Pro-inflammatory cytokines up-regulate MUC1 gene expression in oral epithelial cells.

The membrane-bound mucin MUC1 is expressed ubiquitously on epithelial surfaces and is thought to provide protection from bacterial and chemical injury. The present study was undertaken to determine whether MUC1 was expressed in cultured oral epithelial cells and whether expression is modulated by pro-inflammatory mediators released as part of the host response to infection by oral pathogens. Northern and Western blotting experiments showed that KB cells express MUC1 mRNA and protein. When cells were treated with interleukins (IL-1beta, IL-6), tumor necrosis factor-alpha (TNF-alpha), or interferon-gamma (IFN-gamma), or combinations of these, real-time PCR demonstrated that MUC1 mRNA increased 1.4- to 3.2-fold. Interestingly, a significant increase in levels of MUC1 protein was also observed. While no effect was observed when KB cells were incubated with LPS from Porphyromonas gingivalis, infection of KB monolayers with this oral pathogen caused a 2.85-fold increase in MUC1 transcript levels. These results suggest that increased MUC1 synthesis may be a key element in the host response to infection with oral pathogens.

MG2 and lactoferrin form a heterotypic complex in salivary secretions.

Protein-protein interactions are necessary for homeostasis to be maintained and for biological systems to be integrated. Heterotypic complexes occur in saliva, and a complex between MG2 and SIgA has been suggested to promote microbial clearance from the oral cavity. In this study, we used a peptide display library to investigate previously unrecognized heterotypic complexes involving MG2 and other proteins. The library was panned with MG2 12 times, and analyses of clones identified the sequence Ala-Leu-Leu-Cys-, which occurs in salivary lactoferrin. Blotting experiments confirmed that MG2 and lactoferrin form a heterotypic complex in vitro and in vivo. Periodate treatment of MG2 did not affect the interaction. A synthetic lactoferrin peptide containing the motif Ala-Leu-Leu-Cys-blocked the interaction between MG2 and lactoferrin, confirming the specificity of the interaction identified by panning. This complex may enhance the properties of these salivary components in the oral environment.

Molecular mapping of statherin- and histatin-binding domains in human salivary mucin MG1 (MUC5B) by the yeast two-hybrid system.

MGI is a high-molecular-weight mucin secreted by mucous acinar cells in human submandibular and sublingual glands. We have recently shown that the tracheobronchial mucin MUC5B is a major component of MG1. MUC5B is organized into cysteine-rich N- and C-terminal regions that flank a central tandem-repeat region containing cysteine-rich subdomains and imperfect 29-residue tandem repeats. In earlier work, we have shown that this mucin selectively forms heterotypic complexes with amylase, proline-rich proteins, statherin, and histatins in salivary secretions, and the aim of this study was to identify specific binding domains within MUC5B using the yeast two-hybrid system. Interactions of cysteine-rich domains in the tandem-repeat region (Cys1-Cys4) and C-terminal region (Cys8a, Cys8b, Cys8c) of MUC5B with statherin and histatins were investigated. These studies indicated that histatin 1 selectively bound to Cysl and Cys2, whereas statherin and histatin 1, 3, and 5 selectively bound to Cys8a. Analysis of the primary sequences of the identified binding domains suggests that these domains most probably can fold into globular-like structures in the native mucin. A ProDom blast search revealed that sequences in Cys1, Cys2, and Cys8a exhibit similarity to domains in evolutionarily diverse extracellular proteins known to participate in a wide variety of protein-protein interactions.

The recombinant N-terminal region of human salivary mucin MG2 (MUC7) contains a binding domain for oral Streptococci and exhibits candidacidal activity.

MG2 (the MUC7 gene product) is a low-molecular-mass mucin found in human submandibular/sublingual secretions. This mucin is believed to agglutinate a variety of microbes and thus is considered an important component of the non-immune host defence system in the oral cavity. We have shown that MUC7 can bind to cariogenic strains of Streptococcus mutans and that this binding requires a structural determinant in the N-terminal region. In the present study an expression construct, pNMuc7, encoding the N-terminal 144 amino acids of MUC7 was generated, and the recombinant protein rNMUC7 was expressed in Escherichia coli. Purified rNMUC7 was characterized and the binding of this protein to oral bacteria was investigated in an established assay. The results showed that the recombinant protein bound to S. mutans ATCC 25175 and ATCC 33402, and that alkylation of the two cysteine residues (Cys(45) and Cys(50)) resulted in the complete loss of bacterial binding. This suggests that binding of MUC7 to S. mutans occurs between the N-terminal region of the mucin molecule and the bacterial surface, and that this interaction is dependent on a cysteine-containing domain within this region of MUC7. In addition, the killing activity of rNMUC7 was compared with that of the candidacidal salivary protein histatin 5 in an established Candida albicans (ATCC 44505) blastoconidia killing assay. It was found that the LD(50) values of rNMUC7 and histatin 5 were comparable, and that the recombinant protein displayed significant killing activity at the physiological concentration range of MUC7 in whole saliva. This study is the first to show that the N-terminal region of MUC7 contains a structural determinant for bacterial binding and that this region exhibits candidacidal activity.

