Cell culture

U2OS, HeLa, MCF7, IMR90-tert and HeLa Flp-In cell lines [20] were grown in DMEM (Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (Thermo Fisher Scientific), L-glutamine, and penicillin/streptomycin. For THP-1, RPMI (Thermo Fisher Scientific) supplemented with 10% fetal bovine serum and penicillin/streptomycin was used. All cell lines were grown at 37°C. When required, cells were synchronized using 100ng/ml nocodazole (Sigma-Aldrich M1404) for 14 hours. These cells were subsequently released in fresh medium after washing twice with phosphate buffer saline (PBS) (Thermo Fisher Scientific), and harvested at specific stages of mitosis.

Cloning and mutagenesis

Complementary DNA (cDNA) construct having full-length MLL was cloned into pBabe-FLAG vector as described [21]. pcDNA 3.1-puro-GFP was made by inserting polymerase chain reaction (PCR) amplified GFP into KpnI (New England Biolabs) linearized pcDNA 3.1-puro vector. Similarly, to generate pcDNA 3.1-puro-mCherry, PCR amplified mCherry was cloned into EcoRI (New England Biolabs) linearized pcDNA 3.1-puro vector. GFP-tagged pcDNA5/FRT and pcDNA5/FRT TO vector, having hygromycin as selectable marker, was generated by PCR sub-cloning of GFP into EcoRI and EcoRV (New England Biolabs) site of pcDNA5/FRT, Xho-1 site of pcDNA5/FRT TO vector respectively. MLL_N-GFP, MLL_C-GFP, and MLL_N-mCherry, constructs were made by PCR cloning MLL_N or MLL_C into XhoI (New England Biolabs) linearized pcDNA3.1-GFP, or pcDNA3.1-mCherry vectors, high fidelity Phusion polymerase (New England Biolabs) was used for PCR amplification. To generate GFP-tagged MLL1400 and MLL fusion proteins (MLL-AF4 and MLL-AF9) constructs, inserts were amplified similarly using cDNA encoding MLL and full length MLL-AF4 and MLL-AF9 as a template (MLL-AF4 and MLL-AF9 cDNAs were a kind gift from J. Hess), and cloned into BamHI linearized pcDNA/FRT TO GFP vector. All the PCR cloning was done using In-Fusion HD Cloning Kit (Takara) as per the manufacturer’s instructions. pcDNA-MLL_N-mCherry and pcDNAFRT-MLL_C-GFP were subjected to inverse PCR using KOD FX high fidelity enzyme (Toyobo KOD-201) to generate MLL_N-∆FYRN (∆1989–2102 amino acid) and MLLC-∆ FYRC (∆3671–3755 amino acid) constructs [13]. All constructs were verified by sequencing the entire insert.

Transfection and stable cell-line generation

pBabe FLAG-MLL cell line has been described before [21]. MLL_N-GFP, MLL_N-mCherry, MLL_C-GFP, MLL_N∆FYRN-mCherry, FRT-MLL_C∆FYRC-GFP, MLL_N1400-GFP, MLL_N-mCherry + MLL_C-GFP, MLL_N∆FYRN-mCherry + FRT-MLL_C∆FYRC-GFP, MLL fusion proteins FRT-MLL-AF4 and FRT-MLL-AF9 cell lines were generated by transfecting above constructs using Lipofectamine-2000 (Thermo Fisher Scientific) into U2OS or HeLa Flp-In cells. The colonies positive for recombinant protein expression were selected using puromycin (4 µg/ml) or Hygromycin (20µg/ml).

siRNA transfections

Human MLL siRNA #1 (against the ORF) and control Luciferase siRNA transfections were performed as described [21]. Cells were harvested after 72 hours after the first transfection and subjected to Western Blot or IFS.

Immunofluorescence staining

To detect the endogenous MLL, cells were fixed with pre-chilled acetone for 90 seconds for centrosome, and 4 minutes for nucleolus; for spindle and midbody detection, pre-chilled methanol was used for 3 minutes at −20°C for fixation. For the detection of ectopically expressed MLL, MLL mutants, and MLL fusion proteins the same fixation conditions were used as endogenous protein, except for the nucleolus where all ectopically expressed MLL cell lines and MLL fusion protein expressing cells were fixed with 4% paraformaldehyde for 10 minutes followed by permeabilization using 0.2% triton-X-100 for 5 minutes.

