Identification of RNase-sensitive LINE-1 Ribonucleoprotein Interactionsby Differential Affinity Immobilization

Long Interspersed Nuclear Element-1 (LINE-1, L1) constitutes a family of autonomous, self-replicating genetic elements known as retrotransposons. Although most are inactive, copious L1 sequences populate the human genome. L1s proliferate in a ‘copy-and-paste’ fashion through an RNA intermediate; a full-length L1 transcript is ~6,000 nucleotides long and functions as a bicistronic mRNA that encodes and assembles in cis with two main polypeptides, ORF1p and ORF2p, forming a ribonucleoprotein (RNP); L1 RNPs also interact with a wide range of host factors in positive and negative regulatory capacities. The following protocol describes an approach to affinity enrich ectopically expressed L1 RNPs and, using RNases, release the fraction of protein that depends upon the presence of intact RNA for retention in the immobilized macromolecules.


Procedure
The protocol below anticipates that suspension-grown, SILAC-labeled cells will be used to enable a final readout by quantitative MS. In the case that SILAC labeling is not necessary for your experimental design, effective standard procedures for L1 expression in HEK-293TLD cells are thoroughly described in Taylor et al., 2016; review of this reference is recommended in any case. In the case that adherent cells will be SILAC labeled, see Taylor et al., 2013. A. Express ORF2p-3xFLAG-tagged L1 in SILAC labeled HEK-293TLD suspension cultures SILAC-labeled cells will be transfected in suspension using pLD401 plasmid DNA, and/or other 3xFLAG-tagged constructs such as pMT302 (Taylor et al., 2013), and PEI (polyethylenimine); after transfection, L1 expression will be induced by the addition of doxycycline (Dox; typically 1 μg/ml).
High-quality endotoxin-free DNA is critical to success. pLD401 contains a synthetic, recoded human L1 sequence (Orfeus-Hs) encoding a carboxy-terminal 3xFLAG-tagged ORF2p, expressed under the control of the tetracycline response element promoter Taylor et al., 2013).
Suspension culture yields large human cell pellets with minimal work, waste, and cost when compared to monolayer adherent growth. We typically obtain ~1.5 g HEK-293TLD cells, wet cell weight, per 100 ml culture. In the past we have used square Corning Pyrex bottles in an orbital platform shaker at 130 rpm, installed inside of a 37 °C, humidified incubator, maintained at 5%-8% CO2 (Muller et al., 2005;Taylor et al., 2016;Domanski and LaCava, 2017); however, more recently we have obtained comparable results using Erlenmeyer flasks at the same speed. For instructions on setting up the hardware for such a system, refer to Taylor et al., 2016. In addition to procedures we describe, standard mammalian cell culturing techniques and best practices apply (Freshney, 2011;Uphoff and Drexler, 2013).
If starting from adherent HEK-293TLD cells, follow the HEK cell suspension adaptation protocol presented in Domanski and LaCava, 2017; however, unlike the cells described in that procedure, frozen stocks made from suspension-adapted HEK-293TLD cells can be used to directly inoculate new suspension cultures successfully. If starting from a frozen stock or recently suspensionconditioned HEK-293TLD cells, proceed as detailed below. Cells may grow to a density of ~10 million/ml but growth slows after ~5 million/ml and transfection efficiency drops when cells are not in logarithmic growth.

IMPORTANT:
We have observed that, over time, the Dox-inducibility of HEK-293TLD can markedly decline. This has been determined to be caused by loss of expression of rtTA. Therefore, these cells should be periodically re-selected for blasticidin resistance (the BSD gene [Kimura et al., 1994]) was co-incorporated with rtTA), ensuring a transcriptionally active rtTA locus. Although keeping cells under continuous selection is the surest way to ensure a maximally Dox-responsive cell population, cells selected prior to making frozen stocks respond in sufficient proportion and uniformity for the work described below; it is also gentler on the cells overall. We therefore periodically re-select HEK-293TLD cell populations under challenge of 10 µg/ml blasticidin S prior making stocks, but do not maintain cells under continuous selection. 9 www.bio-protocol.org/e3200 3. Grow to ~5 million/ml.
Obtaining accurate cell counts can be challenging because HEK-293TLD tends to grow in small clumps of up to ~30 cells in suspension. Accurate counting requires dissociation of the clumps by gentle trituration using a pipette.
i. Using a 1 ml Serological pipette, aliquot 200 μl of culture to a clean microcentrifuge tube.
ii. Mix by flicking, and pipette 10 μl onto one side of the hemocytometer (also called hemacytometer).
iii. With a 200 μl pipette, set the volume to ~150 μl and triturate 30 times to break up clumps.
Try not to foam. iv. Pipette 10 μl onto the remaining half of the hemocytometer.  5. Grow to ~2.5 million/ml in 200 ml medium (~4 doublings). We have successfully transfected at up to 4 million/ml but recommend this concentration to ensure logarithmic growth.