Heterogeneity of high-molecular-weight human salivary mucins.

The existence of high-molecular-weight glycoproteins in saliva and salivary secretions has been recognized for nearly 30 years. These proteins, called mucins, are essential for oral health and perform many diverse functions in the oral cavity. Mucins have been intensively studied, and much has been learned about their biochemical properties and their interactions with oral micro-organisms and other salivary proteins. In the past several years, the major high-molecular-weight mucin in salivary secretions has been identified as MUC5B, one of a family of 11 human mucin gene products expressed in tissue-specific patterns in the gastrointestinal, respiratory, and reproductive tracts. MUC5B is one of four gel-forming mucins which exist as multimeric proteins with molecular weights greater than 20-40 million daltons. The heavily glycosylated mucin multimers form viscous layers which protect underlying epithelial surfaces from microbial, mechanical, and chemical assault. Another class of mucin molecules, the membrane-bound mucins, is structurally and functionally distinct from the gel-forming mucins. These proteins do not form multimers and can exist as both secreted and membrane-bound forms, with the latter anchored to epithelial cell membranes through a short membrane-spanning domain. In the present work, we show that two of the membrane-bound mucins, MUC1 and MUC4, are expressed in all major human salivary glands as well as in buccal epithelial cells. While the functions of these mucins in the oral environment are not understood, it is possible that they form a structural framework on the cell surface which not only is cytoprotective, but also may serve as a scaffold upon which MUC5B, and possibly other salivary proteins, assemble.

A recombinant bovine gallbladder mucin polypeptide binds biliary lipids and accelerates cholesterol crystal appearance time.

BACKGROUND & AIMS: Mucin has a central role in the pathogenesis of cholesterol gallstones, in part because of its ability to bind biliary lipids and accelerate cholesterol crystal appearance time. Previous studies have localized these properties to nonglycosylated mucin domains, and we have recently shown that these domains contain a series of 127-amino acid, cysteine-rich repeats. The aim of this study was to express a recombinant mucin polypeptide containing these repeats and investigate its lipid-binding and pronucleating properties. METHODS: A recombinant mucin polypeptide was expressed as a glutathione S-transferase fusion protein in Escherichia coli, purified by affinity chromatography, and compared with native bovine gallbladder mucin in lipid-binding and cholesterol crystal appearance time assays. RESULTS: The recombinant mucin polypeptide bound a hydrophobic fluorescent probe and cholesterol in a concentration-dependent manner. It accelerated the appearance of cholesterol crystals from lithogenic model bile, an effect that was both time and concentration dependent. CONCLUSIONS: The cysteine-rich repeats in the recombinant mucin polypeptide correspond to the protease-sensitive hydrophobic domains identified in earlier biochemical studies. Further delineation of the lipid-binding site(s) in these repeats will provide new insights into the mechanism of cholesterol crystal nucleation and stone growth.

The amino-terminal sequence of MUC5B contains conserved multifunctional D domains: implications for tissue-specific mucin functions.

The MUC5B mucin gene product is expressed in a wide variety of secretory epithelia including the gallbladder, salivary glands, trachea, and colon. Previous studies by us and others have described the C-terminal region as well as the central tandem repeating domain of this mucin. In an effort to understand the functional role of MUC5B in diverse human tissues, the sequence encoding the N-terminal region of this mucin was determined from the sequences of exons in three overlapping genomic clones. Primer extension mapped the site of transcription initiation 25 bp downstream from a putative TATA box element. The N-terminal region of MUC5B contained 1321 amino acids organized into a signal peptide, a short serine-threonine rich region, and three von Willebrand factor-like D domains. Comparison of this sequence with the N-terminal sequences of MUC2 and MUC5AC revealed a much higher degree of identity (46-59%) than that observed in the C-terminal regions of these mucins (33%). The N-terminal sequence of MUC5B also contains a number of sequence motifs common to several groups of extracellular ligand binding and adhesion proteins not previously recognized in mammalian gel-forming mucins. The N-terminal D domains in MUC5B are likely to have important roles in both mucin assembly and in the protective functions of the secreted mucin.