Fixed cells were then blocked with 1% bovine serum albumin (BSA) at room temperature (RT) for one hour. The cells were subsequently incubated with primary antibodies against MLL_N (1:100, Bethyl A300-086A), MLL_C (1:100, Bethyl A300-374A), FLAG (1:100, Sigma-Aldrich F7425), α-Tubulin (1:250, Sigma T9026), γ-Tubulin(1:250, Sigma T6557), GFP(1:100, Thermo Fisher Scientific A-11122), mCherry (1:100, Abcam ab125096), Pericentrin (1:250, Abcam ab4448), Centrin1 (1: 100, Abcam ab11257), B23 (1:200 Sigma-Aldrich BO556), UBF(1:100 Santa Cruz Biotechnology sc-13,125) in PBS with 0.2% triton-X-100 at RT for 3 hours or at 4°C overnight. Following primary antibody incubation, the cells were washed and incubated with the Alexa 488 (1:1000, Thermo Fisher Scientific A11034), Alexa 633 (1:500, Thermo Fisher Scientific A21050) and/or Alexa 594 (1:500, Thermo Fisher Scientific A11030) conjugated anti-rabbit or anti-mouse secondary antibodies at RT for 1 hour. Nuclei were counterstained with 4', 6-diamidino-2-phenylindole (DAPI) (1:5000, Sigma D9542) and cells were mounted using mounting medium containing anti-fade (Phenylenediamine, Sigma 695106). Images were captured on either Zeiss LSM 510 META or Zeiss LSM 700 inverted confocal microscope and analyzed with LSM or ZEN software respectively.

Live-cell imaging

For time lapse imaging, U2OS cells stably expressing GFP tagged MLL_N were grown in 35 mm glass bottom dishes (Arrow labs). The cells were imaged at 60x magnification on NIKON ECLIPSE Ti-E inverted microscope by acquiring 1 frame per minute. The metaphase and subsequent stages were imaged to observe centrosome, spindle, and midbody localizations of MLL_N-GFP. To detect the nucleolar localization of MLL_N-GFP, the cells were transfected with B23 DsRed (to mark the nucleolar region). Subsequently the cells co-expressing MLL_N and B23 proteins were imaged in interphase. The movies were subjected to Gaussian convolution (kernel size 5) followed by Gauss-Laplace sharpening using NIS elements software.

Fluorescence recovery after photo bleaching (FRAP)

The FRAP was performed using Zeiss LSM 700 inverted confocal microscope. U2OS cells stably expressing MLL_N-GFP and GFP alone were bleached at selected region by 488 nm laser with 20 iterations with 100% laser power. The images before and after bleaching were taken at 2 sec intervals. Fluorescent recovery was plotted as relative intensity Vs. time interval by normalizing the pre-bleach intensity as 1 [22].

Localization of MLL subunits in mitosis

We have previously reported that MLL localizes to spindle microtubules during mitosis [17]. To investigate further, we checked the localization of MLL in live cells in mitosis. We imaged MLL_N-GFP-expressing cells and observed MLL on condensed chromatin as reported before [17,23–26], on centrosome and spindle during metaphase, on central spindle during anaphase, and on midbody during telophase and cytokinesis (Figure 1(a), Movie S1).

Novel sub-cellular localizations and intra-molecular interactions may define new functions of Mixed Lineage Leukemia protein

All authors

Amit Mahendra Karole, Swathi Chodisetty, Aamir Ali, Nidhi Kumari & Shweta Tyagi http://orcid.org/0000-0002-8349-6843

https://doi.org/10.1080/15384101.2018.1553338

Published online:

