Multi-interface licensing of protein import into a phage nucleus

Abstract

Bacteriophages use diverse mechanisms to evade antiphage defence systems. ΦKZ-like jumbo phages assemble a proteinaceous, nucleus-like compartment that excludes antagonistic host nucleases and also internalizes DNA replication and transcription machinery^1-4. The phage factors required for protein import and the mechanisms of selectivity remain unknown, however. Here we uncover an import system comprising proteins highly conserved across nucleus-forming phages, together with additional cargo-specific contributors. Using a genetic selection that forces the phage to decrease or abolish the import of specific proteins, we determine that the importation of five different phage nuclear-localized proteins requires distinct interfaces of the same factor, Imp1 (gp69). Imp1 localizes early to the nascent phage nucleus and forms discrete puncta in the mature phage nuclear periphery, probably in complex with direct interactor Imp6 (gp67), a conserved protein encoded in the same locus. The import of certain proteins, including a host topoisomerase, additionally requires Imp3 (gp59), a conserved factor necessary for proper Imp1 function. Three additional non-conserved phage proteins (Imp2 and Imp4/Imp5) are required for the import of two queried nuclear cargos (nuclear-localized protein 1 and host topoisomerase, respectively), perhaps acting as specific adaptors. We therefore propose a core import system that includes Imp1, Imp3 and Imp6, with multiple interfaces of Imp1 licensing transport through a protein lattice.

Extended Data Fig. 1

a, Efficiency of plating (EOP) of WT ΦKZ on EcoRI fusions, quantified as the number of plaque forming units (PFU)/mL on PAO1 expressing the indicated EcoRI fusion relative to PFU/mL on a nontargeting EcoRI strain. Closed circles indicate plaques that could be counted. Open circles indicate that no plaques were observed, and an arbitrary value of 1 plaque was recorded to calculate a non-zero limit of detection. Variation in plaquing efficiencies and limits of detection on a given strain result from different titers of the phage stock when the test was done. Bar heights represent the mean EOP between replicates where plaques could be counted. EcoRI-Nlp1 n=5 independent biological replicates, Imp2[EcoRI-Nlp1] n=4, EcoRI-Nlp2 n=5, EcoRI-Nlp3 n=3, EcoRI-Nlp4 n=3, EcoRI-Imp2 n=3, Imp1[EcoRI-Imp2] n=4, EcoRI-Imp1_ΦPA3 n=3, EcoRI-TopA n=3, EcoRI-Imp1_ΦKZ n=2. b–d, Plaque assays with WT ΦKZ or WT 14–1 (EcoRI-sensitive phage) (b) or WT ΦKZ and mutant phage (c, d) spotted in 10-fold serial dilutions on a lawn of PAO1 expressing indicated sfCherry2 fusions or an empty vector control (EV), with or without expression of a phage gene in trans from the bacterial attTn7 site (Tn7::impX) (c, d). e, Representation of EOP of WT ΦKZ and EcoRI-Nlp1 resistant mutant phages on PAO1 expressing EcoRI-Nlp1–4. EOP was calculated as the number of PFU/mL on the EcoRI-Nlp strain relative to PFU/mL on the non-targeting dead EcoRI-Nlp1 (dEcoRI-Nlp1) strain. EOP for each phage/strain pair is colored by mean EOP between three independent biological replicates. f, Imp2 phylogenetic tree. g, Plaque assays with WT ΦPA3 or mutant phage spotted in 10-fold serial dilutions on a lawn of PAO1 expressing indicated sfCherry2 fusions. Plaque assays were performed two (b, g) or three (c, d) independent times in biological replicates with similar results. Please see Supplementary Figure 4 for source data underlying graphical representations (a, e).

Extended Data Fig. 2

a–e, Representative images of live-cell fluorescence microscopy of PAO1 expressing the indicated sfCherry2 fusions, infected with WT or indicated mutant ΦKZ (EcoRI+WT ΦKZ, n=118 cells. EcoRI-Nlp2+WT ΦKZ, n=102. EcoRI-Nlp2+imp1 E310G ΦKZ, n=230. EcoRI-Nlp1+WT ΦKZ, n=69. EcoRI-Nlp1+imp2 K45N, n=58). Scale bars, 1 µm. “Excluded” refers to localization of sfCherry2-fused proteins outside of the phage nucleus. Microscopy was performed as in Fig. 1c and replicated two (EcoRI-Nlp1+imp2 K45N) or three (EcoRI+WT ΦKZ, EcoRI-Nlp2+WT ΦKZ, EcoRI-Nlp2+imp1 E310G ΦKZ, EcoRI-Nlp1+WT ΦKZ) independent times in biological replicates with similar results.

