Conserved Outer Tegument Component UL11 from Herpes Simplex Virus 1 Is an Intrinsically Disordered, RNA-Binding Protein

Abstract

A distinguishing morphological feature of all herpesviruses is the multiprotein tegument layer located between the nucleocapsid and lipid envelope of the virion. Tegument proteins play multiple roles in viral replication, including viral assembly, but we do not yet understand their individual functions or how the tegument is assembled and organized. UL11, the smallest tegument protein, is important for several distinct processes in replication, including efficient virion morphogenesis and cell-cell spread. However, the mechanistic understanding of its role in these and other processes is limited in part by the scant knowledge of its biochemical and structural properties. Here, we report that UL11 from herpes simplex virus 1 (HSV-1) is an intrinsically disordered, conformationally dynamic protein that undergoes liquid-liquid phase separation (LLPS) in vitro Intrinsic disorder may underlie the ability of UL11 to exert multiple functions and bind multiple partners. Sequence analysis suggests that not only all UL11 homologs but also all HSV-1 tegument proteins contain intrinsically disordered regions of different lengths. The presence of intrinsic disorder, and potentially, the ability to form LLPS, may thus be a common feature of the tegument proteins. We hypothesize that tegument assembly may involve the formation of a biomolecular condensate, driven by the heterogeneous mixture of intrinsically disordered tegument proteins.IMPORTANCE Herpesvirus virions contain a unique tegument layer sandwiched between the capsid and lipid envelope and composed of multiple copies of about two dozen viral proteins. However, little is known about the structure of the tegument or how it is assembled. Here, we show that a conserved tegument protein UL11 from herpes simplex virus 1, a prototypical alphaherpesvirus, is an intrinsically disordered protein that undergoes liquid-liquid phase separation in vitro Through sequence analysis, we find intrinsically disordered regions of different lengths in all HSV-1 tegument proteins. We hypothesize that intrinsic disorder is a common characteristic of tegument proteins and propose a new model of tegument as a biomolecular condensate.

Keywords: RNA-binding protein; biomolecular condensate; conformational flexibility; herpesvirus; intrinsically disordered protein (IDP); liquid-liquid phase separation (LLPS); small‐angle X‐ray scattering (SAXS); structural model; tegument; viral assembly; viral protein.

FIG 1

Alignment and characteristics of UL11 homolog sequences. (A) Sequences of UL11 homologs from herpesviruses aligned to HSV-1 UL11 with HSV-1 residue numbers marked. Human virus sequences used include HSV-1 strain 17 UL11 (RefSeq accession no. YP_009137085.1), HSV-2 strain HG52 UL11 (RefSeq YP_009137162.1), VZV strain Dumas ORF49 (RefSeq NP_040171.1), EBV strain B95-8 BBLF1 (RefSeq YP_401686.1), CMV strain AD169 UL99 (RefSeq P13200.3), HHV-6A strain Uganda-1102 U71 (RefSeq NP_042964.1), HHV-6B strain Z29 U71 (RefSeq NP_050250.1), HHV-7 strain JI U71 (RefSeq YP_073811.1), and KSHV strain GK18 ORF38 (RefSeq YP_001129391.1). Other representative animal virus sequences used include Marek’s disease virus (MDV) strain Md5 UL11 (RefSeq YP_001033939.1), pseudorabies virus (PRV) composite strain UL11 (RefSeq YP_068364.1), murine cytomegalovirus (MCMV) strain Smith UL99 (RefSeq YP_214100.1), saimiriine herpesvirus 2 (SaHV-2) ORF38 (RefSeq NP_040240.1), and equine herpesvirus 2 (EHV-2) strain 86/67 myristoylated tegument protein (RefSeq NP_042635.1). All sequences show the NCBI reference sequence accession number or code in parentheses. Conserved residues are marked with an asterisk. Myristoylated glycines and palmitoylated cysteines, experimentally determined or predicted, are shown in italicized type in cyan text. Acidic clusters, experimentally defined or predicted, are boxed in red. Groups of basic residues are boxed in blue. Beta strands and alpha helices predicted by PSIPRED (http://bioinf.cs.ucl.ac.uk/psipred/) are highlighted in yellow and light teal, respectively. Disorder-promoting residues (A/G/S/P/Q/E/R/K) are colored in magenta. Residues predicted to be disordered by DISOPRED3 (http://bioinf.cs.ucl.ac.uk/psipred/) are underlined in gray. (B) Representation of disorder by residue in HSV-1 UL11 as predicted by FoldUnfold (http://bioinfo.protres.ru/ogu/). Residues predicted to be in natively folded regions are shown in blue, and residues predicted to be in unfolded regions are shown in red. Residues scoring below the threshold (black line) but surrounded by folded residues are to be treated as folded and are shown in cyan.

FIG 2

Expression and purification of UL11 constructs. (A) Schematic representation of UL11 sequence and UL11 constructs used in this work. Protease recognition sites are shown in gray. Construct names and molecular masses calculated from the sequence (MM_Calc) and determined experimentally (this paper) (by SEC, MS, and SEC-SAXS) are shown. WT, wild type; H₆, His₆ tag; GST, glutathione S-transferase tag; StII, Strep-tag II. (B) Purification scheme for UL11 constructs. (C) Coomassie blue-stained gels of purified UL11 constructs.

