Posttranslational Modifications in Lens Fiber Connexins Identified by Off-Line-HPLC MALDI-Quadrupole Time-of-Flight Mass Spectrometry

Abstract

purpose. Gap junction intercellular communication is necessary for the development and maintenance of the lens. Lens fiber connexins are known to be posttranslationally modified, but little detail is available regarding the nature of some of these modifications and the specific amino acids affected. The purpose of this study was to identify posttranslational modification in the bovine lens fiber connexins, Cx44 and Cx49.

methods. Crude preparations of bovine lens membranes were isolated by centrifugation. The membrane preparations were digested with trypsin or chymotrypsin, and the entire mixture of peptides produced was separated by reverse-phase high performance liquid chromatography and then analyzed by mass spectrometry and tandem mass spectrometry.

results. Coverage of significant portions of the cytoplasmic domains of both Cx44 and Cx49 were successfully obtained. Several of the Ser and Thr residues in the C-tail of Cx44 were phosphorylated, whereas in Cx49 only Ser phosphorylation was detected; however, in this connexin, the phosphorylated residues were located in both the C-tail and the central cytoplasmic loop. The data also show that the N-terminal Met residue in each connexin is removed and that the newly exposed N termini become acetylated. In addition, cleavage sites were identified in both proteins.

conclusions. The study documented the nature and locations of several previously unknown posttranslational modifications in lens fiber connexins. This detailed knowledge of the specific posttranslationally modified sites will allow further work to elucidate the mechanisms that different signaling pathways use to regulate connexins in lens fiber cells.

Gap junctions are structures in the plasma membranes of two adjacent cells that contain aqueous channels connecting the cytoplasm of the cells. ^{1 2} The channels allow molecules with a molecular mass of <1.5 kDa to be directly exchanged between cells. Because they allow the transfer of ions and second messengers, gap junctions play important regulatory roles in development, differentiation, apoptosis, electrical and metabolic coupling, and tissue homeostasis. ^{3 4}  

In vertebrates, gap junctions are composed of integral membrane proteins called connexins. There are 20 or more known members of the connexin gene family in humans and mice. ⁵ Connexin proteins have four transmembrane regions, linked by two extracellular loops and a cytoplasmic loop. ^{6 7} The N and C termini are located in the cytoplasm as well. Six connexin proteins oligomerize to form a connexon or hemichannel that is transported to the plasma membrane. One hemichannel docks with a hemichannel in an adjacent cell to form a complete channel. These channels are gated in response to various factors including pH, voltage, and phosphorylation. ²  

The mouse and the human lens expresses three connexins: Cx43, Cx46 (the bovine orthologue is known as Cx44 ^{8 9} ), and Cx50 (the bovine orthologue is known as Cx49 ¹⁰ ). Cx43 expression is restricted to the anterior layer of lens epithelial cells, ¹¹ where Cx50 is also expressed. ^{12 13} Both Cx46 and Cx50 are expressed in the bulk of the ocular lens, ^{14 15 16 17} which is a mass of concentrically arranged, elongated fiber cells. ¹⁸ The cells throughout the lens are metabolically active, and it has been suggested that an internal circulatory system exists to supply nutrients and remove wastes from throughout the avascular lens. ¹⁹ This proposed circulatory system depends, in part, on gap junctions. This hypothesis is supported by the observations that, in humans, mutations in either Cx46 or Cx50 are associated with the formation of cataracts, ^{20 21} and, in mice, targeted ablation of Cx46 ²² and Cx50 ²³ produces cataracts as well. 

Both Cx44/Cx46 and Cx49/Cx50 are posttranslationally modified, including a truncation at the C terminus as the lens fibers age. ^{24 25} Two cleavage sites, Glu290 and Ser300, have been reported in ovine Cx49. ²⁴ There is evidence that Cx44 is cleaved as well, ^{9 26} but the site has not been identified. Both proteins are also phosphorylated in vivo. ^{27 28 29 30 31 32} While one phosphorylation site in the chicken orthologue of Cx49/Cx50 has been identified, ³¹ for the most part the phosphorylation sites in the lens fiber connexins are unknown. 

Cleavage of the C-tail is expected to have profound functional consequences, as this region serves as the main site of interaction with other proteins ^{33 34} and also harbors multiple potential phosphorylation sites. Connexin phosphorylation has been shown to affect gap junction intercellular communication via multiple mechanisms. ^{35 36 37} Indeed, an association has been observed between the cleavage of the C-tail of Cx46 and changes in gap junction plaque distribution and dye coupling in the rat lens. ²⁵ It has also been demonstrated that phosphorylation of Cx50 by PKCγ leads to a reduction in intercellular coupling and disassembly of Cx50 channels. ³⁸ A more thorough understanding of how the lens fiber connexins contribute to lens function will require a detailed picture of the posttranslational modifications these connexins undergo. To this end, we used a proteomics approach. By proteolytically digesting crude membrane preparations from cow lens fibers and separating the resultant peptides by reversed-phase HPLC, we identified several peptides derived from Cx44 and Cx49 by matrix-assisted laser desorption/ionization (MALDI) time-of-flight mass spectrometry. We also identified several phosphorylated residues in both lens fiber connexins. Other modifications detected include acetylation at the N terminus, a deamidated asparagine residue in Cx49, and potential cleavage sites in both connexins. 

