Skip to main page content

• HOME
• Current Issue
• Archive
• Alerts
• About ASM
• Contact us
• Tech Support
• Journals.ASM.org

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Misaghi, S. Articles by Ploegh, H. Search for Related Content PubMed PubMed Citation Articles by Misaghi, S. Articles by Ploegh, H.

 Previous Article  |  Next Article 

Journal of Virology, January 2004, p. 413-423, Vol. 78, No. 1
0022-538X/04/$08.00+0     DOI: 10.1128/JVI.78.1.413-423.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.

Structural and Functional Analysis of Human Cytomegalovirus US3 Protein

Shahram Misaghi,^1, Zhen-Yu J. Sun,^2, Patrick Stern,¹ Rachelle Gaudet,³ Gerhard Wagner,² and Hidde Ploegh^1*

Department of Pathology,¹ Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical School, Boston, Massachusetts 02115,² Department of Molecular and Cellular Biology, Harvard University, Cambridge, Massachusetts 02138³

Received 28 July 2003/ Accepted 22 September 2003

Top Abstract Introduction Materials and Methods Results Discussion References

 
Human cytomegalovirus (HCMV) unique short region 3 (US3) protein, a type I membrane protein, prevents maturation of class I major histocompatibility complex (MHC) molecules by retaining them in the endoplasmic reticulum (ER) and thus helps inhibit antigen presentation to cytotoxic T cells. US3 molecules bind to class I MHC molecules in a transient fashion but retain them very efficiently in the ER nonetheless. The US3 luminal domain is responsible for ER retention of US3 itself, while both the US3 luminal and transmembrane domains are necessary for retaining class I MHC in the ER. We have expressed the luminal domain of US3 molecule in Escherichia coli and analyzed its secondary structure by using nuclear magnetic resonance. We then predicted the US3 tertiary structure by modeling it based on the US2 structure. Unlike the luminal domain of US2, the US3 luminal domain does not obviously interact with class I MHC molecules. The luminal domain of US3 dynamically oligomerizes in vitro and full-length US3 molecules associate with each other in vivo. We present a model depicting how dynamic oligomerization of US3 may enhance its ability to retain class I molecules within the ER.

Top Abstract Introduction Materials and Methods Results Discussion References

 
Major histocompatibility complex (MHC) molecules activate the immune system by presenting fragments of foreign proteins to T lymphocytes. After the interaction of class I MHC molecules with T-cell receptors of the appropriate specificity, the activated T cells may induce the destruction of the infected cell. Activation of T cells also releases different cytokines that trigger and use other components of the immune system to eliminate the pathogen. Despite this rather complex and multilayered defense, some viruses have evolved the means to evade immune detection by disabling peptide presentation by class I MHC molecules. Human cytomegalovirus (HCMV), a member of the Herpesviridae family, downregulates surface expression of class I MHC molecules (for a review, see reference 24). Although HCMV can be well controlled by the immune system of a healthy individual, it causes serious disease in congenitally infected infants and in immunocompromised/immunosuppressed adults.

HCMV encodes at least four proteins that interfere with surface expression of class I MHC molecules. These proteins are encoded in the unique short (US) region of the HCMV genome, in particular US2, US3, US6, and US11. US2 and US11 target class I molecules from the endoplasmic reticulum (ER) to the cytoplasm for proteasomal degradation. US6 blocks the transport of peptides into the ER via the transporter associated with antigen presentation (TAP) complex. US3 is the product of an immediate-early gene. US3 retains the class I MHC molecules in the ER and prevents their transport through the Golgi and subsequently to cell surface (9, 24). US3 is expressed at a very early stage of HCMV infection and constitutes an early means of viral evasion from the host's immune system. US3 is a 186-amino-acid type I membrane protein, comprised of an ER-luminal domain bearing a single N-linked glycan at Asn 60, a single transmembrane domain and a short cytoplasmic tail (1). Domain swap experiments have shown that the luminal domain of US3 is sufficient for ER retention of US3 itself, whereas both luminal and transmembrane domains are required for retention of class I MHC molecules in the ER (15).

Glycoproteins, such as class I MHC molecules, are exported from the ER to the Golgi complex in vesicles coated with coatomer complex COP II (21). However, many of the ER-resident proteins utilize specific retrieval signals to recycle from the Golgi complex back to the ER. Two such retrieval mechanisms have been identified. The sequence KDEL is a common tetrapeptide signal at the carboxy terminus of some ER resident proteins (19) and is recognized by KDEL receptor (erd2), a protein involved in returning escaped proteins back to the ER (16, 17). The KKXX motif is recognized by the COP I coatomer complex, which mediates retrograde transport of vesicles from the Golgi to the ER (23). Although US3 is apparently capable of retaining itself in the ER, it does not utilize either of the two known retrieval mechanisms. To stay in the ER, US3 likely binds to ER resident protein(s), the identity of which is not known.

Interaction of US3 with class I molecules is transient and iterative. Newly synthesized US3 molecules rapidly replace those already bound (10). However, these transient interactions suffice to prevent class I MHC molecules from forward transport to the Golgi. Eventually, some of the US3 molecules escape the ER and are destroyed, perhaps in lysosomal compartments (10). Ser58, Glu63, and Lys64 in the luminal domain of US3 are required for its ER retention. Mutation of any of these residues to alanine causes localization of US3 molecules to the cis-Golgi, based on immunofluorescence analysis (14). Although the mutant US3 molecules remained endo-ß-N-acetylglucosaminidase H (Endo H) sensitive, they lost their ability to suppress surface expression of class I MHC molecules. The S/EK sequence may be recognized by a putative receptor located within the lumen of the ER (14).

We have previously reported the structure of luminal domain of US2 in a complex with the luminal domain of a peptide-loaded class I MHC molecule (HLA-A2) (8). For these luminal domains, the class I-US2 interaction is strong and survives native electrophoresis, gel filtration, and other chromatographic separations. US2 has an immunoglobulin-like fold composed of seven ß-strands arranged into a ß-sandwich structure (7, 8). Sequence alignment suggests that US3 may have an immunoglobulin-like fold similar to US2, but US2 and US3 have little sequence similarity (20% identity). Since immunoglobulin-like folds mediate contact with other protein surfaces, this structure may be utilized by US3 to retain itself in the ER or to otherwise facilitate its function. Here, we analyzed the structure of luminal portion of US3 by nuclear magnetic resonance (NMR). We determined the secondary structure of US3's luminal domain and predicted its tertiary structure by modeling based on the structure of US2. We show that the luminal domain of US3 does not interact with peptide-loaded soluble HLA-A2, even at very high concentrations of the interacting partners. Despite overall structural similarity, the mode of interaction with class I MHC molecules must therefore be very different for US3 and US2. Unexpectedly, US3 is capable of oligomerization in vitro and in vivo and its luminal domain is probably responsible for this behavior. Based on these results we present a model to explain how different properties of US3 may be involved in carrying out its function.

