-
Notifications
You must be signed in to change notification settings - Fork 4
/
idr0061-study.txt
141 lines (102 loc) · 20.7 KB
/
idr0061-study.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
# FILL IN AS MUCH INFORMATION AS YOU CAN. HINTS HAVE BEEN PUT IN SOME FIELDS AFTER THE HASH # SYMBOL. REPLACE THE HINT WITH TEXT WHERE APPROPRIATE.
# STUDY DESCRIPTION SECTION
# Section with generic information about the study including title, description, publication details (if applicable) and contact details
Comment[IDR Study Accession] idr0061
Study Title Live imaging screen reveals that TYRO3 and GAK ensure accurate spindle positioning in human cells
Study Type high content screen
Study Type Term Source REF EFO
Study Type Term Accession EFO_0007550
Study Description Proper spindle positioning is crucial for spatial cell division control. Spindle positioning in human cells relies on a ternary complex comprising Gαi1-3, LGN and NuMA, which anchors dynein at the cell cortex, thus enabling pulling forces to be exerted on astral microtubules. We developed a live imaging siRNA-based screen using stereotyped fibronectin micropatterns to uncover components modulating spindle positioning in human cells, testing 1280 genes, including all kinases and phosphatases. We thus discover 16 components whose inactivation dramatically perturbs spindle positioning, including Tyrosine receptor kinase 3 (TYRO3) and cyclin G associated kinase (GAK). TYRO3 depletion results in excess NuMA and dynein at the cell cortex during metaphase, as upon blocking Phosphatidylinositol 3-kinase (PI3K), which TYRO3 activates. Furthermore, depletion of GAK leads to impaired astral microtubules, as upon down-regulation of Clathrin, with which GAK interacts. Overall, our work uncovers novel components and mechanisms governing spindle positioning in human cells.
Study Organism Homo sapiens
Study Organism Term Source REF NCBITaxon
Study Organism Term Accession 9606
Study Screens Number 1
Study External URL
Study Public Release Date 2019-07-03
# Study Publication
Study PubMed ID 31253758
Study Publication Title Live imaging screen reveals that TYRO3 and GAK ensure accurate spindle positioning in human cells.
Study Author List Wolf B, Busso C, Gönczy P
Study PMC ID PMC6599018
Study DOI https://doi.org/10.1038/s41467-019-10446-z
# Study Contacts
Study Person Last Name Wolf
Study Person First Name Benita
Study Person Email Benita.Wolf@chuv.ch
Study Person Address CHUV, Lausanne EPFL, Lausanne
Study Person ORCID 0000-0001-5673-4239
Study Person Roles submitter
# Study License and Data DOI
Study License CC BY 4.0
Study License URL https://creativecommons.org/licenses/by/4.0/
Study Copyright Wolf et al.
Study Data Publisher University of Dundee
Study Data DOI https://doi.org/10.17867/10000126
Term Source Name NCBITaxon EFO CMPO Fbbi
Term Source File http://purl.obolibrary.org/obo/ http://www.ebi.ac.uk/efo/ http://www.ebi.ac.uk/cmpo/ http://purl.obolibrary.org/obo/
# SCREEN SECTION
# Screen Section containing all information relative to each screen in the study including materials used, protocols names and description, phenotype names and description.
# For multiple screens this section should be repeated. Copy and paste the whole section below and fill out for the next screen.
Screen Number 1
Comment[IDR Screen Name] idr0061-wolf-spindlepositioning/screenA
Screen Sample Type cell
Screen Description Live imaging screen for regulators of mitotic spindle positioning, screening a library of 1200 kinases, phosphatases and associated proteins (4 siRNAs per gene), in a 96 well format. Cells were incubated 48 hours in the presence of siRNA and than transferred to micropatterned 96 well plates containing L-shaped, fibronectin coated, micropatterns, helping to stereotype cell organelle positioning. On L-shapes, mitotic DNA (metaphase plates) are positioned in a 45° angle with respect to the arms of the Ls, thereby creating a ""reference"" phenotype. We screened for candidates whose depletion led to a significant offset from this ""reference"" or normal position during metaphase. We used HeLa cells carrying the mCherry::H2B transgene, visualizing DNA. The L-micropatterns were coated with fibronectin containing Alexa 555, thereby Ls were visible in the same fluorescent channel. After transfer of cells to micropatterned plates, cells attached within 6 hours to the micropattern. We imaged each plate for 24 hours (framerate 8min), imaging 2 visual fields per well. Each plate contained negative and positive control wells (8 scrambles siRNA wells, 8 LGN siRNA wells). Movies were analyzed using the ImageJ based algorithm TRACMIT.
