Highlights d KIF3A/KIF3B/KAP is the sole intraflagellar transport motor in mammalian cells d Engineered, inhibitable KIF3A/KIF3B motors rescue wildtype motor function d Acute inhibition of KIF3A/KIF3B blocks anterograde IFT and leads to cilium loss d KIF3A/KIF3B/KAP is essential for ciliogenesis and cilium maintenance in mammals

The trafficking of components within cilia, called intraflagellar transport (IFT), is powered by kinesin-2 and dynein-2 motors. Loss of function in any subunit of the heterotrimeric KIF3A/KIF3B/KAP kinesin-2 motor prevents ciliogenesis in mammalian cells and has hindered an understanding of how kinesin-2 motors function in IFT. We used a chemicalgenetic approach to engineer an inhibitable KIF3A/KIF3B (i3A/i3B) kinesin-2 motor that is capable of rescuing WT motor function in Kif3a/Kif3b double-knockout cells. Inhibitor addition blocks ciliogenesis or, if added to ciliated cells, blocks IFT within two minutes, which leads to a complete loss of primary cilia within six hours. The kinesin-2 family members KIF3A/KIF3C and KIF17 cannot rescue ciliogenesis in Kif3a/Kif3b double-knockout cells nor delay the disassembly of full-formed cilia upon i3A/i3B inhibition.These data suggest that KIF3A/KIF3B/KAP is the sole and essential motor for cilia assembly and function in mammalian cells, indicating a species-specific adaptation of kinesin-2 motors for IFT function. E n g e l k e e t a l . 2 Δ 11 ), 12 (i3A Δ 12 ), or 13 (i3A Δ 13 ) amino acids. In similar fashion, we fused the DmrB domain to the N-terminus of KIF3B truncated by five (i3B Δ 5 ), six (i3B Δ 6 ), or seven (i3B Δ 7) amino acids (Fig. 2b). We compared the ability of each i3A construct to pair with each i3B construct and generate primary cilia in the absence but not in the presence of B/B inhibitor. Fusion of the Δ 12 /i3B Δ 6 motor resulted in ~200-fold luciferase induction (Fig. 3c). Addition of B/B inhibitor resulted in a greater increase in luciferase activity induction only in cells expressing the i3A Δ 12 /i3B Δ 6 motor, reflecting pathway hyper-activation only in cells that lack a functional primary cilium (Fig. 3c).In summary, we find that expression of the i3A Δ 12 and i3B Δ 6 constructs in Kif3a -/-;Kif3b -/cells results in a bona fide inhibitable kinesin-2 motor. The engineered i3A Δ 12 /i3B Δ 6 motor is referred to as i3A/i3B throughout the rest of the manuscript. Inhibition of KIF3A/KIF3B results in the stalling of IFT trains and their exclusion from ciliaThe generation of an inhibitable kinesin-2 motor enables us, for the first time, to directly examine the role of KIF3A/KIF3B/KAP motors during IFT in fully-formed cilia. To investigate this, Kif3a -/-;Kif3b -/cells were transfected with plasmids for expressing the inhibitable i3A/i3 motor together with a fluorescently-tagged subunit of the IFT-B complex (IFT88-mNG; mNeonGreen) as in previous studies [34, 35]. Analysis of kymographs generated from live-cell imaging experiments revealed that IFT88-marked IFT trains moved processively towards the tip of the cilium, paused for variable durations, and then trafficked back towards the base (Fig. 4a), similar to what has been observed previously [34, 35]. Anterograde and retrograde speeds were on the order of 0.7 μ m/s, consistent with previously E n g e l k e e t a l . 5