Molecular characterization of a major high molecular weight mucin from human sublingual gland.

Human submandibular/sublingual gland secretions contain a multimeric high molecular weight mucin (MG1) and a monomeric low molecular weight mucin (MG2). MG2 is the product of the MUC7 gene, whereas the gene for MG1 has not been identified. Previously, we isolated a clone (pSM2-1) from a human sublingual gland cDNA expression library using an antibody against deglycosylated MG1 (Troxler et al., Biochem. Biophys. Res Commun., 217, 1112-1119, 1995). In order to identify the mucin gene from which pPM2-1 was derived, Northern blots of human submandibular and sublingual gland RNA were hybridized with a series of probes for tandem repeats found in the high molecular weight secreted mucins MUC2, MUC3, MUC4, MUC5AC, MUC5B, and MUC6. The only known mucin expressed at high levels in sublingual gland was MUC5B, and no known mucin was expressed at high levels in submandibular gland. A series of overlapping clones was obtained by rescreening the sublingual gland cDNA library and by reverse transcriptase-polymerase chain reaction. The resulting clones connected pSM2-1 to a series of MUC5B tandem repeats at the 3' end of the repeat domain and provided the complete nucleotide and deduced amino sequence of the carboxyl terminal region of MG1. This region is enriched with respect to cysteine (approximately 10 mol %) and contained a D domain and a carboxyl terminal domain that could be aligned with the corresponding domains in human intestinal MUC2, human tracheobronchial MUC5AC, and human von Willebrand factor. The limited expression of known mucin genes, together with the considerable mucin synthesizing capacity of submandibular gland, suggests that a novel (previously not described) mucin gene is expressed in this gland and constitutes a portion of MG1 in salivary secretions.

Molecular cloning of a major human gall bladder mucin: complete C-terminal sequence and genomic organization of MUC5B.

Gall bladder mucin has been shown to play a central role in the pathogenesis of cholesterol gallstone disease. While cloning and sequencing studies have provided a wealth of information on the structure of other gastrointestinal and respiratory mucins, nothing is known about the primary structure of human gall bladder mucin. In this study, we show that the tracheobronchial mucin MUC5B is a major mucin gene product expressed in the gall bladder. Antibodies directed against deglycosylated human gall bladder mucin were used to screen a gall bladder cDNA expression library, and most of the isolated clones contained repetitive sequences nearly identical with those in the tandem repeat region of MUC5B. An additional clone (hGBM2-3) contained an open reading frame coding for a 389 residue cysteine-rich sequence. The arrangement of cysteine residues in this sequence was very similar to that in the C-terminal regions of MUC2, MUC5AC and human von Willebrand factor. This cysteine-rich sequence was connected to a series of degenerate MUC5B tandem repeats in a 7.5 kb HincII genomic DNA fragment. This fragment, with ten exons and nine introns, contained MUC5B repeats in exon 1 and a 469 residue cysteine-rich sequence in exons 2-10 that provided a 152 nucleotide overlap with cDNA clone hGBM2-3. Interestingly, the exon-intron junctions in the MUC5B genomic fragment occurred at positions equivalent to those in the D4 domain of human von Willebrand factor, suggesting that these proteins evolved from a common evolutionary ancestor through addition or deletion of exons encoding functional domains.

Molecular cloning of a novel high molecular weight mucin (MG1) from human sublingual gland.

A human sublingual gland cDNA library was screened with a polyclonal antiserum against deglycosylated MG1 and a positive clone, pSM2-1, was isolated which codes for 196 amino acids in the carboxyl-terminal region of this mucin. This region is cysteine-rich and contains a C2-like domain upstream of the extreme carboxyl-terminal domain in which the arrangement of cysteines is nearly identical to that in human von Willebrand factor, human intestinal mucin MUC2, human tracheobronchial mucin MUC5 and porcine and bovine submaxillary gland mucins. Northern analyses with pSM2-1 showed MG1 transcripts are abundant in sublingual gland and barely detectable in submandibular gland. This study provides the first primary sequence data on human salivary mucin MG1 and the significance of the results is discussed with respect to the biosynthesis and differential expression of MG1 in human salivary glands.

Bovine gall-bladder mucin contains two distinct tandem repeating sequences: evidence for scavenger receptor cysteine-rich repeats.

