MYO5A-NTRK3, a motor fused to a kinase

Have motor, will travel. Lu Wang of Memorial Sloan Kettering Cancer Center and a very large group of coworkers published an extremely interesting report of a MYO5A-NTRK3 gene rearrangement of the long arm of chromosome 5 in a childhood melanocytic neoplasm. What makes this fusion so interesting is that the motor protein (myosin 5) that transports melanosomes on actin filaments has had its cargo binding domain replaced with the active kinase domain of TrkC. In this cancer we are dealing with not only loss of regulation of the TrkC kinase but also the cargo regulated transport activity of myosin 5A.

Looking for new genomic alterations

The Wang study was a retrospective analysis in a search of non conventional melanoma genomic alterations in Spitzoid neoplasms. Two unusual features listed in the Wang (2017) introduction were the pediatric occurrence and the lack of genomic alternations in conventional oncogenes.

  • 207 FFPE preserved pediatric melanocytic neoplasms with ambiguous features were submitted for DNA copy number analysis as part of a search for non conventional melanoma genomic alterations.
  • Four tumors had evidence of a TrkC fusion.
  • Three of these children were females ( 1, 7, and 15 years old). The fourth child was an 8 year old male.
  • This blog is about tumor #930 that came from a child’s foot. The other FFPE preserved tumors were of insufficient quality for sequencing.

Tumor 930 mRNA was sequenced allowing for determination of the fusion partner involved. Unlike TrkC, myosin 5A is expressed in comparatively large amounts in the skin.

Graph showing Myosin 5A expression in skin cells as compared to no TrkC expression in skin cells

Figure 1.  Protein expression of wild type fusion partners from Red arrows point to the skin

Both MYO5A and NTRK3 are on the reverse stand of chromosome 15, Fig. 2.

Cartoon showing location of MYO5A and NTRK2 genes on chromosome 6 and 9

Figure 2. Location of the MYO5A and NTRK2 genes on the q arm of chromosome 15.

The authors also looked at changes in copy number. In the tumor 930 fusion, it is not a simple matter of the region of chromosome 15, between the MYO5A and NTRK3 genes, missing. Something far more complicated is going on that is addressed in the supplemental material of Wang 2017. We may address this complication in a separate blog.

Graph showing copy number variations around NTRK3 gene in a tumor

Figure 3. Copy number variations around the NTRK3 gene in tumor #930. From Wang 2017.

The protein part of the story

Another interesting story is in the protein domains of the two fusion partners and where the fusion occurs in tumor 930. In this particular fusion we see all of the motor domain of myosin (Fig 4A), the transmembrane helix of TrkC, as well as the kinase domain of TrkC (Fig. 4B)

Diagram showing domain maps of Myo 5A fusion partners

Figure 4. Domain maps of fusion partners. Arrows point to the break points. A. The fusion contains the N-terminus of the Myo 5A protein up the the point of the arrow. B. The fusion contains the C-terminus of TrkC beginning at the arrow.

The first TrkC (Q16288) amino acid in the fusion is Asp411. The authors list a valine as a the first Trk in the fusion. The last myosin 5A in the fusion is Glu1444.

Myo 5A- TrkC  running wild in a  in a tumor?

Kathleen Trybus of the University of Vermont has a really nice 2008 review on myosin 5. The lever arm is the part that moves when ATP is hydrolyzed to ADP + Pi. Calmodulin (CaM) binds to the IQ motifs in the lever arm. The rod region connects the action part of myosin 5 to the cargo carrying domains in the C-terminal globular tail (Fig. 5A). Unlike in the normal cell (Fig. 5B) there is no cargo carrying ensemble in the fusion. In the fusion we have two juxtaposed kinase domains of TrkC plus the hydrophobic transmembrane region (Fig. 5C).

cartoon showing domains of myosin 5A homodimer

Figure 5. The myosin 5A dimer, adapted from Trybus 2008 A. The domains of he myosin 5A homodimer. Note the rod region and calmodulin (CaM) binding domains on the lever arm.  B. The motor domains of myosin 5A bind to filaments of actin. The C-termini of the dimer binds to a cargo ensemble. C. In the fusion protein, kinase domains of TrkC replace the melanosome cargo.