Screen Size Plates: 32 plates 5D Images: Planes: 1 focal plane Average Image Dimension (XYZCT): Total Tb:
Screen Example Images
Screen Imaging Method fluorescence microscopy
Screen Imaging Method Term Source REF Fbbi
Screen Imaging Method Term Accession FBbi_00000246
Screen Technology Type RNAi screen
Screen Technology Type Term Source REF EFO
Screen Technology Type Term Accession EFO_0007551
Screen Type primary screen
Screen Type Term Source REF EFO
Screen Type Term Accession EFO_0007556
Screen Comments
# Library section. The library file should be supplied separately and it should contain the reagents description including, at the absolute minimum: reagent ID, sequences and position in the layout (= plate + position in the plate)
Library File Name idr0061-screenA-library
Library File Format tab-delimited text
Library Type siRNA library
Library Type Term Source REF EFO
Library Type Term Accession EFO_0007564
Library Manufacturer Dharmacon
Library Version Dharmacon ON-TARGETplus® SMARTpool® siRNA Library, Human Druggable Subsets, G-104675-02 Lot 09159
Library Experimental Conditions
Library Experimental Conditions Term Source REF
Library Experimental Conditions Term Accession
Quality Control Description screening in duplicate, 4 siRNAs per gene, 8 positive and 8 negative control wells per 96 well plate, confirmation of screening hits with 2 different siRNAs plus CRISPR/CAS9 k.o. cell lines
# Protocols
Protocol Name growth protocol treatment protocol HCS library protocol HCS image acquistion and feature extraction protocol HCS data analysis protocol
Protocol Type growth protocol treatment protocol HCS library protocol HCS image acquistion and feature extraction protocol HCS data analysis protocol
Protocol Type Term Source REF EFO EFO EFO EFO EFO
Protocol Type Term Accession EFO_0003789 EFO_0003969 EFO_0007571 EFO_0007572 EFO_0007573
Protocol Description Growth Protocol - HeLa mCherry::H2B cells (gift from Arnaud Echard, Institut Pasteur, Paris6) were used for the screen. Cells were maintained in high-glucose DMEM with GlutaMAX (Thermofisher) medium supplemented with 10% FCS in a humidified 5% CO2 incubator at 37°C. For live imaging, the medium was supplemented with 25mM HEPES (Thermofisher) and 1% PenStrep (Thermofisher).
Treatment Protocol - For reverse transfection, 20 nM (0.1 µl per well) of siRNAs were distributed in 96-well plates (Greiner Cellstar, flat bottom) using a robotic liquid handling system (Biomek F). Negative and positive control siRNAs were distributed in the outer 8 wells on each side of the 96-well plates (see Fig. 1b). 10 µl of ddH2O was added to each well for a 5 min incubation of siRNAs, and Lipofectamin RNAi Max 2000 (Invitrogen, 0.3 µl per well) was also incubated in ddH2O for 5 minutes (10 µl per well) before 10 µl of the Lipofectamin solution was added to each well for a 20 min incubation. Subsequently, ~6'000 HeLa cells expressing mCherry::H2B were seeded per well and incubated for 48 h. CYTOO 96-well imaging plates containing L-shaped micropatterns were pre-incubated at 37°C for 72 h prior to cell transfer, and then in DMEM, 10% FBS for 30 min prior to transfer. Cells were washed once in PBS and detached from the well bottom using 40µl Accutase per well for 8 min (400-600U/ml, Gibco). To stop the detachment process, DMEM, 10% FBS was then added and the cell density adjusted to 6'000/100 µl, which was determined to be the optimal number to obtain maximal coverage of micropatterns by single cells. After pipetting up and down using large aperture tips (Finntip, Thermo Fisher) to obtain a suspension of single cells, cells were transferred to the micropatterned imaging 96-well plates. Cells were allowed to settle until full attachment, which took usually 4 h, before starting live imaging, during which plates were sealed (Breathe-Easy sealing membrane, SIGMA Z380059).
HCS Library Protocol - We choose a library of all human kinases, phosphatases and associated proteins. Those targets are drugable and given the fast changes during mitosis it is most likely that phosphatases and kinases play key roles.