Gall-bladder mucin is a densely glycosylated macromolecule which is the primary secretory product of the gall-bladder epithelium. It has been shown to bind cholesterol and other biliary lipids and to promote cholesterol crystal nucleation in vitro. In order to understand the molecular basis for mucin-lipid interactions, bovine gall-bladder mucin cDNAs were identified by expression cloning and were isolated and sequenced. The nucleotide sequences of these cDNAs revealed two distinct tandem repeating domains. One of these domains contained a 20-amino acid tandem repeating sequence enriched in threonine, serine and proline. This sequence was similar to, but not identical with, the short tandem repeating sequences identified previously in other mammalian mucins. The other domain contained a 127-amino acid tandem repeating sequence enriched in cysteine and glycine. This repeat displayed considerable sequence similarity to a family of receptor- and ligand-binding proteins containing scavenger receptor cysteine-rich repeats. By analogy with other proteins containing these cysteine-rich repeats, it is possible that, in gall-bladder mucin, this domain serves as a binding site for hydrophobic ligands such as bilirubin, cholesterol and other biliary lipids.

Localization of the gene for human heart fatty acid binding protein to chromosome 1p32-1p33.

Heart fatty acid binding protein (hFABP) is an abundant 14-kDa cytosolic protein thought to be involved in trafficking of fatty acids from the plasma membrane to sites of beta-oxidation in mitochondria and peroxisomes and to the endoplasmic reticulum for lipid synthesis. A human hFABP cDNA isolated by polymerase chain reaction was used as a probe for in situ hybridization to metaphase chromosomes. A fragment of the gene for human hFABP was used as a probe for fluorescence in situ hybridization to metaphase chromosomes. The cDNA and genomic probes both localized the gene for human hFABP to chromosome 1p32-1p33.

Amino acid sequence of human plasma galactoglycoprotein: identity with the extracellular region of CD43 (sialophorin).

The amino acid sequence of galactoglycoprotein purified from human plasma was elucidated to 75% completeness by using chemical degradation of peptides and glycopeptides derived from digests of the protein with seven specific proteases. This sequence represents a polypeptide chain of approximately 220 amino acid residues including a high content of serine, threonine, alanine, and proline with one N-linked and multiple O-linked glycans. Comparison of peptide sequences from the native galactoglycoprotein and the deglycosylated derivative demonstrated the locations of 25 sites of O-glycosylation and three serine sites that are not glycosylated. The homogeneous N terminus was established as serine. C-terminal analysis revealed multiple C-terminal residues, suggesting that galactoglycoprotein molecules are of varying lengths. A search of a protein data base revealed that the galactoglycoprotein polypeptide is identical to the N-terminal (extracellular) polypeptide region of the blood-cell surface molecule CD43 (sialophorin, leukosialin). Further support of the relatedness of these molecules was obtained by immunoprecipitation of 125I-labeled galactoglycoprotein by monoclonal anti-CD43 antibodies. The composition and properties of the molecules together with the known structure of the gene encoding CD43 suggest that galactoprotein is derived by proteolytic cleavage from transmembrane "hexasaccharide CD43," known to be expressed on neutrophils, activated T lymphocytes, and platelets.

Characterization of bovine gallbladder mucin. Amino acid sequences of tryptic peptides from the glycosylated domain of the protein core.

Gallbladder mucin is a densely glycosylated macro-molecule that promotes cholesterol gallstone formation in experimental animals and in humans. Bovine gallbladder mucin structure was studied after chemical deglycosylation by treatment with anhydrous hydrogen fluoride at 23 degrees C for 3 hours. Deglycosylated mucin contained less than 5% of the amino sugar and neutral hexose content of native mucin. Electrophoretic and molecular sieve chromatographic analyses indicated that significant cleavage of the mucin polypeptide core had occurred during deglycosylation. Deglycosylated mucin was separated into three major fractions by reverse-phase chromatography, one of which was enriched with respect to threonine and proline. Tryptic peptides prepared from this fraction were purified by molecular sieve and reverse-phase chromatography, and the amino acid sequences (8-20 residues) of the four principal tryptic peptides were determined. These peptides contained 65%-75% threonine and proline residues and demonstrated 80%-100% sequence similarity. These data provide the first information on the primary structure of gallbladder mucin and suggest that repeating amino acid sequences occur in this protein. Comparison of gallbladder mucin peptide structure with the consensus repeat sequence of human intestinal mucin showed approximately 60% sequence similarity. It was concluded that mammalian gastrointestinal mucins may be derived from a common ancestral gene.

