Virology 432 (2012) 353–360

Contents lists available at SciVerse ScienceDirect

Virology

0042-68

http://d

n Corr

Wintert

E-m1 Cu

Visceral

journal homepage: www.elsevier.com/locate/yviro

The herpes simplex virus 1 US3 regulates phospholipid synthesis

Peter Wild a,b,n, Anna Paula de Oliveira b, Sabrina Sonda c,1, Elisabeth M. Schraner a,b,Mathias Ackermann b, Kurt Tobler b

a Institute of Veterinary Anatomy, University of Zurich, Switzerlandb Institute of Virology, University of Zurich, Switzerlandc Institute for Parasitology, University of Zurich, Switzerland

a r t i c l e i n f o

Article history:

Received 23 April 2012

Returned to author for revisions

7 May 2012

Accepted 12 June 2012Available online 11 July 2012

Keywords:

Herpes virus

Nuclear membranes

Golgi membranes

Phospholipid synthesis

[3H]-choline

TEM

Morphometry

22/$ - see front matter & 2012 Elsevier Inc. A

x.doi.org/10.1016/j.virol.2012.06.020

esponding author at: Institute of Virology,

hurerstrasse 266a, Zurich CH-8057, Switzerla

ail address: pewild@access.uzh.ch (P. Wild).

rrent address: Swiss Hepato-Pancreato-Bil

and Transplantation Surgery, University Hos

a b s t r a c t

Herpes simplex virus type 1 capsids bud at nuclear and Golgi membranes for envelopment by

phospholipid bilayers. In the absence of US3, nuclear membranes form multiple folds harboring virions

that suggests disturbance in membrane turnover. Therefore, we investigated phospholipid metabolism

in cells infected with the US3 deletion mutant R7041(DUS3), and quantified membranes involved in

viral envelopment. We report that (i) [3H]-choline incorporation into nuclear membranes and

cytoplasmic membranes was enhanced peaking at 12 or 20 h post inoculation with wild type HSV-1

and R7041(DUS3), respectively, (ii) the surface area of nuclear membranes increased until 24 h of

R7041(DUS3) infection forming folds that equaled �45% of the nuclear surface, (iii) the surface area of

viral envelopes between nuclear membranes equaled �2400 R7041(DUS3) virions per cell, and (iv)

during R7041(DUS3) infection, the Golgi complex expanded dramatically. The data indicate that US3

plays a significant role in regulation of membrane biosynthesis.

& 2012 Elsevier Inc. All rights reserved.

Introduction

Capsids of herpes viruses are formed in replication centers(RCs) (Quinlan et al., 1984) in host cell nuclei and transported tothe nuclear periphery where they bud at the inner nuclearmembrane (INM) acquiring the tegument and the envelope(Roizman and Knipe, 2001). The result is a fully enveloped virionlocated in the perinuclear space (PNS). Translocation of virusparticles out of the PNS is assumed to be accomplished via fusionof the viral envelope with the outer nuclear membrane (ONM)releasing capsid and tegument into the cytoplasmic matrix, viavesicle formation or via intraluminal transportation throughcisternae of the rough endoplasmic reticulum (RER) (Roizmanet al., 2007). Alternatively, capsids of pseudo rabies virus (PrV)(Klupp et al., 2011) and of simian agent 8 (Borchers and Oezel,1993) also may overcome the nucleocytoplasmic barrier afterimpairment of nuclear membranes. Impairment of the nuclearenvelope is considered likely to occur at nuclear pores in cellsinfected with herpes simplex virus 1 (HSV-1) (Leuzinger et al.,2005; Wild et al., 2009) and bovine herpes virus 1 (BoHV-1) (Wild

ll rights reserved.

nd. Fax: þ41 1 635 89 11.

iary Center, Department of

pital, Zurich.

et al., 2005) so that capsids can gain direct access to thecytoplasmic matrix (Roizman et al., 2007). In the cytoplasm, capsidsbud at membranes of the trans Golgi network (Mettenleiter, 2004;Mettenleiter et al., 2006) or at membranes of vacuoles (Homman-Loudiyi et al., 2003; Stannard et al., 1996) or at any site of the Golgicomplex (Leuzinger et al., 2005; Wild et al., 2002) or at the RER(Leuzinger et al., 2005).

The Us3 gene encodes a protein kinase present in HSV-1 infectedcells (Purves et al., 1987). The HSV-1 US3 protein kinase blocksapoptosis (Benetti et al., 2003; Galvan and Roizman, 1998; Jeromeet al., 1999; Leopardi et al., 1997; Munger et al., 2001; Munger andRoizman, 2001; Ogg et al., 2004) and enables translocation of virusparticles from the nucleus to the cytoplasm associated withphosphorylation and distribution of the viral proteins UL31 andUL34 on the nuclear rim (Poon et al., 2006; Reynolds et al., 2004,2001, 2002). In cells infected with US3 deletion mutants (HSV-1DUS3), nuclear membranes form folds and invaginations contain-ing large numbers of virions (Poon et al., 2006; Reynolds et al.,2002). Translocation of HSV-1DUS3 virus particles out of theseinvaginations cannot be accomplished because glycoprotein B (gB),which is involved in fusion of the viral envelope with the ONM, isnot phosphorylated (Wisner et al., 2009).

The viral envelope derives from cell membranes. It consists ofa phospholipid bilayer and viral glycoproteins that are insertedduring the process of budding (Roizman et al., 2007). Phospholi-pids have been shown to be newly synthesized for viral envelope

P. Wild et al. / Virology 432 (2012) 353–360354

formation in HSV-1 (Asher et al., 1969; Sutter et al., 2012) and inPrV (Ben-Porat and Kaplan, 1971) infected cells. Folding ofnuclear membranes and accumulation of HSV-1DUS3 virionsbetween nuclear membranes (Poon et al., 2006; Reynolds et al.,2002) raise the question of US3 function in phospholipid meta-bolism. Therefore, we investigated [3H]-choline incorporation intocell membranes in Vero cells infected with the US3 deletionmutant R7041(DUS3) (Purves et al., 1987). We show that [3H]-choline incorporation into lipids of both the nuclear and cyto-plasmic fraction is dramatically increased suggesting enhance-ment of de novo phospholipid synthesis, which led to considerableenlargement of nuclear and cytoplasmic membranes compared towild type (wt) HSV-1 infection. The data indicate that US3 isinvolved in down-regulation of phospholipid biosynthesisinduced by HSV-1.