10 December 2018

Figure 1. MLL localizes to spindle, centrosome and midbody in mitosis. Representative fluorescent time-lapse images of U2OS cells stably expressing MLL_N-GFP are shown (See Movie S1 in Supplemental Data). Arrow shows the localization of MLL_N on subcellular organelles at different time frames, centrosome (frame a, b) spindle (frame b), central spindle (frame c) and midbody (frame d-f). Time in MM: SS format is shown. (B-D) IFS of endogenous MLL_N (panel a) and MLL_C (panel b, c) in mitosis is shown. The U2OS cells were stained with antibodies against MLL_N or MLL_C and α-tubulin to mark spindle (B) and midbody (D), or γ-tubulin to mark centrosome (C). (E) IFS of endogenous MLL_N (panel a) and MLL_C (panel b-c) in interphase is shown. The U2OS cells were stained with anti-MLL and anti-γ-tubulin antibody. (B-E) DNA was stained with 4, 6-diamidino-2-phenylindole (DAPI). The cells were treated with control (panel b) or MLL siRNA (panel c) to check for antibody specificity. The plot profile analysis (shown on left) of area indicated with white arrow, depicts co-localization profile of MLL (green line) with the respective tubulin marker (red line). Scale bar, 5μm. See inset for zoomed-in images.

Display full size

In order to validate these observations, we performed immunofluorescence staining by using antibodies against both – N and C subunits of MLL (Figure S1A). Staining for both subunits of MLL will either reveal their digressions in localization or confirm specificity of binding. Consistent with our previous findings, we could detect both subunits of MLL on the spindle apparatus (Figure 1(b)). Like before [17] we tried different fixation conditions and detected MLL on centrosomes in mitosis with both MLL_N and MLL_C antibodies (Figure 1(c)). The midbody staining of MLL was also clearly observed (Figure 1(d)) and MLL stain was specifically diminished upon treatment with MLL siRNA in both cases (Figure 1(b,d), Figure S1B).

Out of all the mitotic structures reported here, centrosome is present in interphase cells as well. Acetone extracted majority of the MLL from the nucleus, and centrosome staining of MLL was clearly visible (Figure 1(e)). In addition, we also observed peri-nuclear staining of MLL. However, only centrosome staining diminished upon MLL RNAi (Figure 1(e)).

We reproduced the spindle, centrosome and midbody association of MLL in transformed (HeLa and MCF7) as well as non-transformed (IMR90-tert) cell lines indicating that MLL localization to mitotic structures was not limited to one cell type (Figure S2A-C). We also used stable cell-lines expressing FLAG-epitope-tagged MLL [27] and detected FLAG-MLL on the mitotic structures (Figure S2D). Next we verified that MLL also localized to centrosome in interphase in different cells and showed staining with anti-FLAG antibody like the endogenous protein (Figure S2E, F) further confirming that our observed pattern of MLL staining was reproducible and specific.

To check if the two subunits dissociated during mitosis, we made fluorescent protein fusion constructs of MLL where MLL_N was fused with mCherry and MLL_C with GFP (Figure S1B and S3). When tested with specific markers, both subunits were able to localize to the mitotic structures, and centrosome in interphase (Figure 2(a–d)). The low expression from the recombinant proteins raised the possibility of these forming heterodimers with their respective endogenous counterpart, i.e., MLL_N-mCherry may interact with endogenous MLL_C and vice versa. Therefore, we designed MLL_NΔFYRN-mCherry and MLL_CΔFYRC-GFP; mutants which are deficient in hetero-dimerization [13](Figure S1B, Figure 2(e–h)).

Novel sub-cellular localizations and intra-molecular interactions may define new functions of Mixed Lineage Leukemia protein

All authors

Amit Mahendra Karole, Swathi Chodisetty, Aamir Ali, Nidhi Kumari & Shweta Tyagi http://orcid.org/0000-0002-8349-6843

https://doi.org/10.1080/15384101.2018.1553338

Published online:

10 December 2018

Figure 2. MLL_N and MLL_C localize independently to mitotic structures. (A-D) IFS of U2OS interphase (A-B) or mitotic (C-D) cells stably expressing MLL_N-mCherry (A, C) or MLL_C-GFP (B, D), stained with anti-mCherry or anti-GFP antibody are shown. The cells were co-stained with anti-α-tubulin antibody to mark spindle (C-D, panel a) and midbody (C-D, panel c); either anti-Pericentrin (A, C panel b) or anti-γ-tubulin antibody (B, D panel b) to mark the centrosome. (E-H) IFS of U2OS interphase (E-F) or mitotic (G-H) cells stably expressing MLL_NΔFYRN-mCherry (E, G) or MLL_C-ΔFYRC-GFP (F, H) are shown. The cells were stained with anti-mCherry (E, G) or anti-GFP (F, H), anti-α-tubulin antibody (G, H panel a, c); either anti-Pericentrin (E, G panel b) or anti-γ-tubulin antibody (F, H panel b). (A-H) The plot profile analysis shows co-localization of recombinant MLL (red line for mCherry and green line for GFP-tagged subunit) with the respective marker. Scale bar, 5μm. mCh, mCherry.