Conventional myosins are associated with skeletal, cardiac, vascular, and enteric muscle contraction. This involves movement of myosin filaments on actin filaments. The unconventional myosin 5A is capable of what is called processive movement on actin filaments (Fig. 6). This processive walking requires the hydrolysis of ATP, just like muscle contraction. It is controlled by the calcium binding protein calmodulin. Consult the Trybus (2008) review to learn more about calcium and calmodulin regulation of myosin 5A movement.

a cartoon showing hypothetical movement of Myo 5A-TrkC on actin filaments.

Figure 6. Hypothetical movement of the Myo 5A-TrkC fusion protein movement on actin filaments. Adapted from Trybus 2008

In her 2008 review, Kathleen Trybus presented cyroelectron microscopy two-dimensional arrays of crgo-free myosin 5A on lipid monolayers. These arrays resembled six petal flowers. Known X-ray crystal structures were fit into these “flowers.” Note that the pollen producing “pistil” art of the flower was missing in the Trybus (2008) review. The important thing to note is the interaction of the cargo free tail domain and the motor domain of myosin 5A.

crystal structure of crgo-free myosin 5A on lipid monolayers.

Figure 7. Auto inhibition of cargo free myosin 5A. From Trybus 2008. A. Cargo free myosin 5A was crystallized into a 2D, flower shaped array on a lipid bilayer. B. Known X-ray crystal structures were fit into these densities.

The missing pistil of the myosin 5 flower was defined by Velvarska and Niessing (2013). Note the role of two lysines in the tail domain (K1708 and K1781) in cargo free auto-inhibition of the motor domain. This region is missing in the fusion.  Would such autoinhibition exist in the Myo 5A-TrkC fusion? In seems unlikely. Velvarska and Niessing have a nice discussion on how fast the cargo free tail associates with the motor domain and how fast it comes off. It is in those coming off intervals that cargo may be loaded.

Crystal structure showing interaction of the C-terminal, cargo free globular domain of myosin 5A with the motor domain along with auto inhibitory interactions that cannot occur in the fusion.

Figure 8. From Velvarska and Niessing 2013.  Refined interaction of the C-terminal, cargo free globular domain of myosin 5A with the motor domain. A. Fitting into the cryoelectron microscope 2D crystal “flower” B. Auto inhibitory Interactions that cannot occur in the fusion.

One may speculate that the Myo 5A- TrkC fusion protein is free to run wild, phosphorylating substrates it encounters on it jaunts on actin filaments.

Important Information

We have not covered the emerging opportunity of Trk inhibitors like entrectinib, in controlling a potentially “running wild”  fusion kinase in this blog.  NTRK fusions are taking on a more important role, because of new clinical trials of investigation agents that can regulate TRK.

There is an open NTRK fusion clinical trial that is actively enrolling any solid tumor patient with NTRK fusions (STARTRK-2). For more information go to the NTRK trial website.


Trybus KM. (2008) Myosin V from head to tail. Cell Mol Life Sci. 2008 May;65(9):1378-89. PubMed

Velvarska H, Niessing D. (2013) Structural insights into the globular tails of the human type v myosins Myo5a, Myo5b, And Myo5c. PLoS One. 2013 Dec 10;8(12):e82065. PubMed

Wang L, Busam KJ, Benayed R, Cimera R, Wang J, Denley R, Rao M, Aryeequaye R, Mullaney K, Cao L, Ladanyi M, Hameed M. (2017) Identification of NTRK3 Fusions in Childhood Melanocytic Neoplasms. J Mol Diagn. 19(3):387-396. PubMed

Leave a Reply