Results

Incorporation of [3H]-choline into both nuclear and cytoplasmic

membranes is enhanced after R7041(DUS3) infection

Cells infected with the well characterized deletion mutantR7041(DUS3) (Poon et al., 2006; Purves et al., 1987, 1991)typically showed folds of nuclear membranes associated withaccumulation of virions that derive by budding at the INM (Fig. 1).Phospholipids needed for envelopment and folding of membranesare assumed to be supplied by de novo synthesis. Therefore,we exposed cells to [3H]-choline for 4 h prior to harvesting,isolated nuclei and extracted lipids from the nuclear and cyto-plasmic fraction. As shown in Fig. 2, isolated nuclei are sur-rounded by the INM, the ONM and associated virions as well assome additional cytoplasmic membranes whereas the cytoplasmicfraction consisted of membranes. Concentration of virus confirmedthat the majority of virus particles were in the nuclear fraction,including capsids and enveloped virions. However, most of the viralenvelopes seem to have disrupted during the concentrationprocedure.

Scintillation counting revealed significantly higher numbers ofdisintegrations per minutes (DPMs) in the nuclear fraction at 9, 12,16 and 20 hpi with wt HSV-1 or the repair mutant R2641 comparedto mock infection. After infection with R7041(DUS3), DPMs wereeven significantly higher compared to wt HSV-1 or R2641 at 12, 16,20 and 24 hpi, peaking at 20 hpi (Fig. 3A). In contrast to the nuclearfraction, DPMs in the cytoplasmic fraction of virus infected cells

Fig. 1. TEM image taken from a section through a nucleus 20 hpi with

R7041(DUS3) showing multiple folds and invaginations of the INM (i) associated

with accumulation of virions (v), numerous capsids (c) distributed throughout the

nuclear matrix, and replication centers (r). Bar, 1 mm.

were significantly higher only at 12, 16 and 20 hpi compared tomock infection (Fig. 3B). DPMs in the R7041(DUS3) cytoplasmicfraction was only significantly higher at 16 and 20 hpi compared towt HSV-1 or R2641, peaking also at 20 hpi. Interestingly, DPMs inthe cytoplasmic fraction 24 hpi with any virus were substantiallylower compared to mock infection. The data suggest that (i) [3H]-choline was incorporated into membrane phospholipids of thecytoplasmic fraction as well as into membrane phospholipids ofthe nuclear fraction of HSV-1 infected cells at a significantlyhigher level than into that of mock infected cells, (ii) [3H]-cholineincorporation was dramatically enhanced upon R7041(DUS3)infection, (iii) [3H]-choline incorporation into membrane phos-pholipids ceases after 20 h of inoculation with wt HSV-1, and (iv)[3H]-choline incorporation into nuclear membranes proceeds inR0741(DUS3) infection until 24 hpi although at a lower level.

Nuclei expand during HSV-1 and R7041(DUS3) infection

Phospholipids are needed for maintenance of membrane integ-rity in expanding nuclei, for formation of membrane folds and forenvelopment of virions in the course of capsid budding. Nuclei wereshown to expand upon HSV-1 infection as deduced from calcula-tions of data obtained in 2 dimensional images (Monier et al., 2000;Simpson-Holley et al., 2005). The shape of nuclei of Vero cellsgrown as monolayer is a triaxial ellipsoid. Therefore, we determinednuclear size of R7041(DUS3) and HSV-1 infected Vero cells bycalculation of volume and surface area on the basis of the half axesa, b, and c measured on confocal images. As shown in Fig. 4A, themean nuclear volume in wt HSV-1 infection was �390 mm3 largerat 9 hpi compared to mock infection, and �185 mm3 larger at20 hpi compared to 9 hpi. After R7041(DUS3) infection, nuclearvolume was similar as in wt HSV-1 infected cells at 9 and 24 hpi butsomewhat smaller at 12, 16 and 20 hpi. Nuclear surface area, alsocalculated on the basis of the half axes, increased accordingly tonuclear expansion (Fig. 4B). At 9 hpi, the mean nuclear surface was�185 mm2 larger in R7041(DUS3) and wt HSV-1 infected cellscompared to mock infected cells, but only �60 mm2 larger at24 hpi compared to 9 hpi. The data indicate that nuclei expandsubstantially early in infection and continue to expand at a muchslower rate as infection proceeds, and that progress of nuclearexpansion is somewhat slower in R7041(DUS3) infection than in wtHSV-1 infection. Therefore, we conclude that phospholipids areneeded for membrane maintenance in expanding nuclei in largeamounts especially during the first 9 h of infection and that thelarge incorporation of [3H]-choline later in infection is needed forsomething else.

Nuclear membranes form multiple folds and invaginations

The nuclear membrane was shown to form invaginations inwhich virions accumulate upon infection with US3 deletionmutants (Mou et al., 2009; Poon et al., 2006; Reynolds et al.,2002; Wisner et al., 2009). In order to determine the amount ofmembranes involved in folding of membranes, we calculatedthe surface area of folds per mean nuclear volume on the basisof data obtained by morphometric analysis of electron micro-scopic images taken at random. The surface area of folds turnedout to be 90 (736) mm2 per mean nuclear volume of 927(7387) mm3 and mean nuclear surface area of 529 (766) mm2

at 9 hpi in R7041(DUS3) infected cells. It increased to 240(728) mm2 per mean nuclear volume of 983 (7200) mm3 andmean nuclear surface area of 546 (761) mm2 at 24 hpi (Fig. 5). Inwt HSV-1 infection, the surface area of folds was only 26(756) mm2 and 43 (7125) mm2 at 20 and 24 hpi, respectively.Therefore, we conclude that US3 is involved in regulation ofmembrane biosynthesis.

Fig. 2. TEM control of nuclear and cytoplasmic fraction. (A) Section through an isolated R7041(DUS3) infected nucleus with multiple folds and invaginations of the INM

(i) associated with virions (v), impaired ONM (o), a few capsids (c) and replication centers (r). (B) Cytoplasmic fraction consisting merely of cell membranes as revealed by

negative staining with Na-phosphotungstic acid. (C) Negatively stained virus isolates from the nuclear fraction with �1600 virus particles (c) per mesh, and (D) from the

cytoplasmic fraction with �34 virus particles per mesh. Bars, 500 nm.