Materials and Methods

Tissue

Bovine eyes were obtained from a local abattoir. The lenses were removed from the eyes, the lens capsule was removed, and the remaining lens material was frozen at −80°C. 

Isolation of Fiber Membranes

Gap junction–rich lens fiber membranes were isolated as described in Kistler et al. ³⁹ Approximately 500 mg wet weight of thawed lenses was initially homogenized in Tris buffer (5.0 mM Tris [pH 8.0], 5.0 mM EDTA, and 5.0 mM EGTA). Crude membrane preparations were pelleted from the lysates by centrifugation at 83,000g for 20 minutes (25,900 rpm, SW41Ti rotor; Beckman, Fullerton, CA). Proteins adhering to the membranes were extracted with 4 M urea (5.0 mM Tris [pH 9.5] and 5.0 mM each EDTA and EGTA), and the membranes were pelleted by centrifugation and then re-extracted with 20 mM NaOH. The membranes were subsequently pelleted by centrifugation at 83,000g, washed twice with Tris buffer (5.0 mM Tris, [pH 7.0] and 2.0 mM each EDTA and EGTA, and 100 mM NaCl) and finally pelleted by centrifugation 83,000g. 

In-Solution Digest

The membrane pellet was resuspended in 100 μL of 50 mM NH₄HCO₃ in water/methanol (40/60) or in water/acetonitrile (ACN; 80:20). Either 5.0 μL of 200 ng/μL sequencing grade trypsin (Promega Madison, WI) or 5.0 μL of 200 ng/μL type I-s α-chymotrypsin (Sigma-Aldrich, St. Louis) was added, and the sample was placed in a sonicating bath for 5 minutes followed by overnight digestion at 37°C. Water was added to the samples, which were then centrifuged at 14,500 rpm in a centrifuge (EBA 12R; Hettich). The supernatant was removed and dried in a vacuum centrifuge. For each sample, two sequential digestions were performed. After the first digestion, the membranes were pelleted, the supernatant was collected, and the resuspended pellet was redigested. The membranes were pelleted again and the supernatant collected. The first and second supernatants were analyzed separately, with the second digest typically producing a greater number of unique ions. 

HPLC Separation

Separations of the peptides derived from the in-solution digest were performed (microAgilent 1100; Agilent Technologies, Wilmington, DE). Deionized water, ACN, and trifluoroacetic acid (TFA) were used for the preparation of the eluents. Dried peptides were dissolved in 20 μL of water, and 5 μL was injected into a 5.0-μL loop, and onto a 150-μm × 150-mm column (5.0 μm; 218 TP C18; Grace Vydac, Hesperia, CA), which was maintained at 30°C. The sample was eluted from the column with a linear gradient of 1% to 46% ACN (0.1% TFA) in 35 minutes (1.32% ACN per minute at 4 μL/min flow rate), followed by an increase in ACN concentration from 47% to 99% over 5 minutes. PEEK 65 μm inner diameter (ID) and fused silica 50 μm ID tubing were used for pre- and postcolumn liquid connections. The column effluent (4 μL/min) was mixed online with MALDI matrix solution (9 mg 2,5-dihydroxybenzoic acid [DHB], Aldrich, Milwaukee, WI) dissolved in 500 μL ACN and 500 μL methanol) at a flow rate of 2.0 μL/min. Forty fractions were deposited onto a gold target at 1-minute intervals (4 μL eluent + 2 μL matrix solution per fraction) by a computer-controlled robot built in-house. Online mixing was performed (Microtee P775; Upchurch Scientific, Oak Harbor, WA), and the fractions were air-dried and subjected to MALDI-MS analysis. 

MALDI MS and MS/MS

Mass spectra were obtained using a matrix-assisted laser desorption/ionization (MALDI)-quadrupole time-of-flight (QqTOF) mass spectrometer built at the University of Manitoba. ⁴⁰ A mass spectrum of each of the 40 fractions was acquired and combined in a set for each HPLC separation/target plate. The MS and chromatographic data were analyzed with software developed in-house (described later), to identify ion peaks of interest. Lists of precursor masses were then generated and tandem mass spectra (MS/MS) were manually acquired according to the list using the same MALDI-QqTOF instrument. 