Top Abstract Introduction Materials and Methods Results Discussion References

 
Plasmid constructs. A DNA fragment corresponding to the luminal domain of US3 (residues 30 to 136) DNA was amplified from a DNA construct containing the full-length US3 molecule (12) as a template for a PCR. The proper oligonucleotides contained NcoI and XhoI restriction sites at their 5' and 3' ends, respectively. Amplified fragments were digested with NcoI and XhoI and ligated into the NcoI- and XhoI-digested pET-28a(+) vector (Novagen) in frame with the hexahistidine tag (US3L-His₆). We used the primers CATGCCATGGACTCGGAGATCAGATCGGCC (sense) and CCGCTCGAGGTCCATCCTCACCTGAGGAAC (antisense) in this PCR.

Full-length US3 with hemagglutinin (HA) tag (US3-HA) DNA construct was made by PCR by using the original US3 construct (12) as a template. The PCR constructs were flanked by EcoRI and XbaI sites at the 5' and 3' ends, respectively. After digestion, the oligonucleotide fragments were cloned into pcDNA3.1(+) (Invitrogen) and pLNW retroviral vector. We used the primers CCGAATTCACCACCATGAAGCCGGTGTTGGTG (sense) and CTAGTCTAGATTAGTCGAGTGCGTAGTCTGGTACGTCGTACGGATACTCGAGAATAAATCGCAGACGGGCGCTC (antisense) in this PCR.

A standard site-directed mutagenesis protocol (Stratagene) was used to generate the E63A/K64A-US3 with an HA tag (US3-HA*) construct, which was subcloned into the pcDNA3.1(+) and pLNW vectors. The following primers were used for this protocol (substituted codons are underlined): sense (TCGGGCAACTTCACCGCGGCACACTTTGTGAACGTG) and antisense (CACGTTCACAAAGTGTGCCGCGGTGAAGTTGCCCGA). The accuracy of the all constructs was confirmed by DNA sequencing.

Protein expression and purification. BL21(DE3) Escherichia coli bacteria were transformed with US3L-His₆ plasmid. A single colony was grown overnight in 20 ml of M9 medium containing 30 µg of kanamycin. This was used to inoculate 2 liters of M9-kanamycin medium made in D₂O (Cambridge Isotope Labs) and supplemented with 2 g of [¹³C]glucose and 1 g of ¹⁵NH₄Cl/liter for isotope labeling (Cambridge Isotope Labs). This culture was grown at 37°C to an optical density at 600 nm of 0.6 and was induced by adding 1 mM IPTG (isopropyl-ß-D-thiogalactopyranoside). The cells were harvested after 12 h and lysed under denaturing conditions in 40 ml of buffer A (10 mM Tris [pH 8], 100 mM NaCl, 7 M urea) in a French press. The lysate was spun for 15 min at 12,000 rpm, and the supernatant was filtered (0.45-µm pore size) and applied to a 10 ml of Ni²⁺-charged nitrilotriacetic acid column (Qiagen) equilibrated with binding buffer (buffer A plus 5 mM imidazole). The column was washed with 10 column volumes of binding buffer, followed by 6 column volumes of wash buffer (buffer A plus 13 mM imidazole). US3L-His₆ was eluted with 5 column volumes of elution buffer (buffer A plus 500 mM imidazole). The protein was refolded by two-step dialysis in buffer B (10 mM Tris [pH 8], 350 mM NaCl, 1 mM EDTA [pH 8.0]) containing 3.5 and 0.8 M urea at successive steps. The refolded protein was concentrated by Centriprep-10 spin columns (Amicon) and further purified by Superdex-200-Prep-Grade 16/600 sizing column (Pharmacia). Buffer C (20 mM NaPO₄ [pH 7.75], 190 mM NaCl) was used to elute US3L-His₆ and deuterated glycerol (Cambridge Isotope Labs) was added to a final concentration of 5%. These fractions were concentrated to ca. 0.3 mM and used for NMR studies. HLA-A2/TAX trimer was expressed and purified as described previously (6, 9), concentrated to 0.17 mM, and used in a 1:1 molar ratio with ¹⁵N-labeled US3L-His₆ for NMR interaction experiments.

NMR spectroscopy and structure determination. Standard triple-resonance NMR experiments (5) were carried out on a Bruker 500 MHz, and a Bruker 600 MHz spectrometer was used for the assignment of backbone atom chemical shifts by using U-¹³C-¹⁵N two-dimensionally labeled US3L-His₆ protein samples. The NMR samples contain 5% deuterated glycerol, 10% D₂O, 20 mM NaPO₄, and 190 mM NaCl at pH 7.75. TROSY data were also acquired for HNCACB and HNCOCACB experiments on Varian 750 MHz spectrometer. The protein concentrations were ca. 0.3 mM, and the experiments were done at 12°C to extend protein stability in solution. The NMR data were analyzed by using the software XEASY (2), and assisted with the automatic assignment software IBIS (11). The two-dimensional ¹⁵N-¹H correlated HSQC NMR experiments studying interaction between US3L-His₆ and class I MHC molecule were carried out on a Varian 500-MHz spectrometer at 12°C.

Cell lines and antibodies. The human embryonic kidney cell line (HEK-293), U373-MG astrocytoma cells, US3+/U373-MG cells (US3+), US3-HA/U373-MG cells (US3-HA), and US3-HA*/U373-MG cells (US3-HA*) were transfected, infected, selected, and maintained as described previously (20). Neomycin at 0.5 mg/ml was used to select infected cells. Rabbit polyclonal antisera against HCMV US3 protein and antisera against class I MHC free heavy-chain molecules (HC) have been described (3, 12). Anti-Sec61ß antibody was made by conjugating a peptide corresponding to the N-terminal nine residues of Sec61ß, equipped with an extra cysteine (PGPTPSGTNC), to keyhole limpet hemocyanin and then using it as an antigen for raising rabbit polyclonal antibody.