On Day 1-transfect the cells in suspension
Each 200 ml suspension culture will receive 1 µg/ml pLD401 DNA and 3 µg/ml PEI 'Max' c. Transfer a 1 ml aliquot(s) of each culture for Western blotting to a microcentrifuge tube. Spin at ≤ 1,000 RCF for 30 s, aspirate the media, and store on ice until freezing is convenient. 11 www.bio-protocol.org/e3200

Materials and Reagents B25). If the expression levels are very different (greater than
±20%), the samples may not be suitable for quantitative comparisons or will require thoughtful normalizations.
d. Spin the cultures at ≤ 1,000 x g for 10 min at 4 °C to pellet.
This can be done in several 50 ml conical tubes for each batch, or in fewer, larger bottles.
Note: The pellets should appear approximately uniform. If centrifuged too hard, cells will be crushed, forming two different-colored layers. The goal here, and in subsequent centrifugations of these cells, is to obtain a relatively tight-packed cell pellet that maximally excludes buffer without breaking the cells. We consider 1,000 x g to be the maximum nominal centrifugal force to be used. We are commonly using a Sorvall T6000D or Beckman Allegra X-14R with swing-bucket centrifuge rotor at 2,000 rpm (~830-930 x g). One may want to establish their choice operating speeds in a preliminary experiment. ii. Pre-label 50 ml conical tubes. Punch a number of holes in the caps of the tubes using a 16 Ga needle. This will permit the LN2 to be drained from the tubes later without loss of cell material. These caps will later be replaced with un-punched caps. To avoid plastic waste, punch caps taken from used 50 ml conical tubes destined for disposal (wash caps first) and save punched caps for future use. Each tube will hold approximately 15-20 g extruded cells, which we often refer to as 'BBs' because of their round shape.
iii. Transfer the pre-labeled tubes to the rack in LN2. Fill the tubes with LN2. 12 www.bio-protocol.org/e3200 C2f) to homogenize, and then centrifuge the extract for 10 min at 20,000 x g to produce a clarified extract that can be measured by e.g., Bradford or BCA assays.
b. Alternatively, extract proteins in a Tris-buffered solution containing 2% (w/v) SDS and measure protein content by BCA assay-SDS is not well tolerated by the Bradford assay (Zor and Selinger, 1996). SDS-based solutions tend to produce viscous cell extracts requiring a greater degree of sonication than typical IP solutions. 13 www.bio-protocol.org/e3200 i. Add DTT to 25 mM to the 10 µl sample(s) (Step B2a) and heat at 70 °C for 10 min, and then briefly chill samples on ice.
ii. Add iodoacetamide to 100 mM and incubate at room temperature for 30 min in the dark.
iii. Load a range of aliquots-e.g., 1, 2, and 4 µl or 3 and 6 µl-on e.g., a 10-well, 1 mm, 4-12% Bis-Tris gel. Changing the stain and incubating several hours to overnight in fresh (green-hued) blue silver will typically resolve this problem and result in mild, uniform background staining with major bands clearly visible above background.
ii. Destain background thoroughly-several hours to overnight. 1) For blue silver, simply use several changes of de-ionized water.
2) You may choose to take an image of the gel. Always clean surfaces used to image gels that will contribute sample to mass spectrometry analyses.
c. Cut the regions of gel containing the samples from just below the well to just above the dye front (referred to as a 'gel plug;' will appear similar to those depicted in Figure 4C, but more protein rich)-Day 1. Note: Continue to destain as needed (this step can go overnight).
vi. Remove destain when the gel pieces no longer exhibit blue staining.
Note: They should appear whitish and semi-transparent with no obvious blue remaining.
vii. Dehydrate gel pieces with 500 µl acetonitrile, vortex, let the tubes sit at room temperature for 2 min or more.
Note: Gel pieces will appear opaque and thoroughly white.
viii. Remove acetonitrile from all samples.
x. Samples can be stored at -20 °C or proceed to trypsin digest. d. In-gel digestions of proteins to peptides using trypsin-Day 2.
i. Prepare all tubes on ice.
iii. Add enough 12.5 ng/µl trypsin to cover the gel pieces.
Note: This will vary by gel plug dimensions, but in our experience is ~38 µl for gel plugs obtained from 15-well, 1 mm thickness NuPAGE gel, and ~90 µl for a 10-well gel of the same.
iv. Let samples swell for 45 min on ice.
v. After swelling, add 50 mM ammonium bicarbonate to cover gel pieces if necessary.
vi. Move the tubes to a 37 °C incubator to digest overnight (or at least 6 h). e. Terminate the digest; collect and pool the peptides-Day 3.
ii. Add 1/3 rd the digestion volume of 2% (w/v) TFA to the digestion mix (0.5% w/v TFA final concentration) to stop digestion.
iii. Incubate at room temperature for 5 min with shaking in a Thermomixer iv. Remove the acidified digest solutions to microcentrifuge tubes ('protein low-binding' optional); hold at room temperature or 4 °C while pieces are further extracted.
IMPORTANT: Performance may vary, some considerations discussed here (Kraut et al., 2009). Tubes that will be exposed to 40%+ acetonitrile (as below) may benefit from

