Real Time de novo Deposition of Centromeric Histone-associated Proteins Using the Auxin Inducible Degradation system

Measuring protein dynamics is essential to uncover protein function and to understand the formation of large protein complexes such as centromeres. Recently, genome engineering in human cells has improved our ability to study the function of endogenous proteins. By combining genome editing techniques with the Auxin Inducible Degradation (AID) system, we created a versatile tool to study protein dynamics. This system allows us to analyze both protein function and dynamics by enabling rapid protein depletion and re-expression in the same experimental set-up. Here, we focus on the dynamics of the centromeric histone-associated protein CENP-C, responsible for the formation of the kinetochore complex. Following rapid removal and re-activation of a fluorescent version of CENP-C by auxin treatment and removal, we could follow CENP-C de novo deposition at centromeric regions during different stages of the cell cycle. In conclusion, the auxin degradation system is a powerful tool to assess and quantify protein dynamics in real time.


INTRODUCTION
The ability to follow protein dynamics in vivo is extremely important to determine protein function and the role of specific proteins in complexes. This is particularly important for the assembly of large protein complexes, in which the association of specific proteins may rely on the presence of others.
Therefore, the analysis of the dynamics of each complex component may yield essential information on the global function and the assembly of the complex itself.
Centromeres are DNA/protein structures necessary to maintain the balance in genetic information by controlling faithful chromosome segregation during cell division. They are the foundation for the assembly of the kinetochore that, in turn, is required to interact with spindle microtubules. Centromeres are epigenetically identified by the presence of a histone H3 variant named CENP--A. CENP--A is strongly enriched at centromeric regions and is the base for the assembly of the entire centromere/kinetochore complex. Altogether, the centromere/kinetochore can form a complex with over 100 proteins of about 150 nm in human cells [1]. Recently it was proposed that the subunits of the Constitutive Centromere Associated Network (CCAN) interact between each other in an interdependent manner, and that this interdependency is necessary for the stability of the entire complex [2,3]. Although most of the centromeric proteins are present along all stages of the cell cycle, their assembly is highly regulated and often restricted to a limited time window [4]. For example, CENP--A is deposited only once every cell cycle, at the exit of mitosis [5] via a tight regulatory mechanism [6].
Here we describe the Auxin Inducible Degradation (AID) system as a method to track protein dynamics throughout the cell cycle. The AID system is based on the transplantation in eukaryotes [7] of a ligand--induced degradation system found in plants [8,9]. Addition of the plant hormone indole--3--acetic acid (IAA, auxin) mediates the interaction of the AID--tagged protein of interest with the ectopically expressed F--box from plant (Transport Inhibitor Response 1; TIR1), that can associate with Skp1 in eukaryotic cells and form an SCF--TIR1 complex to induce protein ubiquitination and degradation. AID size is about 25 kDa (similar to GFP), but can be reduced by ~1/3 without loss of functionality, as observed in budding yeast [10]. The auxin degradation system has already been successfully used in the past [2, 11--15] to study protein function following rapid and complete protein degradation at every cell cycle stage. As an example, using this system we recently demonstrated that CENP--A is dispensable for maintenance of an already assembled centromere/kinetochore complex [11].
In this chapter, we illustrate how to exploit one of the key characteristics of the AID system, its reversibility. We describe in detail how to ( Figure 1A): i) generate a stable human cell line expressing the E3 ubiquitin ligase TIR1, ii) insert an AID tag coupled with a fluorescent protein (mRFP

Retrovirus infection
Bio--safety measures for working with the virus are required 8. Day 0: seed DLD--1 cells in a 6--well cell culture plate at 20% confluency (3 wells/transfection). 9. Day 1: replace medium with 2 mL of culture medium containing 8 µg/ml Polybrene. 10. Day 2: add 250 µl, 500 µl and 750 µl of the virus to three separate wells and allow virus infection for 2 days. 11. Day 3: Wash--out the virus three times with 2 mL culture medium. Wash cells once with 2 mL 1x PBS, remove PBS and add 300 µl trypsin to detach cells from the tissue culture dish.
Incubate cells in trypsin for 5 min in the incubator. After about 2--5 min shake of the cells.
Confirm cell detachment using a bright light microscope. When all cells are detached from the plate re--suspend cells in 1.7 ml culture media. Pool and seed infected cells in a 15--cm cell culture plate. Add culture medium (total volume: 25 ml).