Metabolic labeling, immunoprecipitation, cross-linking, and light-scattering experiments. Cells were metabolically labeled with 500 µCi of [³⁵S]methionine-cysteine (NEN-Dupont), lysed in digitonin or NP-40 buffers, and immunoprecipitated as described previously (12, 20). For in vivo cross-linking experiments, dithiobis(sulfosuccinimidylpropionate) (DSP; Pierce) was used. For every sample, about 4 to 5 million U373-MG cells were labeled for 1 h. Cells were washed once in cold phosphate-buffered saline (PBS) and resuspended in 0.5 ml of cold PBS. DSP (stock of 25 mM) was added to the final concentration of 750 µM, and the samples were incubated for 1 h on ice. The control group did not receive any DSP. Tris-HCl (pH 7.0) was added to each sample to a final concentration of 20 mM to inactivate the remaining DSP. Cells were centrifuged (5 min, 4,000 rpm), the supernatant was removed, and the cells were lysed as described above. For in vitro cross-linking, we used bis(sulfosuccinimidyl)suberate (BS³) as our cross-linker (Sigma). US3L-His₆ from the NMR experiments was diluted to 1 mg/ml with PBS containing 5% glycerol. A total of 25 µl of the protein was then incubated for 2 h on ice with 0, 10, 100, or 1,000 µM BS³. The cross-linking reaction was terminated by the addition of 2 µl of 1 M Tris-HCl (pH 7.5), and the samples were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). For dynamic light scattering, we used a Dynamics v.5.26.56 instrument from Protein Solutions. For multiangle light-scattering experiments, we used a Superdex-200-HR-10/30 sizing column on an AKTA-FPLC system (Amersham) coupled to a DAWN EOS light-scattering instrument (Wyatt).

Top Abstract Introduction Materials and Methods Results Discussion References

 
Purification and refolding of the luminal domain of US3. US3 is a type I membrane glycoprotein that consists of four segments: an N-terminal signal peptide, a glycosylated luminal domain, a transmembrane domain, and a short cytoplasmic tail (Fig. 1a). To study the structure of US3, its luminal domain, from residues 30 to 136, corresponding roughly to the crystallized luminal domain of US2 (8) was cloned in frame with a His₆ tag in the pET28a(+) bacterial expression vector. We refer to this construct as US3L-His₆. US3L-His₆ was expressed in E. coli and purified under denaturing conditions as explained in Materials and Methods. The eluted protein was refolded by dialysis and subjected to further purification by size exclusion chromatography. US3L-His₆ protein eluted in a single peak at the expected molecular mass of 14 kDa. Peak fractions were pooled and concentrated for NMR spectroscopy. Figure 1b shows a silver-stained gel of the US3L-His₆ NMR sample.

View larger version (38K): [in this window] [in a new window]

FIG. 1. Structure of US3. (a) Kyte-Doolittle hydropathy plot of the US3 molecule, representing the signal peptide, luminal domain, transmembrane, and cytoplasmic tail domain. An index of 1 or higher represents a hydrophobic/transmembrane domain. (b) Silver-stained SDS-PAGE gel of US3L-His6 after size exclusion chromatography. This sample was used for NMR analysis. (c) Sequence alignment between US2 and US3. Identical residues are indicated in red; homologous residues are indicated in green. The yellow bars above the US2 sequence indicate the positions of the ß-strands. The circles below the US3 sequence indicate residues assigned by NMR; filled circles indicate residues in ß-sheet conformation as confirmed by NMR. The asterisks indicate US2 residues that are involved in binding to the class I MHC molecule HLA-A2.

Secondary structure of US3L-His₆ and its modeling. We have assigned ca. 70% of the backbone atom chemical shift values of US3L-His₆ by using standard triple-resonance NMR experiments, as well as by using the more recent TROSY-type NMR experiments. The N-terminal residues up to Arg33, and the C-terminal residues from Arg134 onward appear to be unstructured, since their peak intensities are larger and their peak positions are mostly in the central region of the NMR spectrum. The deviations of assigned chemical shift values relative to the random coil conformation values were analyzed by using the TALOS (4) software to predict the secondary structure of US3L-His₆. The results are shown in Fig. 1c. Of note, the US3 residues that are identified as ß-sheets are consistent with the sequence alignment of US2 ß-strands. Figure 2a shows schematic models comparing US2 and US3 secondary structures, with the assigned residues of the US3 molecule colored orange. The behavior of the US3L-His₆ molecule precluded a secondary structure assignment for the remaining 30% residues of the molecule, although the signals of the individual residues were present. Figure 2b illustrates a homology model of the secondary structure of US3L-His₆ based on sequence alignment and secondary structure analysis. Relative to US2, the US3L-His₆ luminal domain is extended at the ends due to the extensions in the FG and DE loops. The CC' loop is extended by one residue, and the three glycine residues from US2 are replaced with bulkier residues so that the CC' loop forms a bigger protrusion in US3. The N-linked glycosylation site (Asn60) is in the BC loop at the top of the molecule. The residues Ser58, Glu63, and Lys64 at the beginning and end of the BC loop when mutated to alanine affect the ER retention properties of US3.

View larger version (31K): [in this window] [in a new window]

FIG.2. NMR and modeling of US3L-His6. (a) Schematic diagram of US2 and US3 secondary structures. The ß-sheet assigned based on the US2 secondary structure are blue, whereas the assigned US3 segments are orange. The intramolecular disulfide bond is represented by a yellow line. (b) Ribbon structure representing a homology model of US3 based on the US2 structure. The left panel shows modeled US3 with the GFCC' face in front, and the right panel shows the same model rotated 90° clockwise. The residues S58, E63, and K64, mutation of which abolishes ER retention, are shown in blue. The glycosylation site N60 is shown in red, and the disulfide bond is shown in yellow. (c) NMR two-dimensional 15N-1H-correlated HSQC spectra comparing 15N-labeled US3L-His6 in the absence (left panel) or presence (right panel) of the class I MHC molecule HLA-A2 (0.17 mM). These spectra are essentially identical.