Display full size

Even though MLL_NΔFYRN-mCherry was more stable than MLL_N-mCherry, both proteins showed localization comparable to that of endogenous MLL_N (Figure 2(e,g)). Similarly, MLL_C ΔFYRC-GFP showed localization akin to MLL_C-GFP (Figure 2(f,h)). Although both subunits localized to mitotic structures, we could detect subtle differences in their localizations (for instance compare the localization of MLL_N-mCherry and MLL_C-GFP on centrosome in Figure S3C panel b). Co-expression of both subunits together did not alter the localization of individual subunit (compare Figure S3A-D to Figure 2(a–d)). These observations were not made in cells expressing GFP or mCherry alone (Figure S3E-H). To sum up, our results show that both subunits of MLL dynamically localize to different mitotic structures independent of each other.

MLL localizes to nucleolus

When immunostained with the N and C antibodies, majority of MLL protein localized to the nucleus (Figure S4A). Previously it has been observed that MLL localized to the nucleolus [19,24,28]. We also observed that MLL_N and MLL_C displayed nucleolar reactivity when they co-localized with B23, a nucleolar marker protein (Figure 3(a)). Upon MLL siRNA-treatment, the nucleolar staining of MLL was reduced (Figure 3(a) panel c-d). We could detect FLAG-MLL in the nucleolus (Figure S4B) and, in multiple cell lines endogenous MLL could be seen co-localizing with B23, indicating that it was nucleolar (Figure S4C).

Novel sub-cellular localizations and intra-molecular interactions may define new functions of Mixed Lineage Leukemia protein

All authors

Amit Mahendra Karole, Swathi Chodisetty, Aamir Ali, Nidhi Kumari & Shweta Tyagi http://orcid.org/0000-0002-8349-6843

https://doi.org/10.1080/15384101.2018.1553338

Published online:

10 December 2018

Figure 3. Endogenous MLL localizes to the nucleolus dynamically. (A) anti-MLL_N (panel a, b) or anti-MLL_C (panel c, d) antibody staining in U2OS cells is shown. B23 was used to mark nucleolus. The cells were treated with control (panel c) or MLL siRNA (panel d) to check for antibody specificity. Panel b is zoomed-in image of a selected cell from panel a. Scale bar is equal to 10μm (panel a) or 5μm (panel b-d). (B) Fluorescent time-lapse images of U2OS cells stably expressing MLL_N-GFP and transiently transfected with B23-DsRed are shown. Arrows indicate the nucleolar region. Time in MM: SS format is shown. (See Movie S2 in Supplemental Data). (C) Fluorescence recovery after photo bleaching (FRAP) experiment was performed on MLL_N-GFP cells to determine dynamic localization of MLL in the nucleolus. Representative images are shown. Scale bar, 5μm. (D) FRAP recovery curves of MLL_N-GFP in interphase cells. Graph shows relative intensity Vs. time interval. Pre-bleach intensity is normalized to 1. Error bars show standard deviation of the experimental data (n = 4). (Also see Figure S4). (E) U2OS cells stably expressing MLL_N-mCherry (panel a), MLL_N ∆ FYRN-mCherry (panel b), MLL_C-GFP (panel c), and MLL_C Δ FYRC-GFP (panel d) stained with anti-mCherry or GFP antibodies are shown. (F) U2OS cells stably co-expressing MLL_N-mCherry and MLL_C-GFP (panel a), and MLL_N∆FYRN-mCherry and MLL_CΔFYRC-GFP (panel b) are shown. Cells were stained with anti-GFP to check for localization of C subunit. (A, E-F) The plot on left shows co-localization profile of MLL with B23. mCh, mCherry. Scale bar, 10μm.

