Inter-organelle interactions between the ER and mitotic spindle facilitates Zika protease cleavage of human Kinesin-5 and results in mitotic defects

Abstract

Here we identify human Kinesin-5, Kif11/HsEg5, as a cellular target of Zika protease. We show that Zika NS2B-NS3 protease targets several sites within the motor domain of HsEg5 irrespective of motor binding to microtubules. The native integral ER-membrane protease triggers mitotic spindle positioning defects and a prolonged metaphase delay in cultured cells. Our data support a model whereby loss of function of HsEg5 is mediated by Zika protease and is spatially restricted to the ER-mitotic spindle interface during mitosis. The resulting phenotype is distinct from the monopolar phenotype that typically results from uniform inhibition of HsEg5 by RNAi or drugs. In addition, our data reveal novel inter-organelle interactions between the mitotic apparatus and the surrounding reticulate ER network. Given that Kif11 is haplo-insufficient in humans, and reduced dosage results in microcephaly, we propose that Zika protease targeting of HsEg5 may be a key event in the etiology of Zika syndrome microcephaly.

Keywords: Cytoskeleton; Developmental Genetics; Virology.

Graphical abstract

Figure 1

List of potential ZPrC cleavage sites and corresponding secondary structure elements within HsEg5 ranked in order of decreasing likelihood of cleavage Mass spectrometry analysis confirms in vitro cleavage by ZPrC of the sites marked in red. AMPPNP is shown in the active site in stick view, whereas the primary landmark MT-binding element of the motor domain, alpha helix4, is shown in cyan.

Figure 2

In vitro hydrolysis of HsEg5_1-367 motor domain by s-ZPrC Recombinant soluble s-ZPrC (27 kDa) was incubated with purified HsEg5 motor domain (42 kDa) for 24 h. s-ZPrC exhibited a low level of self-cleavage after 24 h marked primarily by a 23 kDa fragment. After 24 h, HsEg5 was cut into 21.7 kDa N-terminal (red arrow) and 20.4 kDa C-terminal (cyan arrow) fragments. In turn, the 20.4 kDa fragment was incompletely cut into 9.7 kDa and 10.7 kDa fragments (asterisk). The 20.4 kDa fragment was excised for sequencing to identify the primary s-ZPrC cleavage site (Figure 3).

Figure 3

Kinetic parameters of the hydrolysis of p-nitroanilide-labeled peptides by s-ZPrC (A) Two labeled peptides based on native ZPrC polyprotein cleavage sites were challenged for proteolysis by s-ZPrC: Ac-FAAGKR-pNA, based on the NS3-NS4A junction, and Ac-GLVKRR-pNA, based on the NS4B-NS5 junction. Three replicate sets of data from each peptide were obtained using 0.1 μM enzyme. Error = ± S.D. (B) The HsEg5 loop8 peptide, Ac-DPRNKR-pNA, exhibited comparable kinetics of hydrolysis to the native-derived peptides. Values are ± S.D. The overall s-ZPrC enzyme kinetics were similar to those already reported in the literature (Hill et al., 2018; Gruba et al., 2016; Lei et al., 2016; Leung et al., 2001).

Figure 4

Microtubule-bound HsEg5 motor domain is not protected from hydrolysis by s-ZPrC HsEg5 motor domain was incubated with or without taxol-stabilized microtubules and in the presence of either AMPPNP or ATP. HsEg5 motor domain plus ATP in the absence of microtubules was readily hydrolyzed by s-ZPrC (ATP). The addition of microtubules to the motor plus ATP (MT + ATP), which would promote transient motor/microtubule complexes, did not protect the motor from hydrolysis by s-ZPrC. The formation of long-term stable motor/microtubule complexes by AMPPNP (Wang et al., 2017; Gigant et al., 2013) also did not protect the motor from hydrolysis by s-ZPrC. Note that cleavage products attributable to tubulin were not observed over the time span of this experiment.

Figure 5

Schematic of the Zika polyprotein showing the segment utilized to create the n-ZPrC-eGFP expression construct The Zika polyprotein coding region from the NS2A-NS2B processing cleavage site (from residue # 1,369) through to the linker region between the NS3 peptidase domain (residue # 1,683) was synthesized (blue feature, Native NS2B-NS3) in-frame with C-terminal eGFP and cloned into the pcDNA3.1 transfection vector. Expressed under the control of the CMV promoter, the fusion protein exhibited ER-specific subcellular localization.