P. Wild et al. / Virology 432 (2012) 353–360 355

R7041(DUS3) virions accumulate in the PNS

Virions of US3 deletion mutants accumulate in invaginations ofnuclear membranes (Mou et al., 2009; Poon et al., 2006; Reynoldset al., 2002; Wisner et al., 2009) as shown for R7041(DUS3) virionsin Fig. 1. To quantify the total amount of membranes of theenvelope of these virions, we determined the total envelopesurface area by morphometric analysis on electron microscopicimages taken at random. Calculation in relation to the meannuclear volume revealed that the total envelope surface area ofthese virions was 56 (785) mm2 at 9 hpi but 300 (730) mm2 at24 hpi with R7041(DUS3). In contrast, the total envelope surfacearea was only 23 (747) mm2 at 20 hpi and 65 (783) mm2 at24 hpi with wt HSV-1. We conclude that [3H]-choline incorpora-tion were used for phospholipids synthesis to provide membranesneeded for envelopment of R7041(DUS3) capsids by budding atthe INM as has been shown recently also for wt HSV-1 capsids(Sutter et al., 2012).

The Golgi complex disassembles and expands during R7041(DUS3)

infection

HSV-1 capsids can bud at membranes of the trans Golginetwork acquiring the envelope and tegument and concomitantlybecome enclosed in a transport vacuole, a process referred to aswrapping (Mettenleiter, 2002, 2004; Mettenleiter et al., 2006).Capsids were also shown to bud at membranes of any site of theGolgi complex (Leuzinger et al., 2005; Wild et al., 2002) or atmembranes of Golgi derived vacuoles (Homman-Loudiyi et al.,2003; Leuzinger et al., 2005). The result of wrapping is a smallconcentric vacuole filled by a single virion (Fig. 6A, B and D) asdescribed in detail (Wild et al., 2002). Large vacuoles were shownto detach from lateral cisternae (Wild et al., 2002) and bear one orseveral virions (Fig. 6B–D). Disturbance in wrapping leads to

malformed virions (Fig. 6D). R7041(DUS3) virions were rarely foundin vacuoles derived from wrapping or in larger vacuoles (Fig. 6A andC). Instead, Golgi fields appeared as multiple stacks comprisingthick electron dense membranes. Cisternae of small stacks werealso found to be dilated to vacuole-like structures occasionallycontaining virions. The high radioactivity detected in the cytoplas-mic fraction prompted us to quantify Golgi size (Fig. 7A). Thesurface area of Golgi membranes including membranes of wt HSV-1virion bearing vacuoles calculated per mean cell volume wasapproximately twice as high as in wt HSV-1 and R7041(DUS3)infected cells compared to mock infected cells 9 to 16 hpi. At 20 and24 hpi, the Golgi surface area in wt HSV-1 and mock infected cellswere similar, that in R7041(DUS3) infected cells, however, approxi-mately by a factor of 2.5 larger. These data suggest that in theabsence of US3 [3H]-choline was incorporated into newly synthe-sized phospholipids used for building of Golgi membranes.

Cell volume increases during wt HSV-1 and R7041(DUS3) infection

Recently, we showed that cell volume had increase at 9 to12 hpi with wt HSV-1 infection (Sutter et al., 2012). Here, weconfirm increment in cell volume peaking at 12 hpi but to a lessextent as reported (Fig. 7B). The data obtained for mock infectedcell in this study are in average about 15% lower. Whether thepresent data represent underestimation or the previous dataoverestimation is difficult to prove. Possibly, the true values aresomewhere in between. Nevertheless, the cell volume tended tobe lower 20 and 24 hpi with wt HSV-1 whereas it remained at thesame level as at 12 hpi with R7041(DUS3). Enlargement of cellvolume by 1100 mm3 as measured at 12 hpi equals an enlarge-ment of the cell surface area by �650 mm2. Hence, phospholipidsare required for maintaining plasma membrane integrity duringboth wt HSV-1 and R7041(DUS3) infection provided that numberand size of microvilli are not reduced.

Fig. 3. Means and standard deviations of disintegrations per minutes (DPMs)

obtained from the nuclear fraction (A) and the cytoplasmic fraction (B) 9, 12, 16,

20 and 24 hpi with wt HSV-1, R7041(DUS3), repair mutant R2641 or after mock

infection. Level of significance compared to mock infected cells is indicated by a

(po0.05), a0 (po0.01), a00 (po0.001). Level of significance between R7041(DUS3)

and wt HSV-1 or R2641 is indicated by b (po0.05), b0 (po0.01), b00 (po0.001);

n¼4. DPMs of the nuclear fraction of HSV-1 or R264 infected cells were higher at

9 to 20 hpi but lower or equal at 24 hpi compared to that of mock infected cells.

DPMs were equal in the cytoplasmic fraction at 9 hpi with wt HSV-1 or R2641

compared to mock but substantially lower at 24 hpi. DPMs after R7041(DUS3)

infection were much higher at 12 to 20 hpi compared to wt HSV-1 infection

peaking at 20 hpi in the nuclear fraction but at 16 hpi in the cytoplasmic fraction.

Fig. 4. Nuclear volume and nuclear surface area (mean and standard deviation)

calculated on the basis of the three axes a, b, and c measured on confocal images

were significantly larger at any time after wt HSV-1 and R7041(DUS3) infection

compared to mock infected cells. Level of significance compared to mock infected

cells is indicated by ***po0.001; n¼25.

Fig. 5. Surface area of nuclear membrane folds and envelope of associated virions

(mean and standard deviation) determined by morphometric analysis and

calculated per mean nuclear volume. Note that the high standard deviation is

due to random sampling that included nuclei with and without folds and virions.