Display full size

Over-expression and/or fixation can influence the results of protein-localization experiments [29]. However, in living cells also, MLL_N-GFP accumulated in nucleolar regions (Figure 3(d), Movie S2). Next, we used fluorescence recovery after photo bleaching (FRAP) approach. Here, we used the MLL_N-GFP stable cell lines to photo bleach a nucleolar region and recovery of green fluorescence was measured as a function of time [22,30]. Fluorescence recovery of MLL_N showed a rapid increase in fluorescence followed by a much slower recovery (Figure 3(c,d)), in direct contrast to the GFP alone expressing cells (Figure S4D-E). Taken together, our experiments show that MLL is nuclear and nucleolar in interphase cells.

MLL_N, but not MLL_C, is nucleolar

It has been reported earlier that MLL_N translocate MLL_C to the nucleus [12,16]. To see if the same was true with the nucleolar localization, we made use of cell lines described above. MLL_N-mCherry and MLL_N ΔFYRN-mCherry like MLL_N-GFP showed nucleolar localization (Figure 3(e) panel a-b). In direct contrast to MLL_N, majority of cells displayed very little nuclear accumulation of MLL_C-GFP and it was absent in the nucleolus (Figure 3(e) panel c). MLL_C ΔFYRC-GFP, showed better expression, hence could be seen inside the nucleus but not the nucleolus.

To test the effect of MLL_N expression on MLL_C protein, we made use of stable cell lines co-expressing (i) MLL_N-mCherry with MLL_C-GFP and (ii) MLL_N ΔFYRN-mCherry with MLL_C ΔFYRC-GFP and checked for localization of the GFP protein. We observed that while MLL_C-GFP now displayed nucleolar localization, MLL_C ΔFYRC-GFP was still incapable of translocating to the nucleolus (Figure 3(f) compare panel a to b). These observations indicate that MLL_C requires hetero-dimerization with MLL_N for its nucleolar localization.

Localization of MLL fusion proteins

How do our observations here affect our knowledge of MLL-fusion proteins? To answer this question, we started by studying the localization of the first 1400 amino acids fragment of MLL_N (MLL1400), as these are common to most direct MLL-fusions [31,32]. We observed that MLL1400-GFP was able to go to nucleolus and centrosome in interphase (Figure 4(a,b)) and spindle, centrosome and midbody in mitosis (Figure 4(c)).

Novel sub-cellular localizations and intra-molecular interactions may define new functions of Mixed Lineage Leukemia protein

All authors

Amit Mahendra Karole, Swathi Chodisetty, Aamir Ali, Nidhi Kumari & Shweta Tyagi http://orcid.org/0000-0002-8349-6843

https://doi.org/10.1080/15384101.2018.1553338

Published online:

10 December 2018

Figure 4. MLL1400 and MLL fusion proteins localizes to the nucleolus, centrosome, spindle and midbody. (A-B) IFS of U2OS cells stably expressing MLL1400-GFP is shown. MLL1400-GFP, stained by anti-GFP antibody localizes to the nucleolus (A) and centrosome (B) as shown. (C) Cells stably expressing MLL1400-GFP were stained with anti-GFP antibody (panel a-c) and anti-α-tubulin (panel a, c) or anti -γ tubulin antibody (panel b). (D) HeLa Flp-In cells expressing MLL-AF4-GFP or MLL-AF9-GFP were co-stained with anti-GFP and anti-UBF antibody. (E, F) MLL fusion protein (MLL-FP) localize to centrosome, spindle, midbody as shown. Cells expressing MLL-AF4-GFP and MLL-AF9-GFP were fixed and stained with anti–GFP and anti-γ-tubulin or anti-α-tubulin antibodies as shown. (G) The cells transfected with MLL fusion proteins were subjected to analysis for mitotic defects. The percentage of the cells showing mitotic defects were quantified in control GFP and MLL-FP transfected cells. The data represents as mean ± SD. Significant P-values (*<0.02) were obtained with Student t-test. (H) IFS of leukemic cell line THP-1 (with MLL-AF9 fusion) stained with anti-MLL_N antibody. THP-1 cells were fixed and co-stained with anti-MLL_N (green) and anti-UBF (red) or Centrin1 (red) antibodies. The anti-MLL_N antibody detects both wild type MLL_N and MLL-AF9 fusion protein (see Figure S1A). UBF and Centrin1 were used to mark nucleoli and centrosome respectively. (A–G) The plot profile analysis depicts co-localization profile of MLL1400-GFP or MLL-FP-GFP (green line) with the respective marker. Scale bar, 5μm.