FIG 3

UL11 copurifies with RNA. (A) Fast protein liquid chromatography (FPLC) shows that UL11-StII is separated from copurifying nucleic acids on heparin resin. Protein, but not nucleic acids (NAs), binds heparin resin and elutes with a salt gradient (conductivity). UL11-StII is present in the eluted fraction, but not in the unbound fraction. The y axis shows absorbance or conductivity (in arbitrary units [AU]). Samples were resolved by SDS-PAGE and stained with Coomassie blue. The A₂₆₀/A₂₈₀ ratio shown below the SDS-PAGE confirms the presence of nucleic acids in the input fraction but not in the eluted fraction. (B) Nucleic acids that copurify with UL11-StII in acidic or neutral phenol-chloroform (P:C) are susceptible to digestion by RNase, but not DNase. The banding pattern is characteristic of E. coli rRNA, as marked. Samples were resolved on a formaldehyde agarose gel and stained with ethidium bromide. The gel in panel A was split to remove unrelated lanes, but contrast settings remain consistent between related gels.

FIG 4

UL11 is an elongated monomer. (A) MALDI-TOF spectrum of UL11-StII. The x axis shows the mass-to-charge ratio (m/z) in daltons and the y axis shows relative abundance (%) scaled for the mass range in x. (B) Size exclusion chromatography (SEC) A₂₈₀ traces and gels. The elution volumes of calibration standards used to calculate apparent molecular masses (kDa) are depicted in gray and with gray arrows. (C) Coomassie blue-strained gel of UL11 incubated with increasing amounts of BS³ cross-linker.

FIG 5

UL11 undergoes liquid-liquid phase separation. (A) Microscopic images of UL11 constructs in solution show representative images of liquid-liquid phase separation (LLPS) seen under a variety of conditions, including purification buffer and for H₆-UL11, crystallization screens (1:1 mix of protein-solution in vapor diffusion chamber). (B) Macroscopic images of UL11-StII in solution in purification buffer show reversibility of UL11-StII phase separation (cloudiness) upon temperature cycling. (C) Time lapse microscopic images of UL11-StII condensates in purification buffer display hallmark liquid-like properties of surface wetting and droplet merging (selected example in orange box). (D) LLPS (separation, orange circles; clear solution, blue triangles) occurs with increasing protein concentration, increasing salt concentration, and increasing molecular crowding.

FIG 6

UL11 lacks defined secondary structure and is sensitive to proteolysis. (A) CD spectra of UL11-StII at several concentrations show primarily random-coil characteristics. UL11-StII undergoes LLPS (LLPS, orange circles; clear solution, blue triangles) with increasing concentration in CD buffer at ambient temperature. See also Table 1. (B) Limited proteolysis of UL11-StII using trypsin, chymotrypsin, V8 protease, or proteinase K with increasing amounts of protease. (C) Sequence of UL11-StII with potential proteolysis sites colored (green, chymotrypsin; pink, trypsin; purple, V8). V8 fragment identified by MALDI-TOF mass spectrometry and N-terminal sequencing (see Fig. S1 in the supplemental material) is underlined in purple. The StII sequence is double underlined.

FIG 7

SAXS analysis of UL11. (A) Trace of integrated scattering intensities for each frame collected in the UL11-StII SEC-SAXS experiment. Buffer curves (frames shaded gray) were averaged and subtracted from each scattering curve to calculate R_g (diamonds) frame by frame. Data from curves giving a consistent R_g within the major SEC peak (frames shaded blue) were averaged, and the buffer average was subtracted from this sample average to give the subtracted SAXS curve in panel B. (B) Subtracted SAXS curve for UL11-StII with Guinier plots (inset, including linear fits) at the low-angle region (qRg < 1.3). These regions were selected with AUTORG such that residuals were evenly distributed around zero. Data points in gray were excluded from the Guinier analysis. (C) Pair distance distribution functions [P(r)] for UL11-StII calculated from scattering curves in panel B, comparing curves calculated with D_max forced to zero (inset) or not forced to zero (main panel). In the main panel, the faded blue line shows the continuation of the UL11 P(r) after the D_max when forced to zero. (D) Normalized (dimensionless) Kratky plot was calculated from scattering curve in panel B. Parameters from these data are also summarized in Table 2. Data for UL11-StII have been deposited into the Small Angle Scattering Biological Data Bank (SASBDB) under code SASDEX4.

FIG 8

Ensemble optimization modeling (EOM) of UL11-StII. (A and B) Histograms comparing the distribution of R_g values (A) and D_max values (B) between a pool of 10,000 randomly generated structures (gray) and three independently optimized representative ensembles of structures (blues in panel A; purples in panel B) indicate the presence of three conformations. (C) Representative structures of the conformations found in the optimized ensembles are shown with the average R_g (blue) and D_max (purple) values and the percentage of the ensemble they represent. These models suggest possible behavior of the flexible protein in solution, but they do not represent the only solution. These values are also summarized in Table 3.

FIG 9

Predicted disorder in HSV-1 tegument proteins. Predicted flexibility of HSV-1 tegument proteins as calculated by FoldUnfold (http://bioinfo.protres.ru/ogu/). Folded residues are indicated in blue, while unfolded residues are indicated in red. Protein length is not shown to scale.