P. Wild et al. / Virology 432 (2012) 353–360356

Discussion

Membranes required for budding of capsids at the nuclear andGolgi level either need to be recruited from membrane compart-ments, which are not involved in envelopment, or supplied by denovo synthesis. Cell membranes are bilayers composed of about14 major phospholipids (van Meer et al., 2008). Two of themcontain choline. Choline is esterified to diacylglycerol formingphosphatidylcholine, and to ceramide forming sphingomyelin.Phosphatidylcholine is the most abundant phospholipid in eukar-yotic cell membranes comprising 40% to 50% of the phospholipidmass. It contributes �58% of the phospholipids to nuclear andRER membranes, and �55% to Golgi membranes (Henneberryet al., 2002; Zinser et al., 1991). Sphingomyelin is present on Golgimembranes contributing �15% of the phospholipids (van Meeret al., 2008). Therefore, [3H]-choline is an ideal compound forstudying phospholipid synthesis in HSV-1 infection.

HSV-1 has been shown to stimulate phospholipid synthesis(Sutter et al., 2012) necessary to facilitate envelopment bybudding of capsids at the INM on a large scale (Wild et al.,2009) and at Golgi membranes. Envelopment at the Golgi com-plex is accomplished either by wrapping at the trans Golgi

network (Mettenleiter, 2004; Mettenleiter et al., 2006), or bybudding at any site of the Golgi complex (Leuzinger et al., 2005;Wild et al., 2002) or at membranes of Golgi derived vacuoles(Homman-Loudiyi et al., 2003; Leuzinger et al., 2005). Alterna-tively, there is evidence that virions can be intraluminallytransported from the PNS into Golgi cisternae (Leuzinger et al.,2005; Wild et al., 2002). Vacuoles containing virions dispersetowards the cell periphery where the vacuolar membrane may fusewith the plasma membrane releasing virions into the extracellularspace. Plasma membrane constituents may be recycled back to theGolgi complex (Orci et al., 1981) or to endosomes. Despite ofmembrane recycling envelopment at the Golgi complex results ina net loss of membranes. The high incorporation of [3H]-cholinepeaking at 12 hpi with HSV-1 indicates that phospholipids synthesisis enhanced early in infection whereas it ceases as infectionproceeds. In contrast, phospholipid synthesis is drastically enhanceduntil 20 hpi with R7041(DUS3) suggesting that de novo synthesizedphospholipids were responsible for doubling of the Golgi membrane

Fig. 6. TEM images from the Golgi area in cryo-fixed and freeze substituted cells 20 hpi. ((A) or (C)) In R7041(DUS3) infections, the Golgi complex formed distinct small

stacks (arrows), or small stacks (C) associated with vacuoles or dilated Golgi cisternae which rarely contain virions (vv). Capsids budding at Golgi membranes may

concomitantly form transport vacuoles (vw in (A)), a process referred to as wrapping (cw in (C)). The number of R7041(DUS3) capsids in the cytoplasm is low in contrast to

wt HSV-1 infection (B) where capsids bud at Golgi membranes (cw) resulting in small concentric vacuoles (vw) as described in detail (Wild et al., 2002). (D) The phenotype

of R2641 envelopment is similar to wt HSV-1. This image was taken to show that vacuoles can contain more than one virion (vv) suggesting another way of envelopment

than wrapping (Homman-Loudiyi et al., 2003; Leuzinger et al., 2005; Wild et al., 2002), and that envelopment may result in malformed virions (arrows). Bars, 500 nm.

P. Wild et al. / Virology 432 (2012) 353–360 357

surface area late in infection. Envelopment of R7041(DUS3) capsidsat the Golgi level is suggested to be markedly reduced because (i)virions were retained in the PNS as discussed below, and (ii) thereare little indications for successful budding at Golgi membranes asshown in Fig. 6 and verified by quantitative analysis (unpublisheddata). The immense enlargement of Golgi membrane surface areamay thus be related to impaired envelopment. However, in wt HSV-1 infection, we included the membranes of vacuoles containing wtHSV-1 virions for determination Golgi membrane surface area.Therefore, the difference in Golgi size cannot be explained by thefailure of R7041(DUS3) capsids envelopment alone indicating thatthe enlargement of Golgi membrane surface area in the absence ofUS3 is not or not only related to impaired envelopment but ratherdue to a failure in regulation of phospholipid synthesis.

[3H]-choline was also incorporated into nuclear membranesthough generally to a slightly lower level than into cytoplasmicmembranes. In contrast to cytoplasmic membranes, [3H]-cholineincorporation into nuclear membranes was already apparent at9 hpi with both wt HSV-1 and R7041(DUS3). Newly synthesizedphospholipids may have been required for maintaining mem-brane integrity in the course of nuclear expansion starting early ininfection, to provide membranes for budding of capsids at theINM, and for formation of membrane folds and invaginations.The nuclear envelope is a double coat. Hence, enlargement of thesurface area by 185 mm2 at 9 hpi requires a total of 370 mm2 thatmust have been supplied early in R7041(DUS3) infection. Onlyslightly less has to be supplied in wt HSV-1 infection. The demandfor membranes increases as infection proceeds. Formation ofmembrane folds required additional 240 mm2 per cell by 24 hpiwith R7041(DUS3). The membranes required for budding ofR7041(DUS3) capsids was estimated to equal 300 mm2 per cell

by 24 hpi. Taken together, the total amounts of membranes beingproduced during 24 h of R7041(DUS3) infection equals 910 mm2

per cell which is �2.6 times more than the nuclear surface ofmock infected cells. Assuming that the number of wt HSV-1capsids budding at the INM during 24 h is similar to that ofR7041(DUS3), the total membrane area being produced wouldequal 610 mm2 per cell or �1.8 times more than the nuclearsurface in mock infected cells. If these data are accurate thequestion arises about the fate of the viral envelope inserted intothe ONM during de-envelopment of virions to release capsids intothe cytoplasmic matrix. As discussed earlier (Sutter et al., 2012)the ONM would form folds unless membranes would be shiftedeither to the INM or to the RER. Indeed, nuclear membranesformed folds in R7041(DUS3) infected cells. However, these foldsare considered very unlikely to be caused by insertion of viralenvelopes during de-envelopment because the envelope of US3deletion mutants is unable to fuse with the ONM (Mou et al.,2009; Reynolds et al., 2001, 2002; Wisner et al., 2009) that resultsin accumulation of virions in the PNS. Therefore, the membranefolds are rather caused by over production of phospholipidssuggesting that US3 is involved in regulation of phospholipidsynthesis. The mechanism needs to be investigated. Possibly, US3influences membrane assembly via phosphorylation of UL31 andUL34, which plays a significant role through interaction withlamins in viral egress (Mou et al., 2008; Reynolds et al., 2004;Simpson-Holley et al., 2004). Alternatively, US3 might interfere inthe biosynthetic pathway of phospholipids that are mainly synthe-sized at the RER and the Golgi complex (Henneberry et al., 2002;van Meer et al., 2008).