then repeat these tests periodically (or, when a problem occurs) because occasional (usually unannounced) manufacturing changes may alter performance. Pre-washing reduces the potential for contamination of the sample by components of tube
manufacturing by first solubilizing them in e.g., high acetonitrile and removing them with the wash. We routinely use ~100 µl acetonitrile to wash 0.65 ml tubes and ~300 µl to wash 1.5 ml tubes as a precaution (although it may not always be necessary-testing is advised for any doubt). Tubes we commonly use: Sorenson low binding (e.g., #11300,
These are also available in low protein binding varieties if desired.
v. Extract with 50 µl of 0.1% (w/v) TFA at room temperature for 45 min in a Thermomixer or vortex mixer equipped with multi-tube holder, speed setting as above.
vi. Remove the extractants and combine them with the cognate digest solutions. The total volume may be ~100 µl for 15-well and ~150 µl for 10-well gel plugs.

IMPORTANT: The manufacturer's protocol stipulates pipette-based desalting. We instead use a centrifugal approach because in our experience loading the samples on top of the
3) Repeat the two steps above, once.
Note: Check the volume of waste collected during centrifugation steps; make sure that the OMIX tip will not touch the waste in the collecting tube. Aspirate the liquid in the collecting tubes when needed.
iii. Wash tips by applying 2x 100 µl of tip washing solution.
3) Repeat the two steps above, once.
3) Repeat the two steps above, three times.
vi. Elute the desalted peptides from the C18 resin. not typically use the "re-quantify" option; and a typical result with these settings will identify and quantify ≥ 300 proteins. The "peptide.txt" output file can be used to calculate the heavy amino acid incorporation (Geiger et al., 2011). Peptides corresponding to known exogenous contaminants such as keratin and trypsin should be removed. Lysine-and argininecontaining peptides are considered separately. The "ratio H/L" for each peptide is and store at -20 °C. The slurry can be stored in this way for at least 1 year without any noticeable loss of performance.

