Diaspirins that cross-link beta chains of hemoglobin: bis(3,5-dibromosalicyl) succinate and bis(3,5-dibromosalicyl) fumarate.

Two double-headed aspirins, bis(3,5-dibromosalicyl) succinate and bis(3,5-dibromosalicyl) fumarate, have been found to be potent acylating agents of intracellular hemoglobin (A or S) in vitro. Furthermore, each of these reagents cross-links beta chains of hemoglobin, probably at the beta cleft. The modified hemoglobins show increased oxygen affinities and reduced gelation or sickling tendencies.     

Equilibrium oxygen binding to human hemoglobin cross-linked between the alpha chains by bis(3,5-dibromosalicyl) fumarate

Oxygen equilibrium curves of human hemoglobin Ao (HbAo) and human hemoglobin cross-linked between the alpha chains (alpha alpha Hb) by bis(3,5-dibromosalicyl) fumarate were measured as a function of pH and chloride or organic phosphate concentration. Compared to HbAo, the oxygen affinity of alpha alpha Hb was lower, cooperativity was maintained, although slightly reduced, and all heterotropic effects were diminished. The major effect of alpha alpha-cross-linking appears to be a reduction of the oxygen affinity of R-state hemoglobin under all conditions. However, while the oxygen affinity of T-state alpha alpha Hb was slightly reduced at physiologic chloride concentration and in the absence of organic phosphates, KT was the same for both hemoglobins in the presence of 2,3-diphosphoglycerate (or high salt) and higher for alpha alpha Hb in the presence of inositol hexaphosphate. The reduced O2 affinity arises from smaller binding constants for both T- and R-state alpha alpha Hb rather than through stabilization of the low affinity conformation. All four Adair constants could be determined for alpha alpha Hb under most conditions, but a3 could not be resolved for HbAo without constraining a4, suggesting that the cross-link stabilizes triply ligated intermediates of hemoglobin.     

Thermal stability of hemoglobin crosslinked in the T-state by bis(3,5-dibromosalicyl) fumarate

Bis(3,5-dibromosalicyl) fumarate was used to crosslink oxyhemoglobin between Lys 82 beta 1 and Lys 82 beta 2 (Walder, J. A., et al. (1979) Biochemistry 18, 4265) and deoxyhemoglobin between Lys 99 alpha 1 and Lys 99 alpha 2 (Chatterjee R.Y., et al. (1986) J. Biol. Chem. 261, 9929). Thermal denaturations demonstrated that alpha crosslinked hemoglobin (alpha 99XLHb A) has the same stability as the beta crosslinked one (beta 82XLHb A). Both alpha and beta crosslinked methemoglobins have a denaturation temperature in 0.9 M guanidine of 57 degrees C compared to 41 degrees C of Hb A. The second product from the T-state crosslinking reaction was found to be crosslinked between the beta chains by chain separation and amino acid analysis. The possible positions for this crosslink are limited to the bisphosphoglycerate binding site in the three-dimensional structure. Its stability is comparable to that of the alpha 99XLHb A or beta 82XLHb A. These modified hemoglobins are potential blood substitutes.     

Structural characterization of human hemoglobin crosslinked by bis(3,5-dibromosalicyl) fumarate using mass spectrometric techniques.

Diaspirin crosslinked hemoglobin (DCLHb) was analyzed by mass spectrometric-based techniques to identify the protein modifications effected by the crosslinking reaction with bis(3,5-dibromosalicyl) fumarate. DCLHb consists of two principal components. These components were isolated by size-exclusion chromatography and identified by measurement of their molecular weight using electrospray mass spectrometry and subsequent peptide mass mapping and mass spectrometric sequence analysis of their individual digests. Three major RP-HPLC fractions were observed from the major hemoglobin in DCLHb. Their MWs matched the MW of heme, intact hemoglobin beta-chain, and two hemoglobin alpha-chains crosslinked by a fumarate moiety, respectively. The minor HPLC peaks of DCLHb were also separated, and characterized by mass spectrometric methods. These minor components revealed additional details of the structural nature of covalent modification of DCLHb.     