Isolation and characterization of peptides from the protein core of bovine gallbladder mucin.

Gallbladder mucin may promote cholesterol gallstone formation by accelerating cholesterol monohydrate crystal nucleation in supersaturated bile. In this study, peptides were isolated from the mucin protein core by protease digestion and molecular-sieve high-performance liquid chromatography. Tryptic peptides were purified by anion exchange or reverse-phase high-performance liquid chromatography, and amino acid compositions were determined. Tryptic peptides were (a) nonglycosylated, (b) selectively enriched in serine, glutamic acid plus glutamine, and glycine, and (c) depleted in threonine and proline compared with native gallbladder mucin. Bilirubin derivatized with Woodward's reagent K covalently bound to purified mucin. Tryptic digestion of the mucin-bilirubin complex yielded low-molecular-weight nonglycosylated peptides with covalently bound bilirubin. These data indicate that the mucin protein core contains at least two distinct domains. One domain is rich in threonine and proline and contains the majority of covalently bound carbohydrate. A second domain, possibly internally located, is nonglycosylated, enriched in serine, glutamic acid plus glutamine, and glycine, and binds hydrophobic ligands such as bilirubin and 1-anilino-8-naphthalene sulfonate. Hydrophobic domains on the mucin protein core may contribute to the pathogenesis of cholesterol cholelithiasis.

Studies on Cyanidium caldarium Phycobiliprotein Pigment Mutants.

Phycobiliprotein biosynthesis was investigated in four strains of the unicellular rhodophyte, Cyandium caldarium, with different pigment phenotypes. All strains were incapable of synthesizing phycobiliproteins when grown in the dark. Western blotting experiments showed that dark-grown cells of the wild-type and mutant GGB synthesized the alpha and beta subunit polypeptides of allophyocyanin and phycocyanin after exposure to light for 24 hours, whereas cells of mutant IIIC and GGBY did not. Similarly, light promoted the appearance of allophycocyanin and phycocyanin mRNAs in the wild-type and GGB but not in IIIC and GGBY. However, Southern blots of restricted genomic DNA from the wild type, IIIC, GGBY, and GGB, all hybridized with heterologous phycobiliprotein gene probes and revealed that all four strains contained identical Pst, EcoRI, and Dral restriction fragments containing allophycocyanin and phycocyanin genes. Cells of the wild type and GGB incubated in the dark with the heme precursor. delta-aminolevulinate, synthesized allophycocyanin and phycocyanin apoproteins providing strong evidence for the role of a tetrapyrrole in regulation of phycobiliprotein gene expression. However, cells of IIIC and GGBY incubated in the dark with delta-aminolevulinate did not contain detectable quantities of allophycocyanin or phycocyanin apoproteins. The possible role of a tetrapyrrole in phycobiliprotein gene expression and basis for the genetic lesion in mutants IIIC and GGBY is discussed.

Cloning of rat aorta lysyl oxidase cDNA: complete codons and predicted amino acid sequence.

Lysyl oxidase cDNA clones were identified by their reactivity with anti-bovine lysyl oxidase in a neonatal rat aorta cDNA lambda gt11 expression library. A 500-bp cDNA sequence encoding four of six peptides derived from proteolytic digests of bovine aorta lysyl oxidase was found from the overlapping cDNA sequences of two positive clones. The library was rescreened with a radiolabeled cDNA probe made from one of these clones, thus identifying an additional 13 positive clones. Sequencing of the largest two of these overlapping clones resulted in 2672 bp of cDNA sequence containing partial 5'- and 3'-untranslated sequences of 286 and 1159 nucleotides, respectively, and a complete open reading frame of 1227 bp encoding a polypeptide of 409 amino acids (46 kDa), consistent with the 48 +/- 3 kDa cell-free translation product of rat smooth muscle cell RNA that was immunoprecipitated by anti-bovine lysyl oxidase. The rat aorta cDNA-derived amino acid sequence contains the sequence of each of the six peptides isolated and sequenced from the 32-kDa bovine aorta enzyme, including the C-terminal peptide with sequence identity of 96%. Northern blots screened with lysyl oxidase cDNA probes identified hybridizing species of 5.8 and 4.5 kb in mRNA of rat aorta and lung, while dot blot analyses were negative for lysyl oxidase mRNA in preparations of rat brain, liver, kidney, and heart. A 258-bp segment of the 3'-untranslated region of lysyl oxidase cDNA is 93% identical with a highly conserved region of the 3'-untranslated sequence of rat elastin cDNA.(ABSTRACT TRUNCATED AT 250 WORDS)

Molecular cloning of human submandibular histatins.

