Second Generation Drosophila Chemical Tags: Sensitivity, Versatility and Speed

Labeling and visualizing cells and sub-cellular structures within thick tissues, whole organs and even intact animals is key to studying biological processes. This is particularly true for studies of neural circuits where neurons form sub-micron synapses but have arbors that may span millimeters in length. Traditionally labeling is achieved by immunofluorescence; however diffusion of antibody molecules (>100 kDa) is slow and often results in uneven labeling with very poor penetration into the centre of thick specimens; these limitations can be partially addressed by extending staining protocols to over a week (Drosophila brain) and months (mice). Recently we developed an alternative approach using genetically encoded chemical tags CLIP, SNAP, Halo and TMP for tissue labeling; this resulted in >100 fold increase in labeling speed in both mice and Drosophila, at the expense of a considerable drop in absolute sensitivity when compared to optimized immunofluorescence staining. We now present a second generation of UAS and LexA responsive CLIP, SNAPf and Halo chemical labeling reagents for flies. These multimerized tags with translational enhancers display up to 64 fold increase in sensitivity over first generation reagents. In addition we developed a suite of conditional reporters (4xSNAPf tag and CLIP-SNAP-Halo) that are activated by the DNA recombinase Bxb1. Our new reporters can be used with weak and strong GAL4 and LexA drivers and enable stochastic, intersectional and multicolor Brainbow labeling. These improvements in sensitivity and experimental versatility, while still retaining the substantial speed advantage that is a signature of chemical labeling, should significantly increase the scope of this technology.

1 Introduction myr::SNAPf in attP40 and attP2, UAS-myr::Halo2 in attP40 (Kohl et al., 2014), for 98 details of the new reporter lines generated in this study see Table S1. All images are 99 of female brains, apart from the brains in Figure 4d which are male, all flies were 100 dissected 3-4 days after eclosion.

115
Single and double channel labeling of Drosophila brains was carried out as previously 116 described (Kohl et al., 2014). For labeling of UAS-LA::Halo2 fillet preparation of 117 wandering third instar larvae were made followed by the same protocol used for la- were optimal for non-saturated images of the new reporters and images acquired at 153 the high setting were optimal for the old reporters so that we had a stack that could 154 be segmented for quantification and then the data from the low stacks were quantified 155 (see below). Confocal .lsm files were then converted to .nrrd files using Fiji. Using 156 Amira 6.0.1 (FEI, Thermo Fisher Scientific) a .nrrd stack, for each brain to be quan-  (Kohl et al., 2014) and the new reporters from this study. (b) Labeling of Mz19-Gal4 neurons using the old and new reporters. Each panel contains information on the dye used and insertion sites. Box plots show the quantification of fluorescence intensity of the axonal terminals of projection neurons in the lateral horn (arbitrary units). Boxplot n numbers were; GJ853 CD4::CLIPf on 2nd n=3, GJ851 CD4::CLIPf on 3rd n=4, P40 myr::4xCLIPf n=4, VK00005 myr::4xCLIPf n=4, P40 myr::4xSNAPf n=4, VK00005 myr::4xSNAPf n=5, P2 myr::SNAPf n=4 and P40 myr::SNAPf n=5 (c) New LexAop2-myr::4xCLIPf/4xSNAPf reporters labeling olfactory projection neuron using the weak GH146-LexA::GAD driver. (d) Orthogonal labeling of olfactory sensory neurons (green) and projection neurons (magenta) using new tags. Shown is the max intensity projection of a confocal stack after deconvolution. Images in panels b and c were acquired using the same microscope settings. Scale bars are 50 mm in whole brain images and 10 mm in higher magnification images of the boxed areas in panel d.
ures S1a and S2a). To bridge this gap and extend the use of chemical labeling to 196 most Gal4 driver lines, weak and strong, we designed a new generation of reporters 197 with greatly increased sensitivity. These reporters differ from the original ones in 198 two ways: first, they have a short 5' UTR (AcNPV) and the 3' UTR from the A.  (Table S1). 205 We tested these new transgenes and compared them to the first generation reporters   (Table S1). We tested these reporters us- localization. In order to improve cellular localization we replaced the N-terminal 257 myristolation with a C-terminal CAAX membrane targeting signal (Choy et al., 1999).