Data Processing

The mass chromatograms were initially processed by software developed at the University of Manitoba, which identifies tryptic peptides based on Mass and Retention Time (MART). ⁴¹ The software has a database of tryptic peptides derived from the Swiss-Prot database (http://www.expasy.org; provided in the public domain by Swiss Institute of Bioinformatics, Geneva, Switzerland), where each peptide has a mass and a hydrophobicity rating. The hydrophobicity is calculated based on the sequence using algorithms developed from the sequence-specific retention calculator. ⁴² The software is then able to identify proteins present in the sample and assign peptides to peaks in the chromatogram. We used a tolerance of ±3 minutes for retention time and 10 ppm for the mass. This method is similar to a mass fingerprint, but the added dimension of retention time allows proteins from a complex mixture to be identified. MS/MS was performed to confirm the sequences of peptides identified by MART, as well as to sequence peptides that MART did not identify. The tandem mass spectra were processed by peak-selection software developed in-house, which produced peak lists in the Mascot Generic Format. The peak lists were then submitted to the GPM Website (http://thegpm.org/ The Global Proteome Machine Organization Proteomics Database, provided in the public domain by The Global Proteome Machine Organization, a consortium) for identification. GPM searches were performed using the ENSEMBL Bos taurus database (ensembl http://www.ensembl.org provided in the public domain by a consortium of the EMBL-EBI and the Sanger Institute, UK). The measurement error allowed was 0.4 Da, and the cutoff for the expectation value was log e < −1. For the complete modifications field we chose either none or carbamidomethyl of cysteines for the samples that were alkylated. Potential modifications selected included oxidation, deamidation, and phosphorylation. GPM automatically includes pyroglutamine formation as well as N-terminal acetylation, and it also scans for peptides with nontryptic cleavages in its refinement step. The device and parent ion method was set to TOF (100 ppm). The spectra were also converted into the wiff format to be manually sequenced with Analyst software to pinpoint modified residues (MDS Sciex, Concord ON, Canada). 

Protein Sequences

The protein sequences used for Cx44 and Cx49 were retrieved from the NCBI (National Center for Biotechnology Information, Bethesda, MD) Website. The accession numbers are AAA50954 for Cx44 and XP_602360 for Cx49. Kinases potentially responsible for phosphorylating residues on Cx44 and Cx49 were identified by the NetPhosK 1.0 web server (www.cbs.dtu.dk/services/NetPhosK/ provided in the public domain by the Center for Biological Sequence Analysis, Technical University of Denmark, Lyngby). 

Results

Identification of Peptides from Cx49 and Cx44

A typical mass chromatogram of the tryptic digest of a crude membrane preparation contained 600 to 900 monoisotopic peaks (Fig. 1) . Using mass and retention time our protein identification software was able to assign several hundred of the peaks to peptides from known lens proteins such as crystallins. The initial processing could assign 10 of the peaks to peptides derived from Cx44 and 18 of the peaks to peptides from Cx49. 

Tandem MS

The peptide assignments based on mass and retention time were confirmed by sequencing the peptides with MS/MS (Figs. 2 3) . In MS/MS a precursor ion with a specific charge to mass ratio is allowed to enter a chamber where it interacts with argon gas. Collision-induced fragmentation of the peptides results in cleavage of amide bonds and the formation of N termini containing b-ions and C termini containing y-ions. Labeling of the b- and y-ions was performed according to nomenclature established by Biemann. ⁴³ The MS/MS mass spectra were sequenced by hand, and submitted to the GPM Website to obtain an objective confidence measure of the identity of a particular peptide (Tables 1 2) . The log GPM e-value reported for each peptide in Tables 1 and 2is a measure of the probability that a spectrum matches the theoretical value by chance (e.g., a log e-value of −2 means that in 1 of 100 experiments, one would expect to see a particular score on the basis of chance). ⁴⁴ Many peptides scored with high e-values and some peptides that were manually identified from Cx44 did not receive log e-values below the threshold of −1. In general, most automated peptide identifying algorithms have some difficulty identifying spectra when the precursor ion is in the +1 charge state characteristic of MALDI. These algorithms do much better with multiply charged precursor ions commonly produced in electrospray. ⁴⁵ The reasons GPM failed to identify some of the manually identified Cx44 peptides are not entirely clear and may be particular to each peptide (Table 1) . One example is the case of the high log e-value for one of the peptides with multiple modifications (peptide 228-242). This peptide was present with many different modifications, so only a small amount of the unmodified peptide was present. In its unmodified form, the peptide consequently produced a low signal resulting in a relatively low-quality MS/MS spectrum. On the other hand, the reason for the high log e-value for the 110-115 peptide is probably its small size. The explanation for the high log e-value obtained for the 110-132 peptide is less certain, but it may be because the GPM does not take into account internal cleavages, and this peptide contains several proline and aspartic acid residues that make peptides prone to such events. It is also interesting in this context to note that the two other peptides with the same C terminus (106-132 and 108-132) also had relatively high log e-values (−2.7 and −1.9, respectively), although the scores reached significance in both cases. Conversely, the three peptides that end at residue 129 (i.e., peptides lacking DDR from their C termini relative to the peptides with C termini at 132) all scored with much lower log e-values. These examples point to some of the challenges of determining the probability that a spectrum is matched to the correct peptide, and indicate that the expectation values GPM reports in some cases underestimate the true probability that a spectrum is correctly matched to a peptide. 

