Calcium responses of purified GECI proteins were measured in the cuvette, in collaboration with Marco Mank in the laboratory of Oliver Griesbeck. Protein purification and spec- troscopy of purified proteins was done by Marco Mank. I did comparative data analysis.
Data acquisition differs between the spectrometer and the 2-photon setup in several re- spects. First, data acquisition in 2-photon microscopy used bandpass filters of 40 and 30
nm spectral width (485/40 for CFP, 535/30 for YFP, 510/50 for GFP). Using the spec- troscope, emitted photons were sampled over the spectrum in bands of 5 nm. Integrating
the such acquired data spectrally over 40 and 30 nm or 50 nm before the calculation of ratio changes, results in a reduction of the fractional ratio changes. Thus, the bandpass filters used for fluorescence detection in 2-photon microscopy were mimicked by integrat- ing spectroscopic data over 40 and 30 nm for CFP and YFP respectively, or 50 nm for GCaMP 1.6 and OGB-1, before fractional fluorescence changes were calculated, to allow a fair comparison of data acquired with both techniques.
Second, the reference point for the calculation of fractional ratio changes differs. In spectroscopic calcium titrations, the applied concentrations ranged from 0 calcium to 39.5
µM. However, the cytosolic calcium concentration at rest has been determined as 31 nM
in presynaptic boutons. Calculation of fractional ratio changes from spectroscopy were thus done relative to the lowest calcium concentrations used (65nM). Single chromophore indicators GCaMP 1.6 and OGB-1 were analyzed accordingly. For D3cpv and OGB-1 the cuvette titrations show steep linear onsets at the lowest calcium concentrations applied (see Fig 3.14). The deviation of the mimicked resting calcium (65 nM) from the real, determined resting calcium (31nM) will thus have strong influences of calculated fractional fluorescence changes for these two indicators. Thus, I interpolated a fluorescence ratio for a ”virtual” calcium concentration of 31 nM and used this as reference for the calculation of fractional fluorescence changes. This interpolation was only done for D3cpv and OGB-1 as for all other GECIs the difference between fluorescence at 0 and 65nM was negligable (see Fig 3.14).
Together, both corrections allow direct comparison of the data acquired under 2-photon microscopy and spectrophotometric measurement with 1-photon excitation (see 2.3.4). The effectivity of these corrections was verified by measuring a calcium titration curve of TN- XXL protein in solution at the 2-photon microscope (see 2.3.4).
Yellow Cameleon 3.60 showed similar stimulus response properties in vitro and in vivo at low calcium concentration changes, however, above 0.5 µM (stimulus frequency > 40
Hz) fractional ratio changesin vivo were smaller thanin vitro (Fig 3.14 (A)). For TN-L15 (B), D3cpv (C), TN-XL (D) and TN-XXL (F) I also found good accordance of in vivo and cuvette measurements for calcium concentrations < 0.5 µM. In contrast to Yellow Cameleon 3.60 however, I found that fractional ratio changes were smaller in the cuvette thanin vivo at concentrations >0.5µM. For GCaMP 1.6 (E),in vivo signals were bigger thanin vitro at high concentration changes (above 1µM) and smaller at low concentration changes (below 1µM). Calcium response curves from OGB-1, recorded in the cuvette and in vivo (G) are similar over the full range of concentrations.
3.2 GECIs at the Neuromuscular Junction 79
Figure 3.14: In vivo vs. in vitro comparison of calcium evoked fluorescence changes of cal- cium indicators. Fractional fluorescence changes are plotted as a function of change in calcium concentration. In vivo responses (black) are directly com- pared to in vitro data from cuvette measurements (gray). GECIs measured in both conditions are (A) Yellow Cameleon 3.60, (B) TN-L15, (C) D3cpv, (D) TN-XL, (E) GCaMP 1.6, (F) TN-XXL. For comparison (G) OGB-1 is plotted.
Summarizing, GECIs with wild-type calmodulin (Yellow Cameleon 3.60 & GCaMP 1.6) show decreased fluorescence responses in vivo. All other GECIs (troponin-based GECIs
Figure 3.15: kDs and hill coefficients of GECIs as determined in vivo and in vitro in cu-
vettes. (A) kD values of GECIs are lower in vivo than in vitro. (Yellow
Cameleon 3.3 in vitro kD from [61]). (B) Hill coefficients for calmodulin-
based GECIs are decreased in vivo compared to in vitro (Yellow Cameleon 3.60 & 2.60 and GCaMP 1.6). Troponin-based GECIs show the opposite tendency: hill coefficients are increased in in vivo (TN-L15, TN-XL and TN- XXL). GCaMP 1.6 in vitro hill coefficient from [73].
and D3cpv) do not show this reduction. Their fluorescence change amplitudes even ap- peared increasedin vivo. Thus, the differences in GECI fluorescence responses under both conditions are probably due to effects of the cellular environment, such as calmodulin inter- actions as demonstrated in [71], because only those GECIs show reduced in vivo responses for which calmodulin related interactions are expected.
kD values were extracted from cuvette data and compared to values obtained from in
vitro experiments (see Fig 3.15 A). In general,kDs appear decreasedin vivo. This was not
true for Yellow Cameleon 2.60, which was excluded from the figure. Cuvette measurements have not been done and the determined in vivo kD is an overestimate as calcium signals
did not reach steady-state during 2 s stimuli, thus the degree of saturation and thus the calcium affinity cannot be deduced from these data.
Hill coefficients change fromin vitro toin vivo analysis as well (see Fig 3.15 B), but with a marked difference between calmodulin-based GECIs and all other GECIs tested: while Yellow Cameleons show reduced hill coefficientsin vivo, troponin-based GECIs and D3cpv display increased hill coefficients in vivo. This is in line with the idea, that heterophilic interactions between Yellow Cameleons and native cellular calmodulin or calmodulin inter- action partners perturb FRET response increases in a calium dependent way. For Yellow Cameleon 3.3 no in vitro hill coefficient is given in the literature.
3.2 GECIs at the Neuromuscular Junction 81