258
In addition we made several reporters with either one, three and seven tandem fusion-

262
We made transgenic flies with insertions in attP40, VK00005 and VK00027 (Table   263   S1). 264 We compared cellular localization and signal intensities from the first and new gener-265 ation of Halo tags in the same way as for CLIPf and SNAP. Nuclear signal is greatly 266 reduced in the new CAAX reporters when compared to the myristoylated ones (See 267 higher magnification images from the first two panels of Figure 2b). In addition, we  (Kohl et al., 2014) and the new reporters from this study. (b) Labeling of Mz19-Gal4 positive neurons using the old myr::Halo2 and new Halo7::CAAX reporters. All images were aquired using the same microscope settings. Lower panels are high magnification single slice images showing differences in reporter localization in the cell bodies (arrowheads) of olfactory projection neurons. Arrows indicate signal in glomeruli. The box plot shows the quantification of fluorescence intensity of the axonal terminals of PNs in the lateral horn (arbitrary units). Boxplot n numbers were; myr::Halo2 n=7, UAS-Halo7::CAAX-P40 n=7, UAS-3xHalo7::CAAX n=8 and UAS-7xHalo7::CAAX n=8. Scale bars in full brain images are 50 mm and higher magnification images of cell bodies 10 mm.
We wanted to explore the performance of chemical labeling in tissues other than the  numbers of neurons. While this approach greatly limits the number of labeled cells, 298 they often have overlapping processes which cannot be resolved by light microscopy.

299
In these cases further labeling refinements, using a number of genetic strategies, are 300 often required (Jefferis and Livet, 2012). We extended the applicability of chemical 301 labeling to these situations by developing reagents to: a) limit the number of la-  , Table S1).

314
As a proof of principle we used the conditional reporters in three experiments to in-  (Figure 4c.v).

333
The second strategy for resolving overlapping processes is multiplexing the label.  (Table S1). 344 We tested the new cassettes by labeling subsets of neurons that express the male

UAS-Syt::Halo7
Mz19-Gal4 Halo-TMR attP40    Figure 2b and 5a). The gradation in signal strength going from monomer to heptamer 371 makes these reporters useful for labeling synapses using drivers ranging from weak to 372 strong.

373
Next, we made a reporter for fast and sensitive labeling of actin filaments by fusing 374 a peptide, LifeAct (LA) that binds actin filaments to Halo2 (Table S1) (Riedl et al., (that is the line that gave fewest labelled cells with no heat shock). Next, newly 565 hatched larvae were heat-shocked for 10 minutes at 37°and allowed to develop into 566 adults. Flies were then processed as follows:

567
• All steps were carried out at room temperature unless stated differently.  Table S1: Transgenic flies generated in this study   Next three panels show chemical labeling of cell membranes and antibody staining of the OR22a receptor. All panels partial projections of confocal stacks that exclude the cuticle. All inset images are the corresponding confocal full projections. All scale bars are 50 mm.    Figure S12: HeatShock-Bxb1. Red restriction enzymes indicate that the site is destroyed during the assembly reaction. Part of the sequence for primer 1698* was not found on the cloned construct; the difference being upstream of the functional sequences does not affect its activity.
-The 2xHalo fragment was re-inserted into pJET p1.2-2xHalo7 digested with SpeI. During this step the XbaI site (red) combines with the SpeI site and gets destroyed while one SpeI site gets retained.
-UAS-3xHalo7::CAAX was treated with SpeI and the 4xHalo fragment ligated. During the cloning the XbaI site gets destroyed and the SpeI site is retained. Figure S16: UAS-7xHalo7::CAAX.
-UAS-Synaptotagmin::3xHalo was treated with SpeI and the 4xHalo fragment ligated. During the cloning the XbaI site gets destroyed and the SpeI site is retained.