The effect of crosslinking by bis(3,5-dibromosalicyl) fumarate on the autoxidation of hemoglobin

Bis(3,5-dibromosalicyl) fumarate was used to crosslink hemoglobin both in the oxy and deoxy states. This double headed diaspirin was known to crosslink oxy Hb A selectively between Lys 82 beta 1 and Lys 82 beta 2 (Walder, J. A., et al. (1979) Biochemistry 18, 4265) and deoxy Hb A between Lys 99 alpha 1 and Lys 99 alpha 2 (Chatterjee R. Y., et al. (1986) J. Biol. Chem. 261, 9929). The autoxidation at 37 degrees C of oxy alpha 99 crosslinked hemoglobin was found to be 1.8 times as fast as that of Hb A while that of the oxy beta 82 crosslinked hemoglobin was only 1.2 times as fast. After 5 hours the formation of methemoglobin in the alpha crosslinked Hb A is 21.3% compared to 10.8% in beta crosslinked Hb A and 6.4% in Hb A. These results may effect the proposed use of alpha 99 crosslinked hemoglobin as a blood substitute by demonstrating the need for protection from autoxidation during storage.     

Effects of crosslinking on the thermal stability of hemoglobin. I. The use of bis(3,5-dibromosalicyl) fumarate

The double-headed aspirin, bis(3,5-dibromosalicyl) fumarate, has been used to crosslink hemoglobin A between Lys 82 beta 1 and Lys 82 beta 2 (J. A. Walder et al. (1979) Biochemistry 18,4265). Denaturation experiments were used to compare the stability of this crosslinked protein to that of hemoglobin A. Thermal denaturations, done in 0.01 M 4-morpholine-propanesulfonic acid, pH 7, containing 0.9 M guanidine to prevent precipitation at high temperatures, were monitored by changes in absorbance between 190 and 650 nm using a diode array spectrophotometer. The sample was heated from 25 to 70 degrees C at 0.3 degrees C/min. The data were analyzed by using both a two-state model and a novel first derivative method. As expected, methemoglobin A had a single, broad transition with a midpoint of 40.7 degrees C. The crosslinked methemoglobin showed a transition at 57.1 degrees C. Two minor transitions, one of which was apparently due to residual unmodified hemoglobin, were also observed in the crosslinked sample. Thus, a single crosslink between only two of the four subunits can lead to a significantly more stable molecule. These results can be explained by Le Chatelier's principle, since crosslinking prevents dissociation of the beta-subunits and, thereby, holds the entire tetramer together.     

Oxygen equilibrium properties of highly purified human adult hemoglobin cross-linked between 82 beta 1 and 82 beta 2 lysyl residues by bis(3,5-dibromosalicyl) fumarate

We investigated oxygen equilibrium properties of highly purified human adult hemoglobin cross-linked between lysine-82 beta 1 and lysine-82 beta 2 by a fumaryl group, which is prepared by reaction of the CO form with bis(3,5-dibromosalicyl) fumarate. The cross-linked hemoglobin preparation isolated by the previous purification method, namely, gel filtration in the presence of 1 M MgCl2 followed by ion-exchange chromatography, was found to be contaminated with about 20% of an electrophoretically silent impurity that shows remarkably high affinity for oxygen. This impurity was separated from the desired cross-linked hemoglobin by a newly developed purification method, which utilizes a difference between the authentic hemoglobin and the impurity in reactivity of the sulfhydryl groups of cysteine-93 beta toward N-ethylmaleimide under a deoxygenated condition. After this purification procedure, the oxygen equilibrium properties of purified cross-linked hemoglobin in the absence of organic phosphate became very similar to those of unmodified hemoglobin with respect to oxygen affinity, cooperativity, and the alkaline Bohr effect. The functional similarity between the cross-linked hemoglobin and unmodified hemoglobin allows us to utilize this cross-linking for preparing asymmetric hybrid hemoglobin tetramers, which are particularly useful as intermediately liganded models. Previous studies on this type of cross-linked hemoglobin should be subject to reexamination due to the considerable amount of the impurity.     

Effects of crosslinking on the thermal stability of hemoglobins. II. The stabilization of met-, cyanomet-, and carbonmonoxyhemoglobins A and S with bis(3,5-dibromosalicyl) fumarate

