Optimizing Fluorescence Microscopy - GM 20250125Sat

[GeoMcNamara]

Optimizing Fluorescence Microscopy - GM 20250125Sat
I have a lot of tips posted online, at http://confocal.jhu.edu/mctips and for new 2025 content, http://confocal.jhu.edu/mctips/mctips_2025
I am putting some tips here in IBEX -- this and hopefully future posts (or maybe edit this?) -- to gain a new audience. My tips assume a some (or sometimes a lot) of fluorescence microscopy experience (can catch up if needed by internet and pubmed accessible literature searches).

1. Use Welcome! #1.5H high precision borosilicate glass coverglasses whenever possible. These should be 170 +/- 5 um thickness, smaller tolerance than standard Welcome! #1.5 (nominally 170 um (range 160 - 190 um) and other thicknesses (see below).
2. Marienfeld may be the main manufacturer -- there are many distributors (various prices!).
   https://www.marienfeld-superior.com/precision-cover-glasses-thickness-no-1-5h-tol-5-m.html
   Most high resolution objective lenses (NA >=0.7) are designed for 170 um borosilicate glass (R.I. 1.518).
   I note that ibidi Gmbh offers many coverslip and imaging dishes made of cyclic olefin polymer (COP), 180 um thick, slightly different refractive index, so comparable optical path thickness. Consistent thickness of all your coverglasses is equally or more useful for your experiments

• prices of coverglasses (and many other products!) can vary enormously, so do due diligence (see mctips_2025 for some leads on Welcome! #1.5H sources).
• if you do not know you and your lab are already using Welcome! #1.5H, you are probably not.
• I think IBEX imaging should be done with 40x/0.95 NA (air) or high NA oil immersion objective lens (1.3 or 1.4 or higher NA) Welcome! #1.5H coverglass - but if you routinely use 20x (please use 0.70 NA or higher, not 0.5 NA or lower), by using Welcome! #1.5H you can always re-image later with higher resolution.
• I am equally fond of -- and strongly recommend either -- widefield -> GPU spatial deconvolution and confocal -> GPU spatial deconvolution. New NVidia RTX 5090 GPU is ~105 Teraflops (FP32) so near instant gratification compared to traditional CPU or older generation GPU (RTX 4090 was ~85 Teraflop, RTX 3090 was ~25 Teraflop). RTX 5090 will also benefit from new workstation (faster other components).

from https://microscopy.arizona.edu/sites/default/files/2023-03/coverslips_for_microscopy.pdf
and https://www.marienfeld-superior.com/cover-glasses-thickness-no-1.html
Coverglass Thickness range
00 60 - 80 um
0 85 - 130 um (Cromey) or 85 - 115 um (Marienfeld)
1 130 - 160 um
1.5 160 - 190 um
1.5H 165 - 175 um (Doug Cromey's table has 170 - 180 um)
2 190 - 230 um
3 250 - 350 um (Cromey) or 280 - 320 um (Marienfeld)
4 430 - 640 um (Cromey) or 380 - 420 um (Marienfeld)
5 550 +/- 50 um
7 700 +/- 50 um

---

4. mStayGold2 (Brightness 138) and AausFP1 (Brightness 179) are the two brightest green fluorescent prtoteins (GFP) to date.
   https://www.fpbase.org/protein/mstaygold2
   https://www.fpbase.org/protein/aausfp1
   AausFP1 is unfortunately an "obligate dimer", so not ideal for enerating fusion protein transgenes.
   Request: could someone please generate, publish and post in FPbase (and email me your news, [email protected]) and post here at IBEX (even though IBEX nominally about antibodies etc):
   (i) tandem dimer AausFP1
   (ii) bright monomeric AausFP1
   (iii) yellow version monomeric AausFP1
   (iv) tandem dimer green AausFP1 - yellow AausFP1 ... goal being a terrific FRET construct.