The surface area of the envelope of a virion with a diameter of200 nm (Grunewald et al., 2003) is 0.125 mm2. The total surface

Fig. 7. Mean and standard deviation of the surface area of Golgi membranes

(A) calculated per mean cell volume (B) as describe in material and methods. Level

of significance compared to mock infected cells is indicated by a0 (po0.05), a00

(po0.01), a000 (po0.001). Level of significance between R7041(DUS3) and wt HSV-

1 or R264 is indicated by b (po0.05), b0 (po0.01), b00 (po0.001); n¼5.

P. Wild et al. / Virology 432 (2012) 353–360358

area of all R7041(DUS3) virions in the PNS was measured to be300 mm2 equalling 2400 virions present in a single cell around thenucleus with a mean volume of �1000 mm3 and �550 mm2

surface area. Volume and surface area were calculated on thebasis of the means of the 3 axes (a b, c) of the ovoid-shapednucleus measured to be 9, 7, and 3.9 mm. In a single image (Fig. 1)taken from a section cut perpendicular to the c-axis �150 virionsare present. If the entire nucleus is serially sectioned in distancesof 200 nm, like the viral diameter, �20 sections will be produced.Assuming that the average number of virions is �150 per section,the total number would be 3000 virions. Therefore, the number ofvirions calculated on the basis of the surface area of the viralenvelopes seems to be fairly appropriate. This high number ofvirions confirms the idea that folds of nuclear membranes arerather unlikely to be caused by fusion of the viral envelope withthe ONM. It also implies that the budding process at the INMtakes place at a large scale in the absence of US3.

In conclusion, labeling of phospholipid precursors in combina-tion with morphometric analysis provided clear evidence thatHSV-1 induces phospholipid synthesis for maintaining membraneintegrity of expanding nuclei early in infection, to providemembranes for capsid envelopment at the nuclear level as shownin detail recently (Sutter et al., 2012), and to provide membranesfor envelopment and vacuole formation by the Golgi complex. Theoverproduction of both nuclear and Golgi membranes inR7041(DUS3) infected cells suggests that US3 plays a significantrole in down-regulation of membrane biosynthesis. The accumu-lation of virions within the PNS proved that translocation ofcapsids into the cytoplasm is impaired raising the question howand where R7041(DUS3) virions become infective. Further studiesare needed to clarify regulation and mechanisms of phospholipid

synthesis during HSV-1 infection, and to clarify the significance ofUS3 on the functionality of the Golgi complex.

Material and methods

Cells and viruses

Vero cells (European Collection of Cell Cultures) were grown inDulbecco’s modified minimal essential medium (DMEM; Gibco,Bethesda, MD) supplemented with penicillin (100 U/ml), strepto-mycin (100 mg/ml) and 10% fetal calf serum (FCS; Gibco). The US3deletion mutant R7041(DUS3) and its repair mutant R2641 werekindly provided by B. Roizman, University of Chicago, IL, USA.Wild-type HSV-1 strain F, R7041(DUS3) and R2641 were propa-gated in Vero cells.

Infection of cells

Vero cells grown for 2 day were inoculated with R7041(DUS3),wt HSV-1 or the repair mutant R2641 at a multiplicity of infection(MOI) of 5 PFU per cell, and kept at 37 1C for 1 h to allowadsorption prior to incubation at 37 1C for 9, 12, 16, 20 or 24 h.

[3H]-choline incorporation

Cells were grown in 75 cm2 tissue culture flasks for 48 h priorto inoculation with virus in low choline content (1 mg/l) Eagle’sminimal essential medium (EMEM; Bioconcept, Allschwil, Swit-zerland). For assessment of [3H]-choline incorporation into cyto-plasmic and nuclear membranes, medium was replaced by 4 mlmedium containing 1 mCi/ml [3H]-choline (PerkinElmers, Wal-tham, MA, USA) 4 h prior to harvesting. Cells were harvestedusing trypsin at 9, 12, 16, 20 or 24 hpi for lipid extraction fromboth the nuclear and cytoplasmic fraction to determine [3H]-choline concentration by scintillation counting. 25 ml aliquotswere removed to determine the actual cell number for normal-ization of counts.

Separation of nuclear and cytoplasmic fractions

Cells grown 75 cm2 tissue culture flasks were detached usingtrypsin and collected in 10 ml medium. Nuclei were isolated in ice-cold solutions using digitonin to enhance fractionation according toBronfmanet al. (1998). Cells were centrifuged at 200 g for 10 min,resuspended in 1.5 ml PBS, transferred to Eppendorf tubes, andcentrifuged in a mini spin Eppendorf centrifuge operated at max-imal speed (10,000 g) for 30 s. The pellet was resuspended in 600 mlbuffer A (0.25 M sucrose and 3 mM imidazole, pH 7.4) containing1 mg/ml digitonin, and immediately centrifuged at maximal speedfor 30 s. The pellet was again resuspended in 600 ml buffer A andtransferred to a Dounce homogenizer. After 6 passes, the suspen-sion was transferred into Eppendorf tubes and centrifuged atmaximal speed for 30 s. The supernatant containing the cytoplas-mic fraction was used for lipid extraction. The pellet containing thenuclear fraction was resuspended in 600 ml H2O for lipid extraction.10 ml aliquots were used to test by phase contrast microscopywhether separation of nuclei was successful.