Hemoglobins A and S were crosslinked between Lys 82 beta 1 and Lys 82 beta 2 using bis (3,5-dibromosalicyl) fumarate (J. A. Walder et al. (1979) Biochemistry 18, 4265). Thermal denaturation experiments were used to compare the stabilities of the met, cyanomet, and carbonmonoxy forms of these crosslinked hemoglobins to the corresponding uncrosslinked proteins. Uncrosslinked carbonmonoxy- and cyanomethemoglobins had transition temperatures about 11 degrees C higher than the corresponding met samples. The increase in denaturation temperature (Tm) due to crosslinking was 15 degrees C for the methemoglobins, 10 degrees C for the cyanomethemoglobins, and 4 degrees C for the carbonmonoxy ones. There was no significant difference in stability between the met and carbonmonoxy crosslinked proteins. In order of increasing stability the samples were: met Hb S less than met Hb A less than CO Hb S less than CO Hb A = CN-met Hb A less than met XL-Hb S = CO XL-Hb S less than met XL-Hb A = CO XL-Hb A less than CN-met XL-Hb A. The slight decrease in the stability of Hb S (beta 6 Glu----Val) compared to Hb A can be explained by the replacement of an external ionic group by a hydrophobic residue in Hb S. In mixtures of crosslinked and normal Hb A, the Tm of the uncrosslinked material was slightly increased by the presence of the more stable crosslinked hemoglobin. The effects of both crosslinking and cyanide or carbon monoxide binding can be explained by Le Chatelier's principle since both would favor the native form of the protein.     

Structural features differentiate the mechanisms between 2S (2 state) and 3S (3 state) folding homodimers

The formation of homodimer complexes for interface stability, catalysis and regulation is intriguing. The mechanisms of homodimer complexations are even more interesting. Some homodimers form without intermediates (two-state (2S)) and others through the formation of stable intermediates (three-state (3S)). Here, we analyze 41 homodimer (25 2S and 16 3S) structures determined by X-ray crystallography to estimate structural differences between them. The analysis suggests that a combination of structural properties such as monomer length, subunit interface area, ratio of interface to interior hydrophobicity can predominately distinguish 2S and 3S homodimers. These findings are useful in the prediction of homodimer folding and binding mechanisms using structural data.     

Structural features of the metal binding site and dynamics of gallium putidaredoxin, a diamagnetic derivative of a Cys4Fe2S2ferredoxin

The first reconstitution of an Fe₂S₂ferredoxin with a diamagnetic prosthetic group was recently described[Kazanis et al. (1995) J. Am. Chem. Soc., 117, 6625–6626]. Thereplacement of the iron–sulfur cluster of the bacterial ferredoxinputidaredoxin (Pdx) by gallium (Ga³⁺) renders the proteindiamagnetic and permits the use of high-resolution NMR methods to identifyresonances near the metal binding site. We now describe structural featuresof the metal binding site that are not observable by standard NMR methods innative Pdx due to paramagnetic line broadening. These results provide thefirst example of high-resolution NMR-derived structural data concerning themetal binding domain of an Fe₂S₂ ferredoxin, andthe first structural information of any sort for the metal binding site of aferredoxin from this class, which includes adrenodoxin, placental ferredoxinand terpredoxin. Assignments were obtained by applying multidimensional NMRmethods to a series of selectively and nonselectively ¹⁵N- and¹³C/¹⁵N-labeled GaPdx samples. For mostexperiments, a mutant of Pdx was used in which a nonligatingCys⁸⁵ is replaced by serine. All of the major structuralfeatures that were identified in native Pdx are conserved in GaPdx. Theoverall protein dynamics is considerably faster in GaPdx than in the nativeprotein, as reflected by amide proton exchange rates. The C-terminalresidue, Trp¹⁰⁶, also exhibits considerable mobility, asindicated by ¹⁵N{¹H} NOE and ¹⁵NT₁ values of the C-terminal residue of the protein.     

Structural features of a zinc binding site in the superantigen strepococcal pyrogenic exotoxin A (SpeA1): implications for MHC class II recognition

Streptococcal pyrogenic exotoxin A (SpeA) is produced by Streptococcus pyogenes, and has been associated with severe infections such as scarlet fever and Streptococcal Toxic Shock Syndrome (STSS). In this study, the crystal structure of SpeA1 (the product of speA allele 1) in the presence of 2.5 mM zinc was determined at 2.8 A resolution. The protein crystallizes in the orthorhombic space group P2(1)2(1)2, with four molecules in the crystallographic asymmetric unit. The final structure has a crystallographic R-factor of 21.4% for 7,031 protein atoms, 143 water molecules, and 4 zinc atoms (one zinc atom per molecule). Four protein ligands-Glu 33, Asp 77, His 106, and His 110-form a zinc binding site that is similar to the one observed in a related superantigen, staphylococcoal enterotoxin C2. Mutant toxin forms substituting Ala for each of the zinc binding residues were generated. The affinity of these mutants for zinc ion confirms the composition of this metal binding site. The implications of zinc binding to SpeA1 for MHC class II recognition are explored using a molecular modeling approach. The results indicate that, despite their common overall architecture, superantigens appear to have multiple ways of complex formation with MHC class II molecules.     