Extended Data Fig. 3

a, Live-cell fluorescence microscopy of PAO1 expressing Imp1-mNeonGreen (Imp1-mNG) from the attTn7 site, infected with WT ΦKZ (Top eight panels, n=154 cells; scale bar, 1 µm) or uninfected (bottom panel, n=82 cells; scale bar, 2 µm). b, Live-cell time-lapse fluorescence microscopy of PAO1 attTn7::Imp1-mNG infected with WT ΦKZ. n=102 cells; scale bar, 1 µm. “t=” indicates time in minutes after injected phage DNA is first seen as puncta at the cell pole. Microscopy was performed three independent times in biological replicates with similar results. c, Plaque assays with the indicated WT or mutant phage spotted in 10-fold serial dilutions on a lawn of PAO1 expressing the indicated sfCherry2 fusions, with or without expression of the appropriate phage gene in trans from the bacterial attTn7 site (Tn7::ImpX). Plaque assays were performed as in Fig. 1b and replicated two independent times in biological replicates with similar results.

Extended Data Fig. 4

a, Plaque assays with WT ΦKZ or WT 14–1 (EcoRI-sensitive phage) spotted in 10-fold serial dilutions on a lawn of PAO1 expressing the indicated sfCherry2 fusions. b, Top output Imp1 model from AlphaFold 2, with mutated residues from isolated imp1 mutant phages color coded by the EcoRI selection from which they were isolated. c, Overlay of all five output Imp1 models from AlphaFold 2, colored by confidence scores. d, Surface map of electrostatic potential (semi-transparent overlay) of the top Imp1 predicted structural model from the same view (right, top), or rotated 90° view (right, bottom). Several mutated residues are indicated. e, Plaque assays with the indicated WT or mutant phage spotted on PAO1 expressing the indicated sfCherry2 fusions, with or without Imp3 or the Imp3 operon (p18-imp3-orf60-orf61) expressed in trans. Plaque assays were performed as in Fig. 1b and replicated two (a) or three (e, using 1 or 2 mM IPTG to induce expression from the attTn7 site) independent times in biological replicate with similar results.

Extended Data Fig. 5

a, Overlay of all five output Imp3 models from AlphaFold 2, colored by confidence scores. b, Imp4 and Imp5 phylogenetic trees. c, Conservation of Imp homologs across nucleus-forming jumbo phages. d, Plaque assays with WT ΦKZ or WT 14–1 (EcoRI-sensitive phage) spotted in 10-fold serial dilutions on a lawn of PAO1 expressing the indicated sfCherry2 fusions. e, Plaque assays with the indicated WT or mutant phage spotted in 10-fold serial dilutions on a lawn of PAO1 expressing the indicated sfCherry2 fusions, with or without individual or combinations of Imp1–5 or the Imp3 operon (p18-imp3-orf60-orf61) expressed in trans. Plaque assays were performed as in Fig. 1b and in two (d) or three (e) independent biological replicates with similar results.

Extended Data Fig. 6

a, Top left, top Imp1-Nlp2 model output from AlphaFold3 is displayed. Imp1 is colored in grey and Nlp2 is colored in purple, and Nlp2 N-terminal domain (NTD) and C-terminal domain (CTD) are indicated. All five models output from AlphaFold3 were aligned and are structurally similar, but only the first model is shown for clarity. Bottom left, top model overlayed with confidence scores, colored from high (blue) to low (red) confidence (ipTM = 0.74, pTM = 0.73). Right, boxed views show Imp1 residue positions mutated under selection with EcoRI-Nlp2 (pink) at the predicted interface between Imp1 and Nlp2. Non-carbon atoms are colored according to identity (oxygen in red, nitrogen in blue, sulfur in yellow). b, Plaque assays with WT ΦKZ or WT F8 (EcoRI-sensitive phage) spotted in 10-fold serial dilutions on a lawn of PAO1 expressing the indicated EcoRI fusions. FL, full length. NTD, N-terminal domain. CTD, C-terminal domain. Plaque assays were repeated three independent times in biological replicates with similar results. c, Nlp2 residues positioned at the predicted interface with Imp1 (residues 345–366, highlighted in blue) and Nlp3 proposed import signal (residues 77–95, highlighted in yellow).