The luminal domain of US3 does not interact with class I MHC in vitro. The two-dimensional NMR ¹⁵N-¹H-correlated HSQC spectra of ¹⁵N-labeled US3L-His₆ were acquired in the presence or absence of class I MHC molecules (Fig. 2c). Each peak in the spectra represents the backbone amide hydrogen atom from each residue (or the nitrogen-bonded side chain hydrogen atoms from asparagine, glutamine, and tryptophan). Should binding occur between US3 and class I MHC, the peak positions of the residues involved in binding will change or disappear from the spectra due to severe peak broadening. As shown in Fig. 2c, there are no peak shifts or broadening of peaks for any US3 luminal-domain residues in the presence of the class I MHC molecule. We therefore conclude that the luminal domain of US3 does not interact with class I MHC in vitro, even at the concentrations of protein (0.17 mM) examined here.

US3 luminal domain reversibly oligomerizes in vitro. The two-dimensional NMR spectra in Fig. 2c show that the resonance peaks are well dispersed, indicating that the US3 luminal domain is in its native fold, but the spectra of the three-dimensional NMR experiments necessary for determining the complete atomic structure of US3 suffer from weak signals. This is surprising, considering that the molecular mass of US3L-His₆ is ca. 14 kDa. Even the TROSY experiments, usually productive for proteins with molecular masses from 30 to 50 kDa, gave poor signals for US3L-His₆. This result suggests that US3L-His₆ may exist as oligomers in solution. The situation is exacerbated by the presence of 5% glycerol and the temperature (12°C) required to improve the stability of the US3L-His₆ NMR sample. Figure 3a shows the electrostatic potential surface map of the US3 model derived from the NMR data. Compared to the surface map of US2 (Fig. 3b), it is clear that most charged residues in US3 are located in the upper half of the protein, whereas the lower half presents an almost neutral face that might be involved in oligomerization. After we completed nearly 70% of the backbone resonance assignments, we discovered that residues Gly84 and Gln86 in the C'D loop, residues Ile100, Gly101, and Gly102 at the lower E strand, and residues Val115 and Thr118 in the EF loop have at least two conformations. Interestingly, these residues are all in the lower half of US3 that could be responsible for oligomerization. It is conceivable that the multiple conformations of these residues could correspond to the different monomer and oligomer states.

View larger version (61K): [in this window] [in a new window]

FIG. 3. Surface electrostatic potential maps comparing the US3 model (a) and the US2 structure (b). The positively charged regions are blue, and the negatively charged regions are red. Figures on the left panel are in the same orientation as the ribbon structure shown in Fig. 2b (left). The surfaces on the right are rotated 180° with respect to the surfaces on the left. The figure was generated by using Molmol (13).

Since NMR analysis implies that US3L-His₆ molecules dynamically associate and dissociate, we decided to further analyze this behavior by using light scattering. The US3L-His₆ sample, used for the NMR analysis, was subjected to dynamic light-scattering measurement. The US3L-His₆ preparation was mostly monodisperse, but the observed molecular mass was more consistent with that of small oligomers (52 kDa, with an error rate of 30%). Since signals from larger oligomers obscure other signals, it was not possible to obtain an accurate molecular mass estimate with a low error for all of the oligomers (data not shown). However, subjecting US3L-His₆ to a sizing column coupled with multiangle light-scattering measurements yielded a single peak and showed its molecular mass to be 14.480 kDa with a percent error of 0.9% (Fig. 4a), the molecular mass of monomeric US3L-His₆. These results are consistent with the notion that US3L-His₆ molecules interact with each other and form oligomers, but their interaction is weak at the concentrations explored for gel filtration.

View larger version (27K): [in this window] [in a new window]

FIG. 4. In vitro light-scattering and cross-linking studies suggest that US3L-His6 oligomerizes in solution. (a) In the left panel is shown an elution profile of the concentrated US3L-His6, used in NMR spectroscopy, from the analytical sizing column. US3L-His6 elutes in a single peak with no evidence of aggregation. In the right panel is shown the multiangle light-scattering profile of the peak shown in the left panel. US3L-His6 scatters with molecular mass of 14.480 kDa with a 0.9% error rate, which is in agreement with the size of a monomer. (b) In vitro cross-linking of US3L-His6 in presence of 0, 10, 100, or 1,000 µM BS3. US3L-His6 at 1 mg/ml was incubated for 2 h with the indicated concentrations of BS3, and the samples were subjected to SDS-PAGE and Coomassie blue staining.

Next we sought to determine whether US3L-His₆ exists as an oligomer of a discrete size. We reasoned that if US3L-His₆ has a preferred oligomeric state, we should be able to stabilize and trap a specific multimer by chemical cross-linking. US3L-His₆ at 1 mg/ml was subjected to different concentrations of BS³ cross-linker for 2 h (Fig. 4b). At higher concentrations of the cross-linker, species corresponding to dimers, trimers, and tetramers of the US3L-His₆ were evident, but we did not detect a single predominant US3L-His₆ multimer. In addition, the majority of the US3L-His₆ molecules remained monomeric, again suggesting that the interaction between US3L-His₆ molecules themselves is weak. Overnight incubation of US3L-His₆ (at 1 mg/ml), with up to a 10 mM concentration of BS³, yielded similar results, with the additional presence of very faint extra bands corresponding to pentamers and hexamers.

Full-length US3 can oligomerize in vivo. Since the luminal domain of US3 oligomerizes in vitro, we sought to determine whether full-length US3 behaves in the same way. To address this question, we made a full-length version of US3 that carries an HA tag (US3-HA) at its carboxyl terminus in pcDNA3.1(+) vector and in the pLNW retroviral vector. If US3 oligomerizes in vivo, we should be able to retrieve US3 molecules in association with HA-tagged US3 by immunoprecipitation via anti-HA antibodies in cells that express both wild-type US3 and US3-HA molecules.