The roles of Tyr(CD1) and Trp(G8) in Mycobacterium tuberculosis truncated hemoglobin O in ligand binding and on the heme distal site architecture

The crystal structure of the cyano-met form of Mt-trHbO revealed two unusual distal residues Y(CD1) and W(G8) forming a hydrogen-bond network with the heme-bound ligand [Milani, M., et al. (2003) Proc. Natl. Acad. Sci. U.S.A. 100, 5766-5771]. W(G8) is an invariant residue in group II and group III trHbs and has no counterpart in other globins. A previous study reported that changing Y(CD1) for a Phe causes a significant increase in the O2 combination rate, but almost no change in the O2 dissociation rate [Ouellet, H., et al. (2003) Biochemistry 42, 5764-5774]. Here we investigated the role of the W(G8) in ligand binding by using resonance Raman spectroscopy, stopped-flow spectrophotometry, and X-ray crystallography. For this purpose, W(G8) was changed, by site-directed mutagenesis, to a Phe in both the wild-type protein and the mutant Y(CD1)F to create the single mutant W(G8)F and the double mutant Y(CD1)F/W(G8)F, respectively. Resonance Raman results suggest that W(G8) interacts with the heme-bound O2 and CO, as evidenced by the increase of the Fe-O2 stretching mode from 559 to 564 cm-1 and by the lower frequency of the Fe-CO stretching modes (514 and 497 cm-1) compared to that of the wild-type protein. Mutation of W(G8) to Phe indicates that this residue controls ligand binding, as evidenced by a dramatic increase of the combination rates of both O2 and CO. Also, the rate of O2 dissociation showed a 90-1000-fold increase in the W(G8)F and Y(CD1)F/W(G8)F mutants, that is in sharp contrast with the values obtained for the other distal mutants Y(B10)F and Y(CD1)F [Ouellet, H., et al. (2003) Biochemistry 42, 5764-5774]. Taken together, these data indicate a pivotal role for the W(G8) residue in O2 binding and stabilization.     

Conserved surface features form the double-stranded RNA binding site of non-structural protein 1 (NS1) from influenza A and B viruses

Influenza A viruses cause a highly contagious respiratory disease in humans and are responsible for periodic widespread epidemics with high mortality rates. The influenza A virus NS1 protein (NS1A) plays a key role in countering host antiviral defense and in virulence. The 73-residue N-terminal domain of NS1A (NS1A-(1-73)) forms a symmetric homodimer with a unique six-helical chain fold. It binds canonical A-form double-stranded RNA (dsRNA). Mutational inactivation of this dsRNA binding activity of NS1A highly attenuates virus replication. Here, we have characterized the unique structural features of the dsRNA binding surface of NS1A-(1-73) using NMR methods and describe the 2.1-A x-ray crystal structure of the corresponding dsRNA binding domain from human influenza B virus NS1B-(15-93). These results identify conserved dsRNA binding surfaces on both NS1A-(1-73) and NS1B-(15-93) that are very different from those indicated in earlier "working models" of the complex between dsRNA and NS1A-(1-73). The combined NMR and crystallographic data reveal highly conserved surface tracks of basic and hydrophilic residues that interact with dsRNA. These tracks are structurally complementary to the polyphosphate backbone conformation of A-form dsRNA and run at an approximately 45 degrees angle relative to the axes of helices alpha2/alpha2'. At the center of this dsRNA binding epitope, and common to NS1 proteins from influenza A and B viruses, is a deep pocket that includes both hydrophilic and hydrophobic amino acids. This pocket provides a target on the surface of the NS1 protein that is potentially suitable for the development of antiviral drugs targeting both influenza A and B viruses.     