To assess the presence of nuclear membranes and the qualityof the isolation procedure, the nuclear fraction was controlled byelectron microscopy (Fig. 2A) employing a conventional fixationprotocol including fixation with 2.5% glutaraldehyde in 0.1 MNa–K phosphate, pH 7.4, at 4 1C for 1 h followed by postfixationwith 1% osmium tetroxide in 0.1 M Na/K-phosphate, pH 7.4, at4 1C, dehydration in a graded series of ethanol followed by2 changes in acetone, and embedded in Epon followed by

P. Wild et al. / Virology 432 (2012) 353–360 359

polymerization at 60 1C for 2.5 day. Sections of 50 to 60 nmthickness were stained with uranyl acetate and lead citrate, andanalyzed in a transmission electron microscope (CM12; Philips,Eindhoven, The Netherlands) equipped with a slow-scan CCDcamera (Gatan, Pleasanton, CA, USA) at an acceleration voltage of100 kV. To control the cytoplasmic fraction, 10 ml aliquots wereused for negative staining with Na-phosphotungstic acid, pH 7.0, for1 min after adsorption for 10 min to carbon coated parlodion filmsmounted on 300 mesh/inch copper grids, as described (Wild, 2008),and examined in a Philips CM12 electron microscope. To determinethe approximate number of viral particles present in the nuclearfraction and cytoplasmic fraction, respectively, we isolated the virusparticles from the two fractions by low speed centrifugation(10,000 g for 30 min) followed by centrifugation at 100,000 g for2 h. The pellet was resolved in 50 ml PBS. 10 ml aliquots were usedfor negative staining with Na-phosphotungstic acid as described(Wild, 2008).

Lipid extraction for determination of radioactivity

Lipids were extracted at room temperature according to theprocedure described by Bligh and Dyer (Bligh, 1959). 500 mlchloroform and 1 ml methanol were added to 300 ml of suspen-sions (nuclear fraction or cytoplasmic fraction) in glass tubes,vortexed 4 times within 10 min, and kept at room temperaturefor at least 30 min. After adding 500 ml chloroform and vortexingfor 1 min, 500 ml of distilled water was added, the samplevortexed again for 1 min, and then centrifuged at 800 g for10 min to separate the lipid containing solvent phase from theaqueous phase. The solvent phase was collected in Eppendorf tubesthat were placed on a warm plate (40 1C) in a safety hood forevaporation of the solvent. Then, lipids were dissolved in 50 mlchloroform/methanol (4:1) immediately prior to using 5 ml aliquotsfor scintillation counting (LS 6500, Beckman Coulter, Fullerton, CA,USA) according to the recommendation of the manufacturer.

Cryo-fixation for transmission electron microscopy

Cells grown on sapphire disks placed in 6-well-plates werefrozen in a high-pressure freezing unit (HPM010; BAL-TEC Inc.,Balzers, Liechtenstein) as described in detail (Wild et al., 2002).The samples were then transferred to a freeze-substitution unit(FS 7500; Boeckeler Instruments, Tucson, Arizona) precooledto �88 1C for substitution with acetone and subsequent fixationwith 0.25% glutaraldehyde and 0.5% osmium tetroxide rising thetemperatures gradually to þ2 1C to achieve good contrast ofmembranes (Wild et al., 2001), and embedded in Epon at 4 1Cfollowed by polymerization at 60 1C for 2.5 day. Sections of 50 to60 nm thickness were stained with uranyl acetate and leadcitrate, and analyzed by TEM at an acceleration voltage of 100 kV.

Confocal microscopy

Cells grown on cover slips were briefly washed with PBS prior tofixation with 2% formaldehyde for 25 min at room temperature,washed with cold PBS and kept at 4 1C until further processing.Then, cells were permeabilized with 0.1% Triton-X-100 at roomtemperature for 7 min and blocked with 3% bovine serum albuminin phosphate-buffered saline (PBS) containing 0.05% Tween 20(PBST). To ascertain infectivity, cells were labeled with antibodiesagainst the tegument protein VP16 (gift from B. Roizman), and antirabbit Alexa 594 (Life Technologies) as secondary antibodies. Afterstaining the nuclei with 40,6-Diamidino-2-phenylindol (DAPI;Roche, Mannheim, Germany) cells were embedded in glycergelmounting media (Dako North America, Carpinteria, CA, USA) and25 mg/ml DABCO (1,4-diazabicyclo [2.2.2] octane; Fluka, Buchs,

Switzerland). For calculation of nuclear size cells images were takenusing a confocal laser scanning microscope (SP2, Leica, Wetzlar,Germany).

Morphometric analysis

Nuclei of Vero cells are triaxial ellipsoids. Therefore, meannuclear volume (Vn) and mean nuclear surface area (Sn) werecalculated on the basis of the half axes (a, b, c) measured on 25deconvolved confocal DAPI stained images according to the equa-tions:

V ¼4

3pabc and S¼ 2pc2þ2pab

Z 1

0

1�u2v2x2ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1�u2x2p ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

1�v2x2p dx,

where by

u¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffia2�c2p

aand v¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffib2�c2

pb

:

On the basis of the absolute mean nuclear volume, theabsolute mean cell volume can be evaluated by electron micro-scopic morphometric analysis of thin sections. For this, 20 imageswere sampled at random per experiment on cryo-fixed cells of5 independent experiments at a final magnification of 6300� .From these images nuclear density (Vvn) and cytoplasmic density(Vvcy) were estimated applying the point counting method(Weibel, 1979) according to the equations Vvn¼Pn/(PnþPcy) andVvcy¼Pcy/(PnþPcy), whereby Pn are points hitting the nucleus, andPcy are points hitting the cytoplasm. On the basis of the absolutenuclear volume obtained from confocal images, the absolutemean cell volume (Vc) was calculated as Vc¼Vn/Vvn.

To determine changes in the size of membrane compartmentsincluding INM, ONM, viral envelope, Golgi complex and vacuolesimages were collected at a final magnification of 87’500� . Onthese images the surface density of folds of the INM (Svi) the ONM(Svo) and the envelope of virions (e) in the PNS (Sve) wereestimated using the equations Svi,o,e¼4Ii,o,e/d� Pn, whereby Ii,o,e

are the number of intersections of the test lines d with the INM,ONM and viral envelope, respectively. From the surface density,the area of membrane folds and viral envelope were calculatedper mean nuclear volume: Si,¼Svi�Vn, So,¼Svo�Vn and Se,¼

Sve�Vn. Similarly, the surface density of the Golgi complex (Svg)was determined on images taken at random at a final magnifica-tion of 115,000 from 5 experiments applying the equation Svg¼ Ig/d� Pcy, whereby Ig are intersections of the test lines with Golgimembranes, and Pcy are points hitting the cytoplasm. The meansurface area of the Golgi complex (Sg) was calculated per meancell volume using the equation Sg¼Svg�Vvcy�Vc.