To show that US3-HA is functional and, like US3, is capable of downregulating class I MHC molecules, we expressed it via a retrovirus vector in U373 cells. US3-HA was recognized by both anti-US3 and anti-HA antibodies. There was a visible shift in its molecular mass due to the presence of the HA tag (Fig. 5a). To test its functionality, we performed a pulse-chase experiment on US3-HA-expressing cells and immunoprecipitated class I MHC molecules, with anti-HC antibody, followed by Endo H treatment. In control U373 cells, at the zero time point, all of the class I MHC heavy chains are Endo H sensitive and, after 90 min, acquire Endo H resistance (compare lanes 1 and 2 in Fig. 5b), a finding consistent with their passage from the ER through the Golgi. In contrast, in cells expressing US3 or US3-HA, even after 90 min of chase, the majority of class I MHC molecules remain Endo H sensitive. We conclude that both US3 and US3-HA molecules can retain class I MHC heavy chains in the ER (Fig. 5b, lanes 3 to 6).

View larger version (38K): [in this window] [in a new window]

FIG. 5. US3-HA molecules are functional and can oligomerize with US3 molecules in vivo. (a) Expression of US3-HA in U373-MG cells. Cells infected with pLNW/US3-HA retrovirus vector were pulsed with [35S]methionine for 20 min and subjected to lysis in 1% SDS. US3 and US3-HA molecules were immunoprecipitated by using 4 µl of anti-US3 and/or anti-HA antibodies followed by SDS-PAGE and autoradiography. (b) US3-HA retains the class I MHC heavy chain molecules in the ER. U373 (control), US3+, and US3-HA+ cells were pulsed with [35S]methionine for 20 min and chased for 90 min. Cells were lysed as done for panel a, and 5 µl of anti-HC antibody was added to each lysate to immunoprecipitate the class I MHC heavy-chain molecules. Samples were treated with Endo H and subjected to SDS-PAGE and autoradiography. (c) US3 and US3-HA interact in HEK-293 cells. HEK-293 cells were transiently transfected with pcDNA3.1(+) vectors expressing US3, US3-HA, or both. Cells were pulsed with [35S]methionine for 20 min and lysed in the presence of 1% digitonin. US3 and US3-HA molecules were immunoprecipitated by addition of 3 µl of the relevant antibodies and were subjected to SDS-PAGE and autoradiography.

We then expressed US3 and U3-HA simultaneously in HEK-293 cells by transient transfection. Cells expressing only US3, US3-HA, or both constructs were labeled with [³⁵S]methionine for 20 min. After lysis with digitonin, each sample was split and immunoprecipitated with the anti-HA and anti-US3 antibodies separately. Immunoprecipitation with anti-HA antibody coprecipitates a polypeptide corresponding to US3 in cells that express both US3 and US3-HA (compare lanes 1, 3, and 5 in Fig. 5c). US3 and US3-HA molecules thus associate with each other in HEK-293 cells. This association persists even in the presence of NP-40, a detergent that disrupts the association of US3 with class I MHC molecules (12; data not shown). To ensure that high levels of protein expression that typically occur in transient-transfection experiments do not induce nonspecific interactions between US3 and US3-HA molecules, the experiment described above was then performed with U373-MG cells stably expressing US3 and US3-HA. In these transfectants, anti-HA antibody also recovered US3 molecules from both digitonin and NP-40 lysates, although to a lesser degree than in HEK-293 cells (data not shown).

To visualize the interaction between US3 and US3-HA with a different method, we performed cross-linking experiments with DSP, a cross-linker with a disulfide bond that is cleaved upon reduction. Cells expressing both US3 and US3HA were labeled with [³⁵S]methionine and then incubated with or without DSP for 1 h on ice. The remaining cross-linker was quenched, and the cells were lysed in 1% SDS. After dilution of SDS by addition of NP-40 buffer, the samples were split and immunoprecipitated with the anti-US3 or anti-HA antibodies (Fig. 6a). As expected, the anti-US3 antibody precipitated both US3 and US3-HA molecules in the presence or absence of the cross-linker (Fig. 6a, lanes 1 and 5). However, US3 molecules coimmunoprecipitated with the anti-HA antibody only in the presence of the cross-linker (Fig. 6a, lanes 2 and 6). This result suggests that US3 and US3-HA are not part of a larger aggregate or linked by disulfide bonds, since otherwise they would have remained associated even in the absence of the cross-linker. As a positive control, we used anti-HC antibody and showed that small amounts of US3 can be cross-linked to class I MHC heavy-chain molecules (Fig. 6a, lanes 3 and 7). As a negative control, we used an irrelevant antibody (anti-Sec61ß) to show that US3 and US3-HA do not immunoprecipitate nonspecifically (Fig. 6a, lanes 4 and 8).

View larger version (61K): [in this window] [in a new window]

FIG. 6. US3 oligomerization is independent of the residues involved in ER retention and occurs even when US3 is stably expressed. (a) In vivo cross-linking shows association between US3 and US3-HA. Cells expressing both US3 and US3-HA were pulsed for 1 h with [35S]methionine and incubated with or without DSP for 1 h. Cells were then lysed in 1% SDS, and the lysate was diluted 10-fold with NP-40 buffer. The lysate was divided equally and subjected to immunoprecipitation with 4 µl of the indicated antibodies, followed by SDS-PAGE and autoradiography. (b) US3-HA* also associates with the US3 molecules. Cells were pulsed with [35S]methionine for 45 min and lysed with 1% digitonin. The lysates were normalized for equal amounts of radioactivity, and 4 µl of the anti-HA antibody was used to immunoprecipitate US3-HA molecules. The precipitate was resuspended in 1% SDS and boiled to disrupt all interactions. The samples were diluted as described above and subjected to reimmunoprecipitation with 3 µl of the anti-US3 antibody, followed by SDS-PAGE and autoradiography.

US3 with E63A/K64A mutations is also capable of oligomerization. Mutation of either the Ser58, Glu63, or Lys64 residue prevents retention of US3 in the ER (14). Do US3 molecules bearing these mutations still oligomerize? Using site-directed mutagenesis, we mutated both Glu63 and Lys64 of the US3-HA to alanine (US3-HA*) and made stable cell lines expressing this molecule alone or along with US3. Cells carrying the different versions of US3 were labeled with [³⁵S]methionine and immunoprecipitated them with the anti-HA antibody. The pellet was then resuspended in 1% SDS and boiled to disrupt protein-antibody interaction. After dilution of SDS with NP-40 buffer, a second round of immunoprecipitation was performed with the anti-US3 antibody (Fig. 6b). Both the wild-type US3-HA and US3-HA* molecules interacted with US3 molecules. The Glu63 and Lys64 mutations therefore do not interfere with oligomerization of the US3 molecules (Fig. 6b, lanes 5 and 6). In none of the control cell lines did we observe recovery of material of similar molecular weight (Fig. 6b, lanes 1 to 4). As an additional control, we mixed the [³⁵S]methionine-labeled lysates of cells expressing US3 with lysate from cells expressing US3-HA. The mixture was incubated on ice for 1 h and subjected to the same sequential immunoprecipitation experiment. No association between US3 and US3-HA was observed, indicating that the association observed in our experiments does not occur postlysis (data not shown). Coimmunoprecipitation in HEK-293 cells also showed that US3HA* is capable of interaction with US3 molecules (data not shown).