5. Possibly best red (orange red) fluorescent protein: mScarlet3 and/or mYongHong
   https://www.fpbase.org/protein/mscarlet3/
   mYongHong - Fu ... Piatkevich preprint online 9/2024.

6. (back to IBEX fluorophores) IBEX already uses several conventional fluorophores, Brilliant Violets (BV421 and more). Many more fluorophores are available. For example, Guo 2024 mention that Bio-Rad StarBright (~30plex) are based on their PDots (Polymer Dots) -- and Guo et al describe improved brightness PDots.
   Guo Z ... Chiu DT, Vaughan JC 2024 Highly multiplexed fluorescence microscopy with spectrally tunable semiconducting polymer dots. Sci Adv 10: eadk8829, https://www.science.org/doi/full/10.1126/sciadv.adk8829

7. One of my favorite 2020 publications is Maillard et al report that D2O enables brighter orange - red - near-infrared fluorescence (total photon yield) than H2O. For a typical (for me) 35 imaging dish area volume of 100 uL, the cost of D2O is ~$0.10 (Sigma-Aldrich, aka MilliporeSigma, Merck KGaA). So: you may want to optimize what IBEX steps use D2O based aqueous media vs wash steps etc. (also photo "bleaching" may be slower or faster in D2O depending on molecules).
   Maillard J ... Furstenberg A 2020 Universal quenching of common fluorescent probes by water and alcohols. Chem Sci 12: 1352-1362. https://pubmed.ncbi.nlm.nih.gov/34163898 , doi: 10.1039/d0sc05431c.
   See especially fig 2g https://pubmed.ncbi.nlm.nih.gov/34163898/#&gid=article-figures&pid=fig-2-uid-1
   I do note that Oxygen radicals diffuse further (i.e. less susceptible to triplet quenching) which could be good or bad depending on context.

[GeoMcNamara]

GM again ... Please take the time to optimally refractive index (RI) match your specimen and mounting medium. If using standard immersion oil objective lens R.I. 1.518, match the immersion oil, the borosilicate glass coverslip, mounting medium AND do your best to permeabilize your specimen and make its internal structures match. THe latter often requires gradual infiltration from one medium to another.

More stuff:

• above I listed coverglass thicknesses. You should use Welcome! #1.5H for THIN specimens, such as 5 um thick (thin!) single human cells (HeLa, HEK, primary cell cultures, most cell lines). For VERY THICK specimens, if you R.I. match everything to 1.518, you could use #0 coverglass, ~100 um thick (or #00 if you can find product or have Marienfeld manufacture, QA/QC and ship to you), which gains you an additional ~70 um tZ-focusing range (you could manually sort through #0's to find the thinnest ones).
• You can MEASURE the thickness of your coverslips, some methods:
  a) "high precision" calipers, i.e. electronic calipers from online stores (amazon.com etc), though if 0.005 mm = 5 um "precision", may not be good enough to help you select the exactly 170 um thick (maybe could measure multiple times, average and SD your multiple measurements???).
  b) stand severa coverslips on end, and image with 10x or 20x objective lens (I did this using a 35 mm imaging dish, inverted microscope - and tweaked them to be a little separated ... effect is like books on a bookcase). One problem: the ends tend to be chipped, so exact measurements problematic).
  c) my favorite though low throughput: record Z values when you image from bottom to top side with (say) 20x/0.75NA objective lens (on a K-frame slide holder, can sequeeze the railings close so the coverglass sits on the detented shelves). Helps to mark (sharpie magic marker or diamond scribe scratches) both sides ("X" marks the spot: \ on bottom, / on top). Need to account for the R.I. of the glass, since the Z-drive values are with respect to air.
• You can use your spare coverslips (ex. #0, Welcome! #1, just retired from your imaging supply drawer) as spacers.

[radtkea]

Thank you George for sharing your knowledge! We are grateful!!!

[radtkea]

@GeoMcNamara Can you please check all the links in your posts? Some of them go to the Welcome page and not your intended destination? Thank you!