HCS Image Acquisition and Feature Extraction Protocol - The HeLa cell cycle lasts ~ 24 h and we thus imaged up to that duration to maximize the number of analyzable cells, finding that the percentage of cells dividing is not strikingly different than when imaging for shorter durations (see Fig. Supplementary Fig. 1b). Moreover, we found that imaging for longer did not yield more analyzable cells because most cells had divided after that time, resulting in micropatterns being covered by more than one cell. We also determined the optimal frame rate that allowed us to capture at least 3 images during metaphase and the metaphase-anaphase transition. Since frame rates of 6 and 8 min did not differ significantly with respect to the correct detection of spindle positioning in the negative controls, we chose the latter to minimize data storage (see Fig. Supplementary Fig. 1a). Micropatterned 96-well plates were imaged at 37°C and 5% CO2 using a 10x (0.45NA) objective on the GE InCellAnalyzer 2200 microscope equipped with a sCMOS CCD camera and hardware autofocus. Imaging was conducted every 8 minutes for 24 h, capturing one focal plane with an exposure time of 100 ms, resulting in 180 frames per visual field; 2 visual fields per well were imaged using the Texas red fluorescent channel.
HCS Data Analysis Protocol - The imaging data was analysed initially using TRACMIT and then with the help of a custom KNIME workflow. In the first round, we screened 16 96-well plates and could include all of them in the final analysis based on them having an rSSMD >2 (see below). In the second round, we had to repeat the screening of 6 plates for the following reasons. In one case, we encountered a microscope focus issue. In another case, the majority of cells were dead. In two other cases, the number of cells to analyze was particularly low. In the two last cases, we repeated the screening for the same reason but finally did not include the repeated plates since they harbored even less analyzable cells than in the first pass. We realized that cell fitness was generally worse in the second screening round, as reflected by the slightly lower numbers or analyzable cells, as well as the initial shape of cells, as imaged by bright field microscopy. One plausible explanation is that the second batch of imaging plates contained a different glue that may have yielded a modest level of cytotoxicity. In order to determine if the data collected from each plate meets quality requirements35, we calculated the robust (r) SSMD (strictly standardized median difference) in each case, as follows: (1) rSSMD = (median(Cpos) – median(Cneg)) / ([MAD(Cpos)] ^2 + ([MAD(Cneg)] ^2). Where Cpos= positive control, Cneg= negative control (Fig. Supplementary Fig. 2c), MAD= median average deviation. As is apparent, rSSMD uses the median instead of the mean and was developed especially for siRNA-based screens with large variability, for which it is more suitable that the Z' score, which is best suited for more streamlined small molecule screens35. According to the rSSMD-based quality control criteria, values between 1 and 2 are considered as inferior, between 2 and 3 as good and over 3 as excellent (criterion 1b in 2,34). In order to decide how many cells should be analyzed per well to get a meaningful result, we performed a sample size calculation based on randomly picked replicates of negative and positive control wells, using plates for which 3 replicates were available. We used data from cells in 10 wells over three biological replicates and calculated their group means (10 for control siRNA and 25 for LGN siRNA), as well as the mean standard deviation between the screen replicates (5.7 for LGN siRNA and 2.7 for control siRNA). Based on this calculation, at least 8 cells per condition should be analyzed to obtain a meaningful result73. This could be confirmed by plotting mean and standard deviations of negative and positive control wells as a function of the number of cells per well (Fig. Supplementary Fig. 1h). Hits were selected as summarized in Figure 2 and described in the main text. In the first step, 14 candidates were excluded because no dividing cells were present in either round (PSMB3, PSMA3, GSK3B, PRSS12, PLK1 PIK3R2, BUB1B -most likely essential genes or genes critical for cell cycle progression), or at least in one of the two rounds (FLJ90650, CTSO, DGKH, PRKAG3, FLT3, CTDSP1, PKN3). The latter group of 7 genes originates from different plates, which made it unpractical to repeat the corresponding full plates. Candidate spindle positioning genes were identified using a step-wise procedure (Fig. 2a, Supplementary Data 1, Methods). In a first step, 14 genes were excluded because no cells were present in at least one round of screening, leaving 1266 genes for further analysis. Second, we set an empiric cut off at ≥30% of cells having to exhibit abnormal spindle positioning when combining the numbers from both screening rounds. We choose ≥30% to select candidates with strong spindle positioning phenotypes; LGN siRNA has a relatively weak phenotype, with ~ 26% of cells (±10%, n=256 wells) exhibiting abnormal spindle positioning2 (Supplementary Fig. 2b). This second filter left 231 genes to consider further (Fig. 2b). Third, we excluded 131 of these genes because their depletion led to abnormal spindle positioning in ≥30% of cells in only one of the two screening rounds. Fourth, movies corresponding to the two rounds for the remaining 100 genes were inspected manually. This led to the exclusion of 84 genes that were false positives owing to single cells being poorly located on the micropatterns, or else to more than one cell being present on one micropattern32. Overall, the above step-wise procedure led to the identification of 16 candidates spindle positioning genes (Fig. 2a-c). Representative movie sequences during mitosis following depletion of these 16 components are reported in Supplementary Figure 3. The 16 candidate genes comprised 5 genes that were described previously as having an impact on spindle positioning in mammalian cells. These are the kinases Aurora A15,16 and PLK218, as well as the phosphatase PPP2CA12. As anticipated, the screen also identified PLK4, whose depletion causes impaired centriole duplication, thereby resulting in spindle pole asymmetry and aberrant spindle positioning36. Moreover, we identified the Myosin Light Chain Kinase (MLCK), which is required for positioning the meiotic spindle in the mouse oocyte37. In addition to these 5 genes, another 5 that might have been expected to be identified based on previous work did not meet the stringent criteria of the above step-wise procedure (Supplementary Table 1). Therefore, we identified 5/10 of the genes that were expected from prior work to be associated with a spindle positioning phenotype. By extension, we estimate the false negative discovery rate of the present screen to be ~50%. Overall, we conclude that our live imaging screening pipeline has identified 16 candidate modulators of spindle positioning in human cells.