MLL1 (KMT2a) gene rearrangements underlie the pathogenesis of aggressive MLL-driven acute leukemia. AF9, one of the most common MLL-fusion partners, recruits the histone H3K79 methyltransferase DOT1L to MLL target genes, constitutively activating transcription of pro-leukemic targets. DOT1L has emerged as a therapeutic target in patients with MLL-driven leukemia. However, global DOT1L enzymatic inhibition may lead to off-target toxicities in non-leukemic cells that could decrease the therapeutic index of DOT1L inhibitors. To bypass this problem, we developed a novel approach targeting specific protein-protein interactions (PPIs) that mediate DOT1L recruitment to MLL target genes, and compared the effects of enzymatic and PPIs inhibition on leukemic and non-leukemic hematopoiesis. MLL-AF9 cell lines were engineered to carry mutant DOT1L constructs with a defective AF9 interaction site or lacking enzymatic activity. In cell lines expressing a DOT1L mutant with defective AF9 binding, we observed complete disruption of DOT1L recruitment to critical target genes and inhibition of leukemic cell growth. To evaluate the overall impact of DOT1L loss in non-leukemic hematopoiesis, we first assessed the impact of acute Dot1l inactivation in adult mouse bone marrow. We observed a rapid reduction in myeloid progenitor cell numbers within 7 days, followed by a loss of long-term hematopoietic stem cells. Furthermore, WT and PPI-deficient DOT1L mutants but not an enzymatically inactive DOT1L mutant were able to rescue sustained hematopoiesis. These data show that the AF9-DOT1L interaction is dispensable in non-leukemic hematopoiesis. Our findings support targeting of the MLL-AF9—DOT1L interaction as a promising therapeutic strategy that is selectively toxic to MLL-driven leukemic cells.

No abstract

Leukemias harboring rearrangements of mixed-lineage leukemia gene (MLL1) are associated with poor clinical outcomes, and new therapeutic approaches are needed. Rearrangement of the MLL1 gene generates fusion oncoproteins which drive the high expression of the clustered homeobox (HOX) genes and induce leukemic transformation. Genome-wide histone methylation studies have revealed that the abnormal expression of MLL1 fusion target genes is associated with high levels of histone H3 lysine 79 (H3K79) methylation. Recruitment of DOT1L (disruptor of telomeric silencing 1-like), a unique histone methyltransferase that catalyzes methylation of H3K79, proved to be essential for the transforming activity of multiple MLL fusion proteins. We have mapped the binding site to a short segment of 10 amino acids in DOT1L and shown that DOT1L mutants lacking these residues did not support transformation by MLL-AF9. We hypothesized that by targeting the AF9-DOT1L protein-protein interactions (PPIs), we would selectively kill MLL-AF9 cells without effecting DOT1L role in normal hematopoiesis. Using established DOT1Lf/f MLL-AF9 with reintroduced WT-DOT1L, DOT1L missing 10aa AF9-binding domain (D10), DOT1L with a point mutation in the AF9-binding domain (I867A) and enzymatically inactive DOT1L (RCR), we were able to demonstrate that by disrupting the AF9-DOT1L PPIs, although we can inhibit leukemogenesis similarly to enzymatic inhibition, this interaction is not essential for normal hematopoiesis. Based on our initial studies to map the DOT1L interaction site and in conjunction with utilizing reported NMR structures of the AF9-DOT1L complex, we investigated the nature of the interactions and the minimum length of the peptide. Using different natural and unnatural amino acids, we successfully designed a 7mer peptide with KD of 10 nM and 25 nM against AF9 and ENL, respectively, showing similar potency as the originally identified and validated 10mer peptide. These results lay the groundwork for further optimization of the 7mer peptide towards developing DOT1L peptidomimetics with improved potency and cellular activity, to further validate the PPIs between DOT1L and MLL-fusion proteins as a potential therapeutic target for MLL rearranged leukemia. Citation Format: Sierrah Marie Grigsby, Jennifer Chase, Bridget Waas, Ann Friedman, Lei Du, Aihong Yao, James Ropa, Justin Serio, Andrew Muntean, Ivan Maillard, Haying Sun, Zaneta Nikolovska-Coleska. Towards peptidomimetics to target DOT1L recruitment in MLL-AF9 leukemia [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 1380.