Statistical analysis

Mean and standard deviation of [3H]-uptake and of morpho-metric data were calculated and compared by a multiple t testusing GraphPad Prism 3.

Acknowledgments

The authors thank B. Roizman for critical discussions and forproviding the deletion mutant R7041(DUS3), the repair mutantR2641 and antibodies, Martin Mueller and R. Wepf for providingthe facilities for high-pressure freezing. This study was supportedby the Foundation for Scientific Research at the University ofZurich, Switzerland.

P. Wild et al. / Virology 432 (2012) 353–360360

References

Asher, Y., Heller, M., Becker, Y., 1969. Incorporation of lipids into herpes simplexvirus particles. J. Gen. Virol. 4 (1), 65–76.

Ben-Porat, T., Kaplan, A.S., 1971. Phospholipid metabolism of herpesvirus-infectedand uninfected rabbit kidney cells. Virology 45 (1), 252–264.

Benetti, L., Munger, J., Roizman, B., 2003. The herpes simplex virus 1 US3 proteinkinase blocks caspase-dependent double cleavage and activation of theproapoptotic protein BAD. J. Virol. 77 (11), 6567–6573.

Bligh, 1959. A rapid method of total lipid extraction and purification. Can. J.Biochem. Physiol. 37 (8), 911–917.

Borchers, K., Oezel, M., 1993. Simian agent 8 (SA8): morphogenesis and ultra-structure. Zentralbl. Bakteriol. 279, 526–536.

Bronfman, M., Loyola, G., Koenig, C.S., 1998. Isolation of intact organelles bydifferential centrifugation of digitonin-treated hepatocytes using a tableEppendorf centrifuge. Anal. Biochem. 255 (2), 252–256.

Galvan, V., Roizman, B., 1998. Herpes simplex virus 1 induces and blocks apoptosisat multiple steps during infection and protects cells from exogenousinducers in a cell-type-dependent manner. Proc. Nat. Acad. Sci. U.S.A. 95 (7),3931–3936.

Grunewald, K., Desai, P., Winkler, D.C., Heymann, J.B., Belnap, D.M., Baumeister,W., Steven, A.C., 2003. Three-dimensional structure of herpes simplex virusfrom cryo-electron tomography. Science 302 (5649), 1396–1398.

Henneberry, A.L., Wright, M.M., McMaster, C.R., 2002. The major sites of cellularphospholipid synthesis and molecular determinants of Fatty Acid and lipidhead group specificity. Mol. Biol. Cell 13 (9), 3148–3161.

Homman-Loudiyi, M., Hultenby, K., Britt, W., Soderberg-Naucler, C., 2003. Envel-opment of human cytomegalovirus occurs by budding into Golgi-derivedvacuole compartments positive for gB, Rab 3, trans-golgi network 46, andmannosidase II. J. Virol. 77 (5), 3191–3203.

Jerome, K.R., Fox, R., Chen, Z., Sears, A.E., Lee, H., Corey, L., 1999. Herpes simplexvirus inhibits apoptosis through the action of two genes, Us5 and Us3. J. Virol.73 (11), 8950–8957.

Klupp, B.G., Granzow, H., Mettenleiter, T.C., 2011. Nuclear envelope breakdowncan substitute for primary envelopment-mediated nuclear egress of herpes-viruses. J. Virol. 85 (16), 8285–8292.

Leopardi, R., Van Sant, C., Roizman, B., 1997. The herpes simplex virus 1 proteinkinase US3 is required for protection from apoptosis induced by the virus.Proc. Nat. Acad. Sci. U.S.A. 94 (15), 7891–7896.

Leuzinger, H., Ziegler, U., Schraner, E.M., Fraefel, C., Glauser, D.L., Heid, I.,Ackermann, M., Mueller, M., Wild, P., 2005. Herpes simplex virus 1 envelop-ment follows two diverse pathways. J. Virol. 79 (20), 13047–13059.

Mettenleiter, T.C., 2002. Herpesvirus assembly and egress. J. Virol. 76 (4),1537–1547.

Mettenleiter, T.C., 2004. Budding events in herpesvirus morphogenesis. Virus Res.106 (2), 167–180.

Mettenleiter, T.C., Klupp, B.G., Granzow, H., 2006. Herpesvirus assembly: a tale oftwo membranes. Cur. Opin. Microbiol. 9 (4), 423–429.

Monier, K., Armas, J.C., Etteldorf, S., Ghazal, P., Sullivan, K.F., 2000. Annexation ofthe interchromosomal space during viral infection. Nat. Cell Biol. 2 (9),661–665.

Mou, F., Wills, E., Baines, J.D., 2009. Phosphorylation of the U(L)31 protein ofherpes simplex virus 1 by the U(S)3-encoded kinase regulates localization ofthe nuclear envelopment complex and egress of nucleocapsids. J. Virol. 83(10), 5181–5191.

Mou, F., Wills, E.G., Park, R., Baines, J.D., 2008. Effects of lamin A/C, lamin B1, andviral US3 kinase activity on viral infectivity, virion egress, and the targeting ofherpes simplex virus U(L)34-encoded protein to the inner nuclear membrane.J. Virol. 82 (16), 8094–8104.

Munger, J., Chee, A.V., Roizman, B., 2001. The U(S)3 protein kinase blocks apoptosisinduced by the d120 mutant of herpes simplex virus 1 at a premitochondrialstage. J. Virol. 75 (12), 5491–5497.

Munger, J., Roizman, B., 2001. The US3 protein kinase of herpes simplex virus1 mediates the posttranslational modification of BAD and prevents BAD-induced programmed cell death in the absence of other viral proteins. Proc.Nat. Acad. Sci. U.S.A. 98 (18), 10410–10415.

Ogg, P.D., McDonell, P.J., Ryckman, B.J., Knudson, C.M., Roller, R.J., 2004. The HSV-1Us3 protein kinase is sufficient to block apoptosis induced by overexpressionof a variety of Bcl-2 family members. Virology 319 (2), 212–224.