Top Abstract Introduction Materials and Methods Results Discussion References

 
US3 is one of several HCMV proteins that can down regulate surface expression of class I MHC molecules. US3 achieves its objective by retaining class I MHC molecules in the ER, thus preventing their maturation and subsequent surface expression. To accomplish this task, US3 should itself be retained in the ER. The luminal domain of US3 is responsible for its ER retention, while its transmembrane domain may be involved in binding to class I MHC molecules (15), indicating that class I MHC binding and ER retention involve two separate regions of the molecule. Furthermore, Ser58, Glu63, and Lys64 of US3 likely interact with an ER-resident protein or receptor (protein X) to retain US3 in the ER (14). However, the US3-class I and US3-protein X interactions are transient, since pulse-chase experiments show that the US3/class I complex dissociates rather rapidly (10). Some US3 molecules exit the ER and are degraded without having acquired Endo H resistance. The exact mode of degradation of this pool of US3 molecules is not known. Because class I MHC molecules are efficiently retained in the ER, continuous production of US3 molecules is required to maintain effective US3-class I interactions (10).

Our efforts to produce crystals of US3 or of US3 in a complex with class I MHC molecules have not been successful so far. Therefore, to obtain structural information that might help us understand its mode of action, we used NMR to study the structure of the luminal domain of US3 in solution. An interaction between the labeled US3L-His₆ and unlabeled HLA-A2 should result in a chemical shift in some of the US3L-His₆ residues involved in the interaction. However, all ¹⁵N-labeled US3L-His₆ residues maintain the exact same chemical shift, peak intensity, and two-dimensional NMR profile in the presence or absence of HLA-A2. We conclude that there are no detectable interactions, even at high and equimolar concentrations (0.17 mM), between US3L-His₆ and HLA-A2 molecules. Other methods, such as native gel electrophoresis assay or gel filtration chromatography, likewise failed to detect an interaction between the luminal domains of US3 and HLA-A2 molecules (data not shown).

Despite all attempts, we have been unable to corroborate a direct interaction between the luminal domain of US3 and the extracellular portion of the class I MHC molecule HLA-A2, even though in living cells US3 associates with fully assembled class I molecules. Our inability to demonstrate interactions of the luminal domain of US3 and class I MHC molecules does not allow the conclusion that such interactions do not occur in vivo. However, our results are certainly consistent with an important role for the transmembrane domains of immunoevasins encoded by US cluster of HCMV. We have shown that a single point mutation in the transmembrane segment of US11 completely abolishes its ability to accelerate dislocation and degradation of class I MHC molecules (18).

NMR analysis further suggests that US3L-His₆ exists as small oligomers. Because of this, it is not possible to determine the atomic structure of US3L-His₆ with NMR. However, we were able to assign 70% of the backbone resonances of US3L-His₆ and thus determined its secondary structure based on the chemical shifts observed for each assigned residue. Similar to the US2 luminal domain, US3L-His₆ is composed mainly of ß-sheets. Because of their sequence homology (20% identity), we modeled US3L-His₆ based on the published US2 luminal-domain structure (7, 8). The charged residues in the main ß-sheets of US3 are fewer than those in US2 and are mostly concentrated at the upper half of the structure (Fig. 3a). The relatively neutral surface in the lower half of US3 could be responsible for the observed oligomerization. The class I MHC binding face of US2 is shown in the left panel of Fig. 3b, with the three positively charged residues (blue) located most prominently in the center. For comparison, Fig. 3a, left panel shows the same face on the US3 model as having a very different charge distribution, including a negatively charged residue at the upper-right corner. Such differences in charge distribution readily abolish protein-protein interactions and may well account for our inability to observe US3 and class I MHC interactions. If a binding site for class I MHC molecules exists in the luminal domain of US3, its mode of interaction with its class I MHC client protein must be very different from that seen for the HLA-A2/US2 complex.

Dynamic light-scattering analysis likewise suggests that the behavior of US3L-His₆ in solution is more consistent with that of a small oligomer. Oligomerization does seem to be reversible, since subjecting US3L-His₆ NMR samples to size exclusion chromatography in combination with multiangle light-scattering analysis showed only a single peak with a predicted mass of 14.480 kDa, a finding consistent with monomeric US3L-His₆. Cross-linking experiments showed that US3L-His₆ molecules do not form stable oligomers of a discrete size. The majority of US3L-His₆ molecules remained monomeric even after prolonged exposure to high concentrations of cross-linker, but the ability to be cross-linked at all of course depends on the exact geometry of the reactive residues involved. We conclude that weak but specific interactions may be responsible for the formation of dynamic and heterogeneous populations of oligomers.

In vivo experiments indicate that intact US3 molecules can also oligomerize within the cell. Immunoprecipitation and cross-linking experiments show that US3 molecules interact with each other and that this interaction is sufficiently strong to survive in NP-40 lysis buffers, a condition that leads to dissociation of class I MHC and US3 molecules (12). Control experiments show that association of US3 molecules is not a lysis artifact, and we thus conclude that this interaction is specific. Mutations that inhibit ER retention of US3 (14) do not affect oligomerization of US3. The sites involved in oligomerization must be distinct from those that mediate ER retention and binding of class I MHC molecules. It appears that the interactions between intact US3 molecules within the cell are more stable than those of soluble US3L-His₆. This is not unexpected, since intact US3 contains a transmembrane domain, which confines US3 molecules to the ER membrane and restricts their diffusion to a two-dimensional rather than a three-dimensional environment. In addition, the luminal domain of the full-length US3 terminates at residue 160, whereas US3L-His₆ ends at residue 136. We obviously cannot exclude the possibility that the additional 24 residues may contribute to stability of intact US3 oligomers.