Display full size

MLL-AF4 and MLL-AF9 fusions are among the most prevalent translocations found in patients [1,31]. Therefore, we cloned and expressed GFP-tagged version of MLL-AF4 and MLL-AF9 and observed both MLL-fusion proteins co-localizing with Upstream binding factor (UBF) in the nucleolus (Figure 4(d)). Interestingly UBF is the key transcription factor associated with RNA polymerase I (RNA Pol I) and is essential for rRNA transcription [33]. The Nucleolar Localization Sequence Detector, NoD [34], predicted 10 nucleolar localization signal (NoLS) in whole of MLL_N, out of which seven were concentrated in MLL1400 (Figure S5A) indicating that the N terminal of MLL_N is most likely targeting the MLL heterodimer and MLL-fusion proteins to the nucleolus. To gain insight into the possible role of MLL in the nucleolus, we treated cells with low dose of actinomycin D (Figure S5B). At this dose, actinomycin D is known to inhibit rRNA synthesis and redistribute major components of RNA Pol I machinery including UBF [35]. Interestingly, MLL_N-GFP as well as both the MLL-FPs showed a redistribution pattern similar to UBF upon actinomycin D treatment indicating the possibility of a transcriptional role of MLL in rRNA synthesis and involvement of MLL-FP in the same process, when present.

We next observed weak localization of MLL-FPs on the centrosome in interphase cells (Figure 4(e)). The MLL-FPs displayed sporadic staining on the spindle apparatus (Figure 4(f), panel a-b). The centrosome and midbody were also stained positive for both MLL-FPs (Figure 4(f), panel c-f). Does the illicit localization of MLL-FP affect that of endogenous MLL protein? Dual staining of cells with GFP and MLL_C antibody revealed that both chimeric and endogenous MLL could be detected in nucleolus and centrosome (Figure S5C-E). The spindle staining was weak for both the MLL-FPs and MLL_C subunit (Figure S5E panel a-b), however both proteins could be clearly seen on the centrosome (Figure S5E panel c-d) and midbody (Figure S5E panel e-f). We also observed several anomalies in mitosis like misaligned chromosomes reminiscent of loss of MLL [17]. Consistently, we observed more number of cells with micronuclei and binucleation in MLL-FP expressing cells compared to vector control (Figure 4(g)).

In order to validate our findings here, we used patient-derived leukemic cell line THP-1 which harbors the MLL-AF9 fusion. We stained these cells with MLL_N antibody, which is capable of detecting both wild-type MLL and MLL-AF9 fusion protein (see Figure S1A). We could detect MLL/MLL-AF9 at the nucleolus and centrosome (Figure 4(h)). We further probed these cells with MLL_C antibody, which only recognizes wild-type MLL but not MLL-AF9 (Figure S5F). Our results show that the wild type MLL_C was present at both the nucleolus and centrosome confirming our previous findings.

Novel localization of MLL in mitosis

MLL is predominantly a nuclear protein although its cytoplasmic localization has been reported previously [12,16–19,24,25,36]. MLL has also been shown to co-localize with the condensed chromosomes during mitosis [17,23–26]. Recently we reported the spindle localization of MLL during mitosis [17]. During this study, we observed that MLL may localizes to the centrosome as well. Here we have confirmed by different IFS and siRNA experiments that not only does MLL specifically localize to spindle and centrosome during mitosis but also to midbody during cytokinesis. Both the subunits of MLL are able to localize to these structures independent of each other, indicating that they may each have the capability to get targeted to the microtubule-rich structures. Interestingly, the Drosophila trithorax, ortholog of mammalian MLL, was identified as microtubule-associated proteins (MAPs) in large scale proteomic analysis [37]. Other members of MLL complex like WDR5 and RbBP5 have been reported on microtubule-rich structures before [17,37–39] and a large number of mitotic proteins associate with them [17,38,40]. The localizations reported here indicate towards potential novel undiscovered roles of MLL in mitosis. Consistent with these indications, we reported several mitotic defects upon loss of MLL by RNAi like prolonged pro-metaphase and metaphase, defects in chromosome alignment and formation of spindle apparatus, formation of micronuclei, defective cytokinesis [17,21].