Affinity purification in conjunction with on-the-beads RNase treatment
A general procedure describing and demonstrating best practices for affinity purification using mammalian cell powder and magnetic media has been previously described (LaCava et al.,   2016); a procedure specifically for anti-FLAG affinity capture of ORF2p-3xFLAG-tagged L1 RNPs has also been detailed (Taylor et al., 2016); the following implementation has been modified to test the effects of RNase A/T1 treatment on affinity immobilized L1 RNPs. Before carrying out a SILAC-based quantitative MS experiment (described below), preliminary studies should be carried out using general protein staining and Western blotting (e.g., see Figure 4A, The mixing of samples a-d is displayed schematically in Figure 4B and the resulting SDSpolyacrylamide gel is shown in Figure 4C b. Samples c and d will be ~70 μl each and may be loaded in two rounds into the same well if needed. First load 40 μl and apply 200 V until the whole sample enters the gel (~30 s-1 min). Now load the remaining sample. Carefully layer the sample at the very bottom of the well using a gel-loading tip; residual glycerol in the well from the first loading will prevent the subsequent sample from "falling" into the well, so layering from the bottom of the well is crucial during the second round of sample loading. 5. Electrophorese until all the loading dye reaches the 6 mm mark and stain (see Figure 4C), as per Step B3b (above).
6. Excise gel plugs, and process as per Steps B3c to B3f (above). i.e., 1/(1 + 1). However, our data exhibit some variability, as shown in Figure 3A. Therefore, we normalized the data such that the peaks were re-centered at 0.5. From this set, 0.5 was subtracted from the data (centering insensitive proteins at the origin, and completely sensitive proteins at 0.5), followed by multiplication by two to expand the data to cover the range from 0 (insensitive) to 1 (completely sensitive); these latter two transformations are encompassed by the functions: where, a = 2; b = -0.5.
b. To determine if these data were normally distributed the distances from the (0, 0) point to protein coordinates were calculated. Proteins with distance less than two median distances were selected. The Shapiro-Wilk normality test (the null-hypothesis of this test is that the population is normally distributed) was applied for the distances (P-value = 0.29). The distribution of the distances was plotted as a histogram displaying the frequency (y-axis) 24 www.bio-protocol.org/e3200 versus RNase sensitivity (x-axis) of a simulation of normally distributed data (shown in black) and the actual data shown in blue; a Q-Q plot was also drawn-both displayed in Figure 3B.
c. The mean value and standard deviation were calculated using the distribution of distances from the origin. The distance threshold for P-value = 0.001 was calculated using the R programming language. A circle with radius equal to the threshold was plotted and points with distances higher than the threshold were marked as black ( Figure 2D). d. The relevant code for the above post-processing can be obtained here: https://bitbucket.org/altukhov/line-1/src/master/src/protein_analysis/RNAse.R The prior I-DIRT analysis (Taylor et al., 2013) is an important aspect of the presented analysis.
SILAC as conducted here-so-called mix after purification or MAP-SILAC (Wang and Huang, 2008)does not reveal any information about which interactions are specific or non-specific; it only reveals what has changed between the two treatments (with RNases or BSA). We used I-DIRT, first, in order to establish which interactions are likely to have formed in vivo, versus which are likely to be post-extraction in vitro artifacts (Tackett et al., 2005). Importantly, the I-DIRT data was collected using comparable experimental conditions (i.e., protein extraction and affinity capture parameters) as for the SILAC data. By using the I-DIRT analysis as a specificity filter on the SILAC analysis, second, we were able to focus our attention on the differential behaviors of proteins most likely to be bona fide physical interactors within L1 RNPs (colored nodes, Figure 4D). Nevertheless, the proteins that were RNase sensitive and not I-DIRT-specific (black nodes, Figure 4D) proved informative: they are all RNA-binding proteins that one might expect to be released by RNase treatment. Thus, they served as a secondary indicator that the treatment is effective. Notably, PABPC1 and C4-poly(A) RNA binding proteins appear insensitive to RNase treatment. This is most likely due to the fact that neither RNase A nor T1 cleave RNA at adenosine residues (Volkin and Cohn, 1953;Yoshida, 2001); hence poly(A) binding proteins may not be ready targets for release 25 www.bio-protocol.org/e3200 from direct RNA binding by the assay implemented here (or generally, using these ribonucleases) -see further discussion in (Taylor et al., 2018). On the basis of these data, we concluded that UPF1, ZCCHC3, MOV10, and ORF1p are at least partly dependent upon the presence of intact RNA to remain stably associated within L1 RNPs. Heavy-untreated (RNase light -with-BSA heavy ). (5) BSA-200 ng. (6) Light-untreated with Heavytreated (BSA light -with-RNase heavy ). D. Processed data displaying the degree of RNase-sensitivity (i.e., differential affinity immobilization) exhibited by constituents of the SILAC-labeled fractions.
Proteins requiring intact RNA to maintain stable interactions with immobilized ORF2p were released from the RNase-treated medium, while the BSA-treated sample controlled for the spontaneous release of proteins from the medium. Results from the assay have been graphed as the fraction of each detected protein present in the BSA-treated sample (RNase-sensitive proteins are more present in the BSA treated sample), normalized as described in step 2 (Data analysis). A cut-off of P = 10 -3 for RNase-sensitivity is indicated by a light gray circle; proteins that are RNase-sensitive with a statistical significance of P < 10 -3 are outside the circle. Proteins previously ranked significant by I-DIRT analysis (Taylor et al., 2013) are labeled and displayed in blue or magenta (as indicated); black nodes were RNase-sensitive but not significant by I-DIRT; gray, unlabeled nodes were neither RNase-sensitive nor significant by I-DIRT.

Recipes
A. Cell culture media