Our results show that the luminal domain of US3 is involved not only in ER retention but also in oligomerization. Since class I MHC molecules are efficiently retained in the ER despite transient interactions between US3/class I and US3 bound to its putative ER receptor, dynamic oligomerization of US3 may be important for efficient retention of class I MHC molecules (Fig. 7). Dynamic oligomerization can mediate formation of patches of US3 oligomers within the ER and US3 molecules in these patches associate and dissociate in dynamic fashion. Binding of a class I MHC molecule to a US3 molecule can bring the class I transmembrane domain in proximity of one of many US3 transmembrane domains present in such patches, thereby reducing the likelihood that class I MHC molecules will dissociate and diffuse away from the cluster of US3 transmembrane domains. In addition, US3 oligomers can recruit more of the putative ER receptor protein X to the site of retention, which in turn can stabilize US3-US3 and US3-class I interactions by providing more anchorage and minimizing lateral movements and diffusion of the protein X/US3/class I complex within the ER membrane. Any US3 molecule that dissociates from the oligomer may associate with another oligomeric patch or be degraded and replaced by a newly synthesized US3 molecule. Hence, reversible oligomerization may be important for US3 to associate with class I molecules and the putative ER receptor protein X in a dynamic fashion.

View larger version (54K): [in this window] [in a new window]

FIG. 7. Schematic model depicting how US3 oligomerization may enhance its function. US3 molecules (purple) bind to the class I MHC trimeric complex (red) via their transmembrane domains (represented by blue and red arrows). The interaction between US3 and class I MHC transmembrane domains is transient. US3 retains itself in the ER by interacting weakly with a putative ER receptor protein X (green) through its Ser58, Glu63, and Lys64 residues (represented by green arrows). The model illustrates dynamic oligomerization of US3 molecules via their interacting luminal domains in the ER (represented by blue arrows). Dynamic oligomerization of US3 molecules may enhance their ability to retain class I MHC molecules. More US3 transmembrane domains in proximity to class I transmembrane domain may make it more difficult for class I MHC molecules to escape the ER. In addition, by recruiting putative ER receptor protein X to the site of retention, both US3-US3 interactions and US3-class I MHC interactions may become more stable due to minimization of lateral diffusion within the ER membrane.

There are numerous wel-documented examples of oligomerization as a means to promote functional efficacy in other biological systems. In Drosophila, for example, differentiation of the R7 photoreceptor cell, involved in development of the compound eye, is induced when the receptor tyrosine kinase Sevenless interacts with its ligand Bride of sevenless (Boss). An attractive model suggests that the Boss ligand could oligomerize through its transmembrane domain, allowing its activation domain to interact with Sevenless as a cluster. In fact, it was shown that oligomerization of the extracellular domain of Boss could enhance its binding to Sevenless receptor (22).

Our present experiments do not address the interesting possibility that, in an HCMV-infected cell, US3 may act in concert with other HCMV-encoded proteins. Although the capacities of US2 and US11 to degrade class I molecules or the ability of US3 to retain class I MHC molecules in the ER do not require the involvement of other HCMV products, we do not know the detailed environment in which these proteins function in an HCMV-infected cell.

We have shown that US3 molecules are capable of oligomerization through their luminal domains in vitro and that they also oligomerize in vivo. Our observation that US3 molecules can self-associate makes the notion of dynamic oligomerization as the means to enhance the affinity of US3 molecule for its targets an attractive model. More detailed structural analysis, as might be accomplished by crystallography, may reveal oligomerization sites and specify the residues that participate in oligomerization. In addition, site-directed mutagenesis may be useful to confirm or identify residues involved in oligomerization. Based on the proposed model, interruption of oligomerization should reduce the efficiency of US3 molecules to retain class I MHC molecules in the ER. A remaining question is the identity of the ER-resident protein or receptor that is utilized by US3 for ER retention. This protein may be an ER chaperone or a novel receptor exploited by HCMV to evade the host immune system.

 
We thank David Jeruzalmi for help with light-scattering experiments and Domenico Tortorella and Amy Hudson for helpful suggestions.

This study was supported by grants from the NIH to G.W. and H.P.

 
* Corresponding author. Mailing address: Department of Pathology, Harvard Medical School, 200 Longwood Ave., Boston, MA 02115. Phone: (617) 432-4777. Fax: (617) 432-4775. E-mail: hidde_ploegh{at}hms.harvard.edu.

S.M. and Z.-Y.J.S. contributed equally to this study.

Top Abstract Introduction Materials and Methods Results Discussion References

 