Role of MLL in nucleolus

We also observed MLL in the nucleolus. Interestingly, nucleolar localizations of MLL has been reported before [19,24,28]. What would be the role of MLL here? The nucleolus is the site of rRNA biogenesis but in addition to this, it also involved in many other cellular processes, which are different from its canonical functions. Nucleolar proteome analysis has shown that about 30% of the nucleolar proteins cannot be attributed to conventional nucleolar functions, which emphasizes the multifunctional role of the nucleolus in cellular homeostasis [41]. Non-nucleolar, as well as nucleolar proteins, have been reported to shuttle between nucleoplasm and nucleolus during various physiological changes [42]. Some proteins get sequestered at the nucleolus during various physiological stresses, for example non-nucleolar proteins like, VHL and HSP 70 get sequestered at nucleolus during acidosis and heat shock respectively [43,44]. Others are known to localize to the nucleolus in a cell cycle-dependent manner [45,46]. Nevertheless, given the role of MLL in transcription, its co-localization with UBF and redistribution pattern similar to UBF upon actinomycin D treatment, it is conceivable that MLL is involved in rDNA transcription. In support of this hypothesis, transcriptionally active rDNA locus has been shown to bear H3K4 tri-methylation marks [47,48]. Recently it was shown that AF4 (MLL fusion partner) interacted with SL1 complex of RNA Pol I transcription machinery and interaction of SL1 with AF4 was needed for the RNA polymerase II (RNA Pol II) mediated transcription initiation [49]. Interestingly, the absence of DNA methyltransferase from the cells can lead to RNA Pol II mediated cryptic transcription from rDNA locus [50]. Whether MLL co-activates RNA Pol I or II inside the nucleolus is a subject of further experimentation. Our results reveal that MLL_C is dependent on MLL_N for its nucleolar localization. Further analysis uncovered the presence of several putative NoLS in N-terminal of MLL_N. Hence, in addition to the chromatin targeting, heterodimerization with MLL_N provides MLL_C capability of nucleolar translocation as well.

MLL fusion proteins

Illegitimate chromosomal rearrangement of mll gene is common in acute lymphoblastic and myeloid leukemia cases. Recent reports have characterized about 80 direct and 120 reciprocal MLL fusions. The fusion partners of MLL comprise of nuclear proteins but also cytosolic enzyme, membrane proteins, extracellular matrix proteins and cytoskeleton proteins. At the molecular level, MLL translocation partners share no common feature or cellular role making it challenging to develop a single model for MLL-fusion protein based leukemogenesis. Nonetheless, due to the transcriptional role of MLL and the predominant occurrence of 7–10 recombinations of MLL gene found in 90% of patients, some unifying “gain-of–function” models have been proposed. The entire focus of these models has been on fusion proteins involved in transcription. All this while fusions with non-transcriptional proteins has been mostly ignored.

Our results here show that not only MLL but also MLL-fusion proteins are targeted to mitotic structures. The presence of MLL-FPs alone in transfected cells, is enough to cause mitotic defects like binucleation and micronuclei, most likely by disturbing the mitotic progression. Moreover, here we have only examined the most obvious phenotypes of deregulated mitosis. A more careful examination of chromosome instability may reveal the true extent of aneuploidy MLL-FP-expressing cells actually undergo. In transgenic flies expressing MLL-FP, alterations in cell-cycle progression has been reported [51]. It is likely that MLL-FPs act by interfering with the function of endogenous MLL protein. We observed more cytoplasmic MLL_C upon expression of MLL-FPs (Karole and Tyagi, unpublished). However, we could not determine if it was present as heterodimer with MLL_N or devoid of it. Interestingly, endogenous MLL undergoes biphasic degradation, but when alone, cytoplasmic MLL_C is highly prone to degradation [16,26]. This is in contrast to the MLL-FPs, which are resistant to degradation undergone by the endogenous MLL protein, causing deregulation of cell cycle progression [26]. Our results suggest that MLL-FPs may have broader role than transcription which may further undermine cellular processes like cell division.