osTIR1 screening
Two to seven days after puromycin addition refresh the cell culture medium in order to remove dead cells and to ensure puromycin efficacy. Two weeks after puromycin addition it is possible to observe the formation of puromycin resistant colonies. For colony isolation, proceed as follows: 13. Sterilize small (~ 1cm 2 ) Whatman paper pieces.
14. Soak the paper in trypsin. Remove the excess of trypsin by squeezing the paper against an empty sterile surface (e.g. cell culture plate). 15. Wash the puromycin resistant colonies with 15 ml 1x PBS. Remove PBS. 16. Place the Whatman paper from step 14 directly onto a colony using sterile forceps. 17. After around 5 min in the cell culture incubator, cells start to detach. At this point, convey the colony into a 24--well plate containing 0.5 ml pre--warmed cell culture medium by transferring the whole paper.
18. On the following day, remove the Whatman paper and replace the medium. Immunoblot Next, use immunoblot using an anti--Myc antibody to screen all single cell--derived colonies for genomic integration and expression of osTIR1--9x--Myc construct. 19. Once the single cell clones are confluent collect cells by trypsinization as described in step 11. Here, use 100 µl of trypsin for each well of a 24--well plate and re--suspend cells with 400 µl culture media. Seed 100 µl of cells into a 12--well plate in duplicate. Add culture media to the 12--well plate (total volume: 1 ml). Use the remaining 400 ul for western blot analysis. Use this cell population for the subsequent genome editing process as described in the following sections.

AID--tag integration at the CENP--C locus
In the next step, introduce an AID tag combined with a fluorescent protein [here: monomeric red fluorescent protein (mRFP) or enhanced yellow fluorescent protein (EYFP)] at the endogenous loci of a centromeric histone--associated protein (in this case CENP--C, but a similar strategy was used to target the histone H3 variant CENP--A). To this end, this procedure describes use of a TALEN genome 8 editing strategy (Transcription Activator--Like Effector binding domain coupled with a FokI Nuclease), however, it can easily be adapted for a CRISPR/Cas9 or any other genome editing strategy (Note 2).
In this case, genome targeting plasmids were delivered by electroporation using a Lonza Cell Line Nucleofector Kit and the Lonza Nucleofector electroporation device (but other transfection methods could be used). Electroporation 1. One day before transfection, grow DLD--1 cells in a 10--cm cell culture dish in antibiotics--free cell culture medium (total volume: 10 ml). 9. Transfer cells carefully with a plastic pipette (provided in the kit) to a 6--well cell culture plate containing 2 ml pre--warmed cell culture medium without antibiotics. Shake the plate gently to dispense cells evenly. 10. One day post--transfection refresh the medium supplemented with antibiotics. Check efficiency of transfection using a fluorescence microscope (e.g EVOS FL Cell Imaging System).
Transfection efficiencies of 70% to 90% with 50--90% cell viability are commonly reached. 11. Grow cells for around three to five days and then expand to obtain a bigger number of cells.

Clones Selection
In order to select for DLD--1 clones in which the AID--mRFP/EYFP tag has been inserted a two--step FACS selection protocol was used (Note 5).
12. Three to five days after transfection, harvest cells in 15 mL Falcon tube by trypsinization as described above.
13. Wash cells with 5 ml FACS buffer and re--suspend cells in 0,5 to 1 ml FACS buffer. Also prepare a negative control with non--transfected DLD--1 cells in the same way. Before performing single cells sorting, screening of cells for positive integration using PCR screening is recommended (see below).

Clones Screening
Following FACS selection there are multiple possibilities to screen for AID--mRFP/EYFP integration.
Here we will present three of them (Note 9).

Cell viability screening
The most straightforward way is to initially screen for cell viability if the protein of interest is essential as in the case of CENP--C. To this end: 16. Seed clones derived from single cell sorting into duplicate 24--well plates containing regular medium or regular medium + IAA (total volume: 0.5 ml).
17. Clones that die within 4 days of IAA treatment are selected. 18. Expand the corresponding untreated single cell clones. 19. Confirm the integration of AID--mRFP in these clones by PCR (as described below) and immunofluorescence microscopy.

PCR examination for AID--mRFP integration
Perform standard PCR to test for the integration of the AID--mRFP (EYFP) tag (Note 10).
DNA extraction using QuickExtract™ DNA Extraction Solution: 20. Collect cells from a 1 mL fully confluent 12--cm cell culture dish by trypsinization and centrifugation in a 1.5 mL Eppendorf tube. 21. Wash cell pellets with PBS and re--suspend in 50 µl QuickExtract™ DNA Extraction Solution (scaling of cell pellet/solution is possible) (Other DNA extraction procedures can be used). 22  Acquire fluorescence microscopy images using a fluorescent microscope such as the DeltaVision Core system (Applied Precision). We use a 100x Olympus UPlanSApo 100 oil--immersion objective (numerical aperture 1.4) and a 250W Xenon light source. The system is equipped with a Photometrics CoolSNAP_HQ2 Camera. Acquire 4 µm z--stacks (step size: 0.2 µm). Quantification of CENP--C using an automated system [16] and ACA to mark centromeres reveals that CENP--C is completely depleted at the centromere after addition of IAA (Figure 1 F).