Orci, L., Montesano, R., Perrelet, A., 1981. Exocytosis-endocytosis as seen withmorphological probes of membrane organization. Meth. Cell Biol. 23, 283–300.

Poon, A.P., Benetti, L., Roizman, B., 2006. U(S)3 and U(S)3.5 protein kinases ofherpes simplex virus 1 differ with respect to their functions in blockingapoptosis and in virion maturation and egress. J. Virol. 80 (8), 3752–3764.

Purves, F.C., Longnecker, R.M., Leader, D.P., Roizman, B., 1987. Herpes simplexvirus 1 protein kinase is encoded by open reading frame US3 which is notessential for virus growth in cell culture. J. Virol. 61 (9), 2896–2901.

Purves, F.C., Spector, D., Roizman, B., 1991. The herpes simplex virus 1 proteinkinase encoded by the US3 gene mediates posttranslational modification ofthe phosphoprotein encoded by the UL34 gene. J. Virol. 65 (11), 5757–5764.

Quinlan, M.P., Chen, L.B., Knipe, D.M., 1984. The intranuclear location of a herpessimplex virus DNA-binding protein is determined by the status of viral DNAreplication. Cell 36 (4), 857–868.

Reynolds, A.E., Liang, L., Baines, J.D., 2004. Conformational changes in the nuclearlamina induced by herpes simplex virus type 1 require genes U(L)31 andU(L)34. J. Virol. 78 (11), 5564–5575.

Reynolds, A.E., Ryckman, B.J., Baines, J.D., Zhou, Y., Liang, L., Roller, R.J., 2001.U(L)31 and U(L)34 proteins of herpes simplex virus type 1 form a complex thataccumulates at the nuclear rim and is required for envelopment of nucleo-capsids. J. Virol. 75 (18), 8803–8817.

Reynolds, A.E., Wills, E.G., Roller, R.J., Ryckman, B.J., Baines, J.D., 2002. Ultrastruc-tural localization of the herpes simplex virus type 1 UL31, UL34, and US3proteins suggests specific roles in primary envelopment and egress ofnucleocapsids. J. Virol. 76 (17), 8939–8952.

Roizman, B., Knipe, D.M., 2001. Herpes simplex viruses and their replication.fourth ed. In: Howly, K.D.M.,.P.M. (Ed.), Fields Virology, Vol. 2. Lipincott-RavenPublishers, Philadelphia, pp. 2399–2460.

Roizman, B., Knipe, D.M., Whiley, R.J., 2007. Herpes simplex viruses. fifth ed. In:Howly, K.D.M.,.P.M. (Ed.), Fields Virology, Vol. 2. Lipincott-Raven Publishers,Philadelphia, pp. 2501–2601 2 vols.

Simpson-Holley, M., Baines, J., Roller, R., Knipe, D.M., 2004. Herpes simplex virus1 U(L)31 and U(L)34 gene products promote the late maturation of viralreplication compartments to the nuclear periphery. J. Virol. 78 (11),5591–5600.

Simpson-Holley, M., Colgrove, R.C., Nalepa, G., Harper, J.W., Knipe, D.M., 2005.Identification and functional evaluation of cellular and viral factors involved inthe alteration of nuclear architecture during herpes simplex virus 1 infection.J. Virol. 79 (20), 12840–12851.

Stannard, L.M., Himmelhoch, S., Wynchank, S., 1996. Intra-nuclear localizationof two envelope proteins, gB and gD, of herpes simplex virus. Arch. Virol. 141(3–4), 505–524.

Sutter, E., deOliveira, A.P., Tobler, K., Schraner, E.M., Sonda, S., Kaech, A., Lucas,M.S., Ackermann, M., Wild, P., 2012. Herpes Simplex Virus 1 Induces de novoPhospholipid Synthesis. Virology 429, 124–135.

van Meer, G., Voelker, D.R., Feigenson, G.W., 2008. Membrane lipids: where theyare and how they behave. Nat. Rev. Mol. Cell Biol. 9 (2), 112–124.

Weibel, E., 1979. Stereological Methods, Vol. I. Practical Methods for BiologicalMorphometry. Academic Press, London.

Wild, P., 2008. Electron Microscopy of Viruses and Virus-Cell Interactions; T. Allen.In ‘‘Introduction to Electron Microscopy for Biologists; Meth. Cell Biol., Else-vier’’, Vol. 88, pp. 497–524.

Wild, P., Engels, M., Senn, C., Tobler, K., Ziegler, U., Schraner, E.M., Loepfe, E.,Ackermann, M., Mueller, M., Walther, P., 2005. Impairment of nuclear pores inbovine herpesvirus 1-infected MDBK cells. J. Virol. 79 (2), 1071–1083.

Wild, P., Schraner, E.M., Adler, H., Humbel, B.M., 2001. Enhanced resolution ofmembranes in cultured cells by cryoimmobilization and freeze-substitution.Microsc. Res. Techn. 53 (4), 313–321.

Wild, P., Schraner, E.M., Cantieni, D., Loepfe, E., Walther, P., Muller, M., Engels, M.,2002. The significance of the Golgi complex in envelopment of bovineherpesvirus 1 (BHV-1) as revealed by cryobased electron microscopy. Micron33 (4), 327–337.

Wild, P., Senn, C., Manera, C.L., Sutter, E., Schraner, E.M., Tobler, K., Ackermann, M.,Ziegler, U., Lucas, M.S., Kaech, A., 2009. Exploring the nuclear envelope ofherpes simplex virus 1-infected cells by high-resolution microscopy. J. Virol.83 (1), 408–419.

Wisner, T.W., Wright, C.C., Kato, A., Kawaguchi, Y., Mou, F., Baines, J.D., Roller, R.J.,Johnson, D.C., 2009. Herpesvirus gB-induced fusion between the virionenvelope and outer nuclear membrane during virus egress is regulated bythe viral US3 kinase. J. Virol. 83 (7), 3115–3126.

Zinser, E., Sperka-Gottlieb, C.D., Fasch, E.V., Kohlwein, S.D., Paltauf, F., Daum, G.,1991. Phospholipid synthesis and lipid composition of subcellular membranesin the unicellular eukaryote Saccharomyces cerevisiae. J. Bacteriol. 173 (6),2026–2034.