# Phenotypes
Phenotype Name control phenotype spindle positioning phenotype
Phenotype Description We refer to the position where the metaphase plate is at 45° from either arm of the L-shape as the normal position, and set it to 0 hereafter Cells with perturbed spindle positioning are expected to exhibit metaphase plate angles away from this position. Analyzing the outcome of the three test plates using genetic programming33 allowed us to establish that a metaphase plate angle ≥40° from the 0 position was the best discriminator between positive and negative controls
Phenotype Score Type automatic in first analysis and manual confirmation upon movie inspection
Phenotype Term Source REF CMPO
Phenotype Term Name mitotic metaphase phenotype
Phenotype Term Accession 208
# Raw Data Files
Raw Image Data Format TIFF
Raw Image Organization 16 96 well plates per screen round (2 rounds = 32 plates total), 2 visual fields per well, framerate 8min, 180 frames (24h)
# Feature Level Data Files
Feature Level Data File Name
Feature Level Data File Description
Feature Level Data File Format
Feature Level Data Column Name
Feature Level Data Column Description
# Processed Data Files
Processed Data File Name idr0061-screenA-processed
Processed Data File Format tab-delimited text
Processed Data File Description
Processed Data Column Name plate index % phenotype TRACMIT 1st n TRACMIT 1st % phenotype TRACMIT 2nd n TRACMIT 2nd Average % phenotype TRACMIT 1st and 2nd n TRACMIT 1st + 2nd % phenotype manual 1st n man 1st % phenotype manuel 2nd n man 2nd phenotypic categories Plate (first round) Well Reagent Identifier Gene Identifier
Processed Data Column Type identifier, plate, first and second round of screening value (%) value (n) value (%) value (n) value (%) value (n) value (%) value (n) value (%) value (n) identifier
Processed Data Column Annotation Level # for data and phenotype columns enter the level of the annotation so that it is clear if the value is derived from data from a single well, is averaged over replicates of a reagent, or is at the gene level. Choose from: well, single replicate of reagent, multiple replicates of reagent, gene, reagent and experimental condition, gene and experimental condition. well (two visual fields) well (two visual fields) well (two visual fields) well (two visual fields) average of 2 wells (4 visual fields) sum of 2 wells well (two visual fields) well (two visual fields) well (two visual fields) well (two visual fields) final result
Processed Data Column Description identifier, plate, first and second round of screening % cells showing spindle positioning phenotype in first round of screening as analyzed by TRACMIT algorithm that has been published under DOI: 10.1371/journal.pone.0179752 cells analyzed by TRACMIT first round of screening % cells showing spindle positioning phenotype in second round of screening as analyzed by TRACMIT algorithm that has been published under DOI: 10.1371/journal.pone.0179752 cells analyzed by TRACMIT second round of screening % phenotype average between first and second round total n first and second round as analyzed by TRACMIT (sum first and second) % phenotype after manual inspection, if performed, in first round of screening n cells in analysis after manual inspection (if performed), for first round % phenotype after manual inspection, if performed, in second round of screening n cells in analysis after manual inspection (if performed), for second round phenotypes found: HIT (according to HIT selection criteria described above), nh (no hit), nh after manual (false positive by TRACMIT, excluded after manual inspection), n to low (not enough cells to analyze)
Processed Data Column Link To Library File plate index Plate well gene name Gene Identifier