Auxin mediated depletion, re--expression and reloading of CENP--C AID--EYFP at the centromere
The auxin degradation system is a useful tool to study consequences of rapid and complete protein depletion. Due to its reversibility, the system can also be used to study de novo re--expression and localization of AID--tagged proteins including histone variants. Here we use the auxin degradation system to follow de novo deposition of CENP--C following its complete degradation. To study the dynamics of CENP--C deposition along the cell cycle we used the CENP--C AID--EYFP/AID--EYFP cell line in which also one allele of CENP--B has been endogenously tagged with an mRFP--tag (the CENP--B tagging was carried out in a similar way as described in the previous section). The presence of mRFP on CENP--B allows us to mark centromere position and to follow de novo CENP--C deposition by live cell imaging.
In order to completely deplete CENP--C AID--EYFP (henceforth referred to as CENP--C AE ) we added IAA to the culture medium. After CENP-- C AE depletion we washed--out auxin and allowed the re--expression of CENP--C AID--EYFP . We followed CENP--C AE reloading at the centromere by live--cell imaging in asynchronous cells or in cells arrested in S--phase (Figure 2 A).

Cell preparation
1. Day 0: seed cells in 5 adjacent wells of an 8--well IBIDI slide (total volume: 300 µl). 6. Replace media with CO 2 independent medium and start live--cell imaging immediately (for the S--phase arrested condition maintain thymidine in the cell culture medium).

Live cell imaging with fluorescence microscopy
7. Prior to image acquisition adjust temperature to 37°C.

For imaging, we use the DeltaVision Core system (Applied Precision) with an Olympus
60X/1.42, Plan Apo N Objective. The microscope is equipped with a CoolSNAP_HQ2 camera.

Conclusion
Several techniques exist to analyze protein dynamics such as Fluorescence Recovery After

Photobleaching (FRAP), photo--activation of proteins, Recombination Induced Tag Exchange (RITE)
and SNAP--based pulse labeling (for a detailed list see table 1 in [17]). The AID system described here has several advantages over some of the aforementioned methods to analyze protein turn over. First and most important, it allows the analysis of both protein function and dynamics in the same experimental setting by inducing rapid protein depletion (by IAA) and by achieving rapid protein re-expression (by IAA removal), both processes within minutes. Auxin is permeable (it passes through the cell membranes) and easy to wash out since it is dissolved in water (no need of DMSO), and it does not require the presence of any additional components (e.g. CRE recombinase in the RITE) to start monitoring protein dynamics. Since specific IAA--mediated protein degradation does not affect mRNA, the AID--tagged protein re--accumulates very rapidly, allowing live measurement of protein turnover at short timescales in every cell (in large numbers) and at every complex (in this case centromeres). Also, it does not require any particular laser or specialized equipment, so a standard fluorescence microscope can be used.
One disadvantage of the AID technique is that an "ad hoc" system is required for every protein of interest, involving extensive gene modification such as gene tagging that may disrupt protein function and the insertion of a transgene (TIR1). Also, target protein dynamics completely depend on protein expression levels, and there could be some photo--bleaching events during time-lapse experiments. However, at least one of these caveats is also found in all of the techniques mentioned above, highlighting again the unique advantage of using the AID system.
In summary, here we presented the auxin degradation system as a unique tool to study the de novo deposition of centromeric proteins. This tool could be adapted to study a wide range of other proteins and protein complexes to gain better insight into their function and dynamics.