Figure 1:. Mutations in previously uncharacterized phage genes imp1 and imp2 reduce protein import into phage nucleus

a, Live-cell fluorescence microscopy of P. aeruginosa strain PAO1 expressing the indicated sfCherry2 fusion proteins infected with WT ΦKZ or the indicated phage mutant (EcoRI+WT ΦKZ, n=118 cells. EcoRI-Nlp2+WT ΦKZ, n=102. EcoRI-Nlp2+imp1 E310G ΦKZ, n=230. EcoRI-Nlp1+WT ΦKZ, n=69. EcoRI-Nlp1+imp2 K45N, n=58). DAPI stain indicates phage DNA within the phage nucleus (white arrow). b, d, f, Plaque assays with the indicated WT or mutant phage spotted in 10-fold serial dilutions on a lawn of PAO1 expressing indicated sfCherry2 fusion. dEcoRI, catalytically inactivated EcoRI. c, e, Live-cell fluorescence microscopy of Imp1 fused to mNeonGreen (mNG) infected with WT ΦKZ (n = 154), imaged at 50–60 min. post-infection (c) or Imp2 fused to mNeonGreen and infected with WT ΦKZ (n = 129) (e). g, Plaque assays with the indicated WT or mutant phage spotted in 10-fold serial dilutions on a lawn of PAO1 expressing the indicated sfCherry2 fusion, with or without expression of Imp2 in trans (+/– Imp2). h, Schematic representing the factors required for nuclear import of proteins queried in the EcoRI selections, as determined by the escape mutations isolated in imp1 and imp2 from each selection (i.e. Nlp2 → Imp1 indicates Nlp2 requires WT Imp1 for its nuclear import). All plaque assays were repeated three independent times in biological replicates with similar results. For plate source images, please see Supplementary Figure 3. All microscopy experiments are representative of two (a, e) or three (a, c) biologically independent experiments with similar results. Scale bars, 1µm.

Figure 2:. Imp1 possesses distinct functional interfaces to enable protein import specificity

a, c, Plaque assays with indicated WT or imp1 mutant phages spotted in 10-fold serial dilutions on a lawn of PAO1 expressing the indicated sfCherry2 fusion. dEcoRI, catalytically inactivated EcoRI. Plaque assays were conducted as in Fig. 1b. b, AlphaFold 2 predicted structural model of Imp1, with mutated residues from isolated imp1 mutant phages color coded by the EcoRI selection from which they were isolated. The top output model is displayed in the figure. d, Phylogenetic tree of Imp1 homologs. Cultured phages are colored by host species, and several model phages are also labeled by name. Uncolored phages are from metagenomic sequencing. A closed black circle indicates that no phage nuclear shell homolog was detected in the corresponding phage genome by 3 iterations of PSI-BLAST or local blastp. An open black circle indicates that the associated genome record is incomplete, and no phage nuclear shell homolog was detected by the same methods. A closed blue circle indicates that no nuclear shell homolog was detected by the same methods, but the tubulin homolog responsible for centering the phage nucleus in the cell, PhuZ, was detected by blastp. All plaque assays were repeated three independent times in biological replicates with similar results.

Figure 3:. Imp3 is required for proper Imp1 function

a, b, Plaque assays with the indicated WT or mutant phage spotted on PAO1 expressing the indicated sfCherry2 fusion, with or without Imp1 expressed in trans (+/– Imp1) (a); Imp1_ΦPA3 indicates the phage ΦPA3 Imp1 homolog (b); dEcoRI, catalytically inactive EcoRI. Plaque assays were performed as in Fig. 1b. c, AlphaFold 2 predicted structural model of Imp3 with mutated DNA bases or residues from isolated imp3 mutant phages color coded by the EcoRI selection from which they were isolated. The top output model is displayed in the figure. d, Plaque assays with the indicated WT or mutant phage spotted on PAO1 expressing the indicated sfCherry2 fusion. e, Schematic representing the factors required for nuclear import of all proteins queried in the EcoRI selections in this work, as determined by the escape mutations isolated in imp1-imp5 from each selection. All plaque assays were repeated three independent times in biological replicates with similar results.

Figure 4:. Imp1 binds to Imp6 and cargo protein Nlp2

a, b, Western blot analysis (anti-Imp1) of a His pull-down of His-tagged Imp1 with FLAG-tagged Imp6, gp70, or Nlp1-His with Imp6-FLAG as a negative control (a) or a FLAG pull-down of Imp1-His with FLAG-Imp6 or FLAG-gp70 (b); I, input; E, elution. c, Size exclusion chromatography (SEC) of Imp1-His alone or in complex with Imp6-FLAG. Peak Imp1-His/Imp6-FLAG SEC fractions (indicated by mL collected off the column) were run on an SDS-PAGE gel and stained with coomassie. d, Estimation of Imp1-His/Imp6-FLAG complex molecular weight by mass photometry. e, Western blot analysis of FLAG-tag pull downs of Nlp2-FLAG, FLAG-Imp6, or sfCherry2-FLAG with Imp-His WT or indicated mutants. RNAP serves as an input loading control. Pull-downs in Fig 4a, b were performed two independent times with similar results. Pull-downs in Fig. 4e were performed three independent times with similar results. Please see Supplementary Figure 1 for uncropped and unprocessed gel source data.