Histatins are a group of histidine-rich polypeptides found in human parotid and submandibular gland secretions. These polypeptides are microbiocidal, possibly involved in maintaining the acquired enamel pellicle, and enhance the glycolytic activity of certain oral micro-organisms. Histatins 1, 3 and 5 are homologous proteins with 38, 32 and 24 amino acid residues, respectively; the cDNAs coding for histatins 1 and 3 have now been isolated and sequenced. The cDNA sequences were highly homologous but contained differences throughout their length, indicating that they arise from different genes that may be derived from a common ancestral gene. Northern blots were hybridized to a series of oligonucleotide probes, designed on the basis of histatin cDNA sequences, and these positively identified mRNAs for histatins 1 and 3. In addition, there was a third mRNA, which hybridized to several histatin oligonucleotide probes, suggesting that histatin 5 might be derived from a distinct mRNA and not by proteolytic processing of histatin 3. A Northern blot of macaque parotid gland total RNA also showed three histatin mRNAs, indicating that similar histatins exist in a non-human primate.

Structural relationship between human salivary histatins.

Histatins are a group of electrophoretically distinct histidine-rich polypeptides with microbicidal activity found in human parotid and submandibular gland secretions. Recently, we have shown that histatins 1, 3, and 5 are homologous proteins that consist of 38, 32, and 24 amino acid residues, respectively, and that these polypeptides kill the pathogenic yeast, Candida albicans. We now describe the isolation and structural characterization of histatins 2, 4, 6, and 7-12, the remaining members of this group of polypeptides. Histatin 2 was found to be identical to the carboxyl terminal 26 residues of histatin 1; histatin 4 was found to be identical to the carboxyl terminal 20 residues of histatin 3; and histatin 6 was found to be identical to histatin 5, but contained an additional carboxyl terminal arginine residue. The amino acid sequences of histatins 7-12 formally correspond to residues 12-24, 13-24, 12-25, 13-25, 5-11, and 5-12, respectively, of histatin 3, but could also arise proteolytically from histatin 5 or 6. These results establish, for the first time, the complete structural relationships between all members of this group of microbicidal proteins in human parotid saliva. The relationship of histatins to one another is discussed in the context of their genetic origin, biosynthesis and secretion into the oral cavity, and potential as reagents in anti-candidal studies.

Heme regulates expression of phycobiliprotein photogenes in the unicellular rhodophyte, Cyanidium caldarium.

Allophycocyanin and phycocyanin in the red alga (Cyanidium caldarium) are chloroplast-encoded, light-harvesting accessory pigments composed of alpha and beta subunit polypeptides (17-19 kDa) to which 1 or more residues of the heme-derived bile pigment chromophore phycocyanobilin are attached by cysteinyl thioether linkages (Offner, G.D., and Troxler, R.F. (1983) J. Biol. Chem. 258, 9931-9940). Western blot experiments utilizing phycobiliprotein antisera revealed that immunoreactive allophycocyanin and phycocyanin apoproteins were absent in cells grown in the dark and present in cells exposed to light. Northern blot experiments using genomic DNA hybridization probes indicated that phycobiliprotein mRNAs were absent in the dark, whereas cells exposed to light contained two allophycocyanin mRNA transcripts, 1.4 and 1.6 kilobases in length, and one phycocyanin mRNA transcript, 3.0 kilobases in length, providing evidence that phycobiliproteins are encoded in photogenes which are only transcriptionally active in the light. Northern and Western analyses demonstrated that cells incubated in the dark with the heme precursor delta-aminolevulinate contained allophycocyanin and phycocyanin mRNAs and apoproteins, indistinguishable in size, number, and quantity from those made in the light. Cells incubated in the dark with delta-aminolevulinate, protoporphyrin IX, or heme, but not biliverdin or phycocyanobilin, synthesized allophycocyanin and phycocyanin alpha and beta apoproteins, suggesting a role for heme in the control phycobiliprotein gene expression. Cells incubated with heme in the dark produced allophycocyanin and phycocyanin mRNA transcripts, but did not produce mRNAs for four other photogenes coding for a P-700 reaction center protein, a 32-kDa herbicide-binding protein, and the large and small subunits of ribulose-bisphosphate carboxylase. These results show, for the first time, that heme is a regulatory factor specifically involved in transcriptional regulation of chloroplast genes for phycobiliproteins.