Figure 6

The subcellular localization of ectopic n-ZPrC-eGFP in HeLa cells is restricted to the endoplasmic reticulum (ER) HeLa cells were transiently co-transfected with n-ZPrC-eGFP (A, green channel) and mScarlet-b5^anchor ER marker (B, red channel) and examined by confocal microscopy. n-ZPrC-eGFP exhibits consistent colocalization with mScarlet-b5^anchor and is only found associated with the ER membranes (C, merge).

Figure 7

Timelapse images of live HeLa cells 24 h after being transfected with S135A-n-ZPrC-RFP670 inactive protease exhibit normal timing and progression through mitosis Mitotic defects, or delays, are not observed after expression of S135A-n-ZPrC-RFP670. Stable-transfected GFP-Tubulin (green) HeLa cells were transiently transfected with mScarlet-H3 (red) and peptidase-inactive S135A-n-ZPrC-RFP670 (ER network, cyan). These cells all completed mitosis within the normal range of 40–60′. Timelapse confocal data were collected 24 h post-transfection. Scale bar, 25μm, time = mm:ss.

Figure 8

HeLa cells exhibit prolonged metaphase delay and spindle positioning defects in the presence of native-like n-ZPrC-RFP670 Stable-transfected GFP-tubulin (green) HeLa cells were transiently transfected with mScarlet-H2A (red) and n-ZPrC-RFP670 (cyan). Timelapse confocal data were collected 24 h post-transfection. Also visible in the field of view (lower right) is a mitotic cell that is co-transformed for GFP-tubulin and mScarlet H2A, but with no detectable n-ZPrC-RFP670 fluorescence, and which divides with normal timing and stable spindle orientation. Scale bar, 25μm, time = mm:ss.

Figure 9

Design of the HsEg5 biosensor for the in vivo detection of specific ZPrC protease activity We modified a fluorescent protein reporter of Caspase activity to instead detect ZPrC activity in vivo. We substituted the original caspase cleavage target sequence with HsEg5 loop8 ZPrC target sequence (DPRNKRG) in an exposed loop of the bacterial phytochrome domain of the reporter. Cleavage of this segment permits integration of the biliverdin chromophore (rendered in “stick” cartoon form) into the phytochrome and activates far-red fluorescence. Biosensor fluorescence is not detectably activated by endogenous cellular proteases in the absence of active ZPrC. Continued association of the cleaved phytochrome domains are sustained by the split-GFP moiety.

Figure 10

ER-localized n-ZPrC-eGFP can cut the soluble HsEg5 loop8 biosensor mimic in MCF7 cells (A) The localization of transient co-transfected HsEg5 biosensor-GFP and n-ZPrC-eGFP in live cells by confocal microscopy. Cells expressing the biosensor display nucleocytoplasmic localization of the reporter GFP moiety in contrast to the reticulate ER network marked by n-ZPrC-eGFP. Asterisk marks a cell expressing n-ZPrC-eGFP without detectable nucleocytoplasmic biosensor. (B) All cells expressing both the HsEg5 biosensor and n-ZPrC protease exhibit cleavage-mediated activation of far-red fluorescence from the biosensor (white signal, n = 100). Cells transfected with either the biosensor alone (n = 100) or co-transfected with protease-inactive S135A-n-ZPrC-eGFP (n = 100) do not exhibit any detectable far-red fluorescence from the biosensor (not shown). Scale bar, 25 μm.

Figure 11

Model depicting the potential impact of ER-tethered Zika protease on asymmetric cell division (A) In a wild-type asymmetric mitosis, the cell division axis is established (dotted line) at metaphase and cell fate determinants (purple and yellow shading) are asymmetrically localized. After anaphase/telophase, the cell fate determinants are partitioned into a neuroblast (purple, NB) and a pluripotent ganglion mother cell (yellow, GMC). (B) An asymmetric mitosis with Zika protease (red) tethered to the reticulate ER membrane differs from wild-type cell division. The protease is able to cleave targets in close proximity to the ER, including HsEg5 and presumably other targets, and causes the spindle to exhibit abnormal mobility and mis-positioning with respect to the metaphase axis. Subsequent anaphase/telophase shows daughter cells with an aberrant mixture of cell fate determinants, which likely results in the termination of cell differentiation.