Structural similarities of micelle-bound peptide YY (PYY) and neuropeptide Y (NPY) are related to their affinity profiles at the Y receptors

Here, we investigate the structure of porcine peptide YY (pPYY) both when unligated in solution at pH 4.2 and when bound to dodecylphosphocholine (DPC) micelles at pH 5.5. pPYY in solution displays the PP-fold, with the N-terminal segment being back-folded onto the C-terminal alpha-helix, which extends from residue 17 to 31. In contrast to the solution structure of Keire et al. published in the year 2000 the C-terminal helix does not display a kink around residue 23-25. The root mean square deviation (RMSD) for backbone atoms of the NMR ensemble of conformers to the mean structure is 0.99(+/-0.35) Angstrom for residues 14-31. The back-fold is supported by values of 0.60+/-0.1 for the (15)N(1)H-NOE and by generalized order parameters S(2) of 0.74+/-0.1 for residues 5-31 which indicate that the peptide is folded in that segment. We have additionally used DPC micelles as a membrane model and determined the structure of pPYY when bound to it. Therein, an alpha-helix occurs in the segment comprising residues 17-31 and the N terminus freely diffuses in solution. The hydrophobic side of the amphipathic helix forms the micelle-binding interface and hydrophobic side-chains extend into the micelle interior. A significant stabilization of helical conformation occurs in the C-terminal pentapeptide, which is important for receptor binding. The latter is supported by positive values of the heteronuclear NOE in that segment (0.52+/-0.1 compared to 0.08+/-0.4 for the unligated form) and by values of S(2) of 0.6+/-0.2 (versus 0.38+/-0.2 for the unligated form). The structures of micelle-bound pPYY and pNPY are much more similar than those of pPYY and bPP with pairwise RMSDs of 1.23(+/-0.21)A or 3.21(+/-0.39) Angstrom, respectively. In contrast to the conformational similarities in the DPC-bound state their structures in solution are very different. In fact pPYY is more similar to bPP, which with its strong preference for the Y(4) receptor displays a completely different binding profile. Considering the high degree of sequence homology of pNPY and pPYY (>80%) and the fact, that their binding affinities at all receptor subtypes are high and, more importantly, rather similar, it is much more likely that PYY and NPY are recognized by the Y receptors from the membrane-bound state. As a consequence of the latter the PP-fold is not important for recognition of PYY or NPY at the Y receptors. To our knowledge this work provides for the first time strong arguments derived from structural data that support a membrane-bound receptor recognition pathway.     

Characterization of yeast U1 snRNP A protein: identification of the N-terminal RNA binding domain (RBD) binding site and evidence that the C-terminal RBD functions in splicing.

The yeast U1A protein is a U1 snRNP-specific protein. Like its human counterpart (hU1A), it has two conserved RNA binding domains (RBDs). The N-terminal RBD is quite different from the human protein, and a binding site on yeast U1 snRNA is not readily apparent. The C-terminal RBD is of unknown function. Using in vivo dimethyl sulfate (DMS) protection of mutant strains, we defined a region in yeast U1 snRNA as the likely U1A N-terminal RBD binding site. This was confirmed by direct in vitro binding assays. The site is very different from its vertebrate counterpart, but its location within yeast U1 snRNA suggests a conserved structural relationship to other U1 snRNP components. Genetic studies and sensitive in vivo splicing measurements indicate that the yeast U1A C-terminal RBD also functions in pre-mRNA splicing. We propose that the N-terminal RBD serves to tether the splicing-relevant C-terminal RBD to the snRNP.     

Structural and enzyme activity studies demonstrate that aryl substituted 2,3-butadienamine analogs inactivate Arthrobacter globiformis amine oxidase (AGAO) by chemical derivatization of the 2,4,5-trihydroxyphenylalanine quinone (TPQ) cofactor

Copper amine oxidases (CAOs) are a family of redox active enzymes containing a 2,4,5-trihydroxyphenylalanine quinone (TPQ) cofactor generated from post translational modification of an active site tyrosine residue. The Arthrobacter globiformis amine oxidase (AGAO) has been widely used as a model to guide the design and development of selective inhibitors of CAOs. In this study, two aryl 2,3-butadienamine analogs, racemic 5-phenoxy-2,3-pentadienylamine (POPDA) and racemic 6-phenyl-2,3-hexadienylamine (PHDA), were synthesized and evaluated as mechanism-based inactivators of AGAO. Crystal structures show that both compounds form a covalent adduct with the amino group of the substrate-reduced TPQ, and that the chemical structures of the rac-PHDA and rac-POPDA modified TPQ differ by the allenic carbon that is attached to the cofactor. A chemical mechanism accounting for the formation of the respective TPQ derivative is proposed. Under steady-state conditions, no recovery of enzyme activity is detected when AGAO pre-treated with rac-PHDA or rac-POPDA is diluted with excess amount of the benzylamine substrate (100-fold K(m)). Comparing the IC(50) values further reveals that the phenoxy substituent in POPDA offers an approximately 4-fold increase in inhibition potency, which can be attributed to a favourable binding interaction between the oxygen atom in the phenoxy group and the active site of AGAO as revealed by crystallographic studies. This hypothesis is corroborated by the observed >3-fold higher partition ratio of PHDA compared to POPDA. Taken together, the results presented in this study reveal the mechanism by which aryl 2,3-butadienamines act as mechanism-based inhibitors of AGAO, and the potency of enzyme inactivation could be fine-tuned by optimizing binding interaction between the aryl substituent and the enzyme active site.     

