand how you dry your lipids overnight. If you don't use any inert gas like N2 or Argon and you let evaporate the chloroform, there is an high chance that your lipids get oxidized and this can explain their later instability with other steps of the protocol.In general lipids are quickly dried with N2/argon and then placed O/N in a desiccator to let all the chloroform evaporate.
@OneScientista @p-stano
It is true the GFP-expression is working, but the rationale that we followed when we were selecting protocols for making vesicles at the workshop, was also to have a preparation that did not have any residual of oil at the interface. In other words, we were expecting a container that had truly a bilayer. If we think to a container as a cell, we would like eventually to be able to incorporate also membrane proteins, transmembrane proteins etc so the presence of additional phases (as the potential residual of chloroform/oil at the . interface) can interfere with these kind of reconstitutions...and yes even with external communication of the vesicles. If we still have oil/chloroform at the interface we still have an emulsion.
I will be curious to know how @npkamat and et al. they optimize the steps of formation of the lipid film and the oil preparation.

[p-stano]

Hello Paola, you are absolutely right. Residual chloroform and oil should be reduced/eliminated as much as possible. Whereas for chloroform straight ways exist (you mentioned in several messages), in the case of oil there is no obvious way if the droplet transfer method is applied.
(But squalene has been indicated by Pautot et al. 2003 as an alternative to mineral oil, because it is "insoluble" in the lipid bilayer. They cited this paper.
I understood that the second (oil-free) protocol is completed and should be tested as well, but testing has not started yet.

[PaolaTorre]

@p-stano I was familiar with the paper, but I didn't know about the squalene.