1
1. Ahn, K., A. Angulo, P. Ghazal, P. A. Peterson, Y. Yang, and K. Fruh. 1996. Human cytomegalovirus inhibits antigen presentation by a sequential multistep process. Proc. Natl. Acad. Sci. USA 93:10990-10995.[Abstract/Free Full Text]
2
2. Bartels, C., T. H. Xia, M. Billeter, P. Guntert, and K. Wuthrich. 1995. The program XEASY for computer-supported NMR spectral analysis of biological macromolecules. J. Biomol. NMR 6:1-10.
3
3. Beersma, M. F., M. J. Bijlmakers, and H. L. Ploegh. 1993. Human cytomegalovirus down-regulates HLA class I expression by reducing the stability of class I H chains. J. Immunol. 151:4455-4464.[Abstract]
4
4. Cornilescu, G., F. Delaglio, and A. Bax. 1999. Protein backbone angle restraints from searching a database for chemical shift and sequence homology. J. Biomol. NMR 13:289-302.[CrossRef][Medline]
5
5. Ferentz, A. E., and G. Wagner. 2000. NMR spectroscopy: a multifaceted approach to macromolecular structure. Q. Rev. Biophys. 33:29-65.[CrossRef][Medline]
6
6. Garboczi, D. N., D. T. Hung, and D. C. Wiley. 1992. HLA-A2-peptide complexes: refolding and crystallization of molecules expressed in Escherichia coli and complexed with single antigenic peptides. Proc. Natl. Acad. Sci. USA 89:3429-3433.[Abstract/Free Full Text]
7
7. Gewurz, B. E., R. Gaudet, D. Tortorella, E. W. Wang, and H. L. Ploegh. 2001. Virus subversion of immunity: a structural perspective. Curr. Opin. Immunol. 13:442-450.[CrossRef][Medline]
8
8. Gewurz, B. E., R. Gaudet, D. Tortorella, E. W. Wang, H. L. Ploegh, and D. C. Wiley. 2001. Antigen presentation subverted: structure of the human cytomegalovirus protein US2 bound to the class I molecule HLA-A2. Proc. Natl. Acad. Sci. USA 98:6794-6799.[Abstract/Free Full Text]
9
9. Gewurz, B. E., E. W. Wang, D. Tortorella, D. J. Schust, and H. L. Ploegh. 2001. Human cytomegalovirus US2 endoplasmic reticulum-lumenal domain dictates association with major histocompatibility complex class I in a locus-specific manner. J. Virol. 75:5197-5204.[Abstract/Free Full Text]
10
10. Gruhler, A., P. A. Peterson, and K. Fruh. 2000. Human cytomegalovirus immediate-early glycoprotein US3 retains MHC class I molecules by transient association. Traffic 1:318-325.[CrossRef][Medline]
11
11. Hyberts, S. G., and G. Wagner. 2003. IBIS: a tool for automated sequential assignment of protein spectra from triple resonance experiments. J. Biomol. NMR 26:335-344.[CrossRef][Medline]
12
12. Jones, T. R., E. J. Wiertz, L. Sun, K. N. Fish, J. A. Nelson, and H. L. Ploegh. 1996. Human cytomegalovirus US3 impairs transport and maturation of major histocompatibility complex class I heavy chains. Proc. Natl. Acad. Sci. USA 93:11327-11333.[Abstract/Free Full Text]
13
13. Koradi, R., M. Billeter, and K. Wuthrich. 1996. MOLMOL: a program for display and analysis of macromolecular structures. J. Mol. Graph. 14:51-55.[CrossRef][Medline]
14
14. Lee, S., B. Park, and K. Ahn. 2003. Determinant for endoplasmic reticulum retention in the luminal domain of the human cytomegalovirus US3 glycoprotein. J. Virol. 77:2147-2156.[Abstract/Free Full Text]
15
15. Lee, S., J. Yoon, B. Park, Y. Jun, M. Jin, H. C. Sung, I. H. Kim, S. Kang, E. J. Choi, B. Y. Ahn, and K. Ahn. 2000. Structural and functional dissection of human cytomegalovirus US3 in binding major histocompatibility complex class I molecules. J. Virol. 74:11262-11269.[Abstract/Free Full Text]
16
16. Lewis, M. J., and H. R. Pelham. 1990. A human homologue of the yeast HDEL receptor. Nature 348:162-163.[CrossRef][Medline]
17
17. Lewis, M. J., and H. R. Pelham. 1992. Ligand-induced redistribution of a human KDEL receptor from the Golgi complex to the endoplasmic reticulum. Cell 68:353-364.[CrossRef][Medline]
18
18. Lilley, B. N., D. Tortorella, and H. L. Ploegh. 2003. Dislocation of a type I membrane protein requires interaction between membrane-spanning segments within the lipid bilayer. Mol. Biol. Cell 14:3690-3698.
19
19. Munro, S., and H. R. Pelham. 1987. A C-terminal signal prevents secretion of luminal ER proteins. Cell 48:899-907.[CrossRef][Medline]
20
20. Rehm, A., P. Stern, H. L. Ploegh, and D. Tortorella. 2001. Signal peptide cleavage of a type I membrane protein, HCMV US11, is dependent on its membrane anchor. EMBO J. 20:1573-1582.[CrossRef][Medline]
21
21. Schekman, R., and L. Orci. 1996. Coat proteins and vesicle budding. Science 271:1526-1533.[Abstract]
22
22. Sevrioukov, E. A., J. H. Walenta, A. Sunio, M. Phistry, and H. Kramer. 1998. Oligomerization of the extracellular domain of Boss enhances its binding to the Sevenless receptor and its antagonistic effect on R7 induction. J. Cell Sci. 111:737-747.[Abstract]
23
23. Springer, S., A. Spang, and R. Schekman. 1999. A primer on vesicle budding. Cell 97:145-148.[CrossRef][Medline]
24
24. Tortorella, D., B. E. Gewurz, M. H. Furman, D. J. Schust, and H. L. Ploegh. 2000. Viral subversion of the immune system. Annu. Rev. Immunol. 18:861-926.[CrossRef][Medline]

---

Journal of Virology, January 2004, p. 413-423, Vol. 78, No. 1
0022-538X/04/$08.00+0     DOI: 10.1128/JVI.78.1.413-423.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.

This article has been cited by other articles:

• Balsara, Z. R., Misaghi, S., Lafave, J. N., Starnbach, M. N. (2006). Chlamydia trachomatis Infection Induces Cleavage of the Mitotic Cyclin B1.. Infect. Immun. 74: 5602-5608 [Abstract] [Full Text]  
• Halenius, A., Momburg, F., Reinhard, H., Bauer, D., Lobigs, M., Hengel, H. (2006). Physical and Functional Interactions of the Cytomegalovirus US6 Glycoprotein with the Transporter Associated with Antigen Processing. J. Biol. Chem. 281: 5383-5390 [Abstract] [Full Text]  
• Zadori, Z., Szelei, J., Tijssen, P. (2005). SAT: a Late NS Protein of Porcine Parvovirus. J. Virol. 79: 13129-13138 [Abstract] [Full Text]  
• Netterwald, J., Yang, S., Wang, W., Ghanny, S., Cody, M., Soteropoulos, P., Tian, B., Dunn, W., Liu, F., Zhu, H. (2005). Two Gamma Interferon-Activated Site-Like Elements in the Human Cytomegalovirus Major Immediate-Early Promoter/Enhancer Are Important for Viral Replication. J. Virol. 79: 5035-5046 [Abstract] [Full Text]  
• Tirosh, B., Iwakoshi, N. N., Lilley, B. N., Lee, A.-H., Glimcher, L. H., Ploegh, H. L. (2005). Human Cytomegalovirus Protein US11 Provokes an Unfolded Protein Response That May Facilitate the Degradation of Class I Major Histocompatibility Complex Products. J. Virol. 79: 2768-2779 [Abstract] [Full Text]  

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Misaghi, S. Articles by Ploegh, H. Search for Related Content PubMed PubMed Citation Articles by Misaghi, S. Articles by Ploegh, H.