NOTES
1. High level of osTIR1 expression is essential for rapid and complete degradation of AID--tagged target proteins. The majority of clones express the osTIR1--9x--Myc transgene due to the puromycin selection after virus infection, however expression levels might change due to the different integration sites within the genome. In order to decide which clones to pick, band quantification on the western blot of osTIR1--9x--Myc expression level is critical. The usage of a positive reference for TIR1 expression level [13] is recommended. We commonly pool all clones that display expression levels that are at least half as strong as the level of the reference clone. An alternative strategy is to AID--tag the gene of interest before adding the TIR1. This strategy will help if different levels of TIR1 are required to degrade the protein of interest (e.g. too much TIR1 might lead to leaky protein degradation in the absence of IAA), however it will preclude the usage of the "Cell viability screening" method described in this book chapter.
2. We used a TALEN--based genome editing approach, which allows site--specific genomic modifications with a very low chance of off--target genome editing effects. The design of left and right TALEN DNA recognition FokI fusion constructs have been extensively described [18] and, due to space limitations, will not be addressed here. Briefly, TALENS are cloned into a modified version of pcDNA3.1 (purchased from Invitrogen) containing also a cDNA sequence for the Fok I endonuclease domain. TALENs CENP--C target sequences: GAGGAAAGTGTTCTTC and GGTTGATCTTTCATC [16].
TALENs cleavage efficiency is tested by using a surveyor assays as described in Ran et al. (2013) [19]. Mis12, a crucial component for kinetochore formation [20], we decided to introduce the AID--mRFP tag at the C terminus. Previously, we have also introduced the AID--tag at the N--terminus of other proteins such as CENP--A with a very similar approach. In our opinion, the amino terminal tagging might be advantageous since following double strand break formation by TALENs or Cas9, DNA will be more frequently repaired via the non--homologous end--joining pathway. This error--prone repair can lead to rearrangements (deletion or insertion) and then, possibly, to de--activation of one of the alleles. In this case, the necessity to introduce the AID tag via HDR is reduced to one instead of two alleles. 7. If the integration of the AID--mRFP tag is very poor or not successful we recommend to use a recombinant adeno--associated virus system (rAAV) to deliver the repair template as described previously [21]. In this case, the efficiency of HDR is expected to be higher since the repair template is delivered as a linear ssDNA instead of a circular dsDNA plasmid.
8. If a single cell sort resulted in poor survival rates we recommend to sort 3 cells in a 96--well plate instead of 1 cell per well. After selection of a 3--cell population a limiting dilution can be performed to obtain a single clone. This will increase the chances of obtaining a positive clone.
9. At this step it is expected that at least one allele of the protein of interest (POI) is tagged. To achieve double heterozygous tagging more clones will likely need to be screened. Alternatively, a double selection strategy may be used (e.g. consecutive mRFP and EYFP tagging). If the protein is not tagged, we recommend to design different CRISPR/Cas9 guides and/or change the donor plasmid as described in Note 6, or to add to the construct a selectable cassette fused to a splicing site (e.g. P2A/T2A) or to a separate promoter. This latter strategy might be necessary if the POI is expressed at low levels and therefore not detectable by FACS. In the case of correct tagging but no protein depletion by IAA, a new target strategy needs to be generated. This might include moving the AID tag to a different protein extremity or adding a longer linker (in the case the AID is sterically hidden by the protein itself). 10. We designed primers that bind up-- (forward) and down--stream (reverse) of the donor template or inside the AID sequence (reverse). HDR--mediated integration of the tag will result in a shift of the PCR product size when using AID--mRFP flanking primers (e.g. with tag integration: 1900 bp, without tag: ~450 bp; Figure 1 C). A PCR product with the reverse primer binding inside the AID sequence is only expected in the presence of the AID tag.
11. Duration of auxin treatment is protein and cell line--dependent. Despite that for CENP--C we can observe complete degradation in only 20 minutes [11], for screening of correct AID integration we prefer to use an overnight treatment to be sure to achieve complete protein degradation.
12. Multiple washing steps are absolutely necessary to remove IAA properly. Washing needs to be done carefully since, if performed too harshly, it can detach cells from the dish.
13. Endogenous protein levels of CENP--C AE and also CENP--B mRFP are low in cells. This represents a problem to perform live--cell imaging since the RFP and EYFP signals are bleached over time. By acquiring images every 10 min we found a good compromise between total duration of acquisition (limited by bleaching) and time resolution to monitor CENP--C AE reloading.
14. CENP--C AE reloading cannot be observed immediately before mitosis (G2), during mitosis or early after mitosis (early G1) (Figure  2  B) as observed by filming asynchronous cells (the phases of the cell cycle were estimated by counting the time that the cells need to reach mitosis). However, we have found that CENP--C AE is reloaded in all S--phase arrested cells (Figure  2  C). This is in agreement with our previously reported findings that CENP--C can only be loaded at the centromere in mid G1--phase (only after CENP--A deposition that occurs immediately at mitotic exit [22]) and in S--phase [8]. We have found that CENP--C levels re--accumulate linearly at the centromere after about 30--60 min