Mitochondrial nicotinamide nucleotide transhydrogenase: nonidentical modification by N,N'-dicyclohexylcarbodiimide and N-(ethoxycarbonyl)-2-ethoxy-1,2-dihydroquinoline at the NAD(H) binding site

The energy-linked nicotinamide nucleotide transhydrogenase (TH) purified from bovine heart mitochondria is inhibited by the carboxyl group modifiers, N,N'-dicyclohexylcarbodiimide (DCCD) and N-(ethoxycarbonyl)-2-ethoxy-1,2-dihydroquinoline (EEDQ). With either reagent, complete activity inhibition corresponds to modification of one carboxyl group per 2 mol (monomers) of this dimeric enzyme, suggesting half-site reactivity toward DCCD and EEDQ [D. C. Phelps, and Y. Hatefi (1984) Biochemistry 23, 4475-4480; 6340-6344]. It has also been shown in the former reference that DCCD appears to modify TH at the NAD(H)-binding site. The present paper presents data suggesting that EEDQ also binds at or near the NAD(H)-binding domain of TH, but at a site not identical to that of DCCD: TH modified with and inhibited approximately 85% by EEDQ could be further labeled with [14C]DCCD to the extent of 70% of the maximum in the same time period that unmodified TH was modified by [14C]DCCD to near saturation (1 mol DCCD/TH dimer); DCCD-modified TH did not bind to NAD-agarose, while EEDQ-modified TH showed partial affinity for NAD-agarose; 5'-AMP completely protected TH against modification by DCCD, but showed only a weak protective effect against EEDQ; by contrast, NMNH, which is a TH substrate and binds to the NADH site, did not protect TH against DCCD, but completely protected the enzyme against attack by EEDQ. The results are consistent with the possibility that DCCD modifies TH where the 5'-AMP moiety of NAD(H) binds, while EEDQ modifies the enzyme where the NMN(H) moiety of NAD(H) resides.     

Phosphorylation of human cytomegalovirus glycoprotein B (gB) at the acidic cluster casein kinase 2 site (Ser900) is required for localization of gB to the trans-Golgi network and efficient virus replication

Human cytomegalovirus (HCMV) glycoprotein B (gB), encoded by the UL55 open reading frame, is an essential envelope glycoprotein involved in cell attachment and entry. Previously, we identified residue serine 900 (Ser900) as a unique site of reversible casein kinase 2 phosphorylation in the cytoplasmic domain of HCMV gB. We have also recently shown that gB is localized to the trans-Golgi network (TGN) in HCMV-permissive cells, thereby identifying the TGN as a possible site of virus envelopment. The aim of the current study was to determine the role of Ser900 phosphorylation in transport of gB to the TGN and in HCMV biogenesis. Recombinant HCMV strains were constructed that expressed gB molecules containing either an aspartic acid (gBAsp900) or alanine residue (gBAla900) substitution at Ser900 to mimic the phosphorylated or nonphosphorylated form, respectively. Immunofluorescence analysis of the trafficking of gB mutant molecules in fibroblasts infected with the HCMV recombinants revealed that gBAsp900 was localized to the TGN. In contrast, gBAla900 was partially mislocalized from the TGN, indicating that phosphorylation of gB at Ser900 was necessary for TGN localization. The increased TGN localization of gBAsp900 was due to a decreased transport of the molecule to post-TGN compartments. Remarkably, the substitution of an aspartic acid residue for Ser900 also resulted in an increase in levels of progeny virus production during HCMV infection of fibroblasts. Together, these results demonstrate that phosphorylation of gB at Ser900 is necessary for gB localization to the TGN, as well as for efficient viral replication, and further support the TGN as a site of HCMV envelopment.