Tandem MS was also performed on peaks that could not be assigned to proteins. Many of these turned out to be due to missed cleavages in crystallin proteins. However, several truncated peptides from Cx44 and Cx49 were also identified. Overall, we were able to obtain coverage of 30% of the complete Cx44 sequence and 46% coverage of the cytoplasmic domains of the protein. For Cx49 we were able to obtain coverage of 47% of the complete protein, and 71% coverage of the cytoplasmic domains. In total, we identified 14 Cx44 peptides, 5 of which included posttranslational modifications (Table 1) , and 32 Cx49 peptides, 12 of which were posttranslationally modified (Table 2) . 

Identification of Posttranslational Modifications

Several common peptide modifications were observed, including the oxidation of Met residues (+15.99 Da) and a shift of Glu to pyroglutamine (−17.03) in peptides where the initial N-terminal residue was glutamine. Asn121 in Cx49 showed a gain of 0.98 Da characteristic of deamidation. 

Both Cx44 and Cx49 were modified at their N termini. The initial Met residues were cleaved from both connexins, and a fraction of both connexins was acetylated at the N terminus as well. The acetylation increased the retention time of the peptides and shifted the mass by 42.01 Da. In the MS/MS spectra of the acetylated peptides, the C termini containing y-ions remained identical with the nonacetylated versions but all of the b-ions, which contain the N termini, were shifted up by 42.01 Da (Fig. 2) . 

In the initial analysis of the MS spectra, we were not able to identify any peptides arising from nontryptic cleavages, such as would be present in a protein that had been proteolytically cleaved in vivo. A list of peaks that could not be assigned to any proteins based on mass and retention time was generated, and tandem MS spectra taken. A portion of these could be assigned to Cx44 and Cx49. Two nontryptic peptides from Cx44 were identified, one from the cytoplasmic loop region ending at Asp124 and another from the C-terminal region ending at Leu255. Three nontryptic peptides from Cx49 were also identified. These terminated at His284, Phe286m, and Glu290. 

The mass of a phosphorylated peptide is shifted from the mass of the nonphosphorylated version by 79.996 Da by each phosphate group (Fig 1) . In the 40 spot runs, the phosphorylated peptides generally eluted in the same fraction as the nonphosphorylated version, or in the fraction immediately preceding it. The presence of a phosphorylated residue was confirmed by MS/MS, and in most cases the specific residue that was phosphorylated was identifiable. Peptides that are phosphorylated on Ser or Thr residues usually show a strong loss of phosphoric acid (98 Da) when using MALDI and a TOF analyzer in reflector mode, because of postsource decay. The resulting shift in the ions by −18 Da compared with the nonphosphorylated peptide can indicate which residue is phosphorylated ⁴⁶ (Fig. 3) . We were able to identify nine phosphorylated Ser residues in Cx49. Three of the phosphorylated residues were from the cytoplasmic loop region: Ser115, Ser118, and Ser134; the other residues were from the C terminus: Ser258, 261, 265, 266, 297, and 300. We detected phosphorylation on two Ser residues, Ser241 and Ser245, and three Thr residues, Thr238, Thr300, and Thr303 in Cx44, all in the C-tail. One additional Cx44 phosphopeptide was observed. This peptide was phosphorylated on Thr328, Ser 329, or Ser330, but we were unable to determine which specific residue was phosphorylated. 

Chymotryptic Digest

Cx44 has large segments in its C terminus that do not contain Lys or Arg residues, which are required for trypsin cleavage. To obtain coverage of this region of the protein, digestion with chymotrypsin was performed. We were able to sequence eight peptides derived from Cx49 but all of them overlapped regions that had already been covered by the tryptic digest. For Cx44 we were able to identify two peptides from the chymotryptic digest that were derived from regions not covered in the tryptic digest. One fragment, residues 377-388 (HAPPEPPADPGR), was probably generated by residual trypsin in the chymotrypsin preparation. The other peptide, 281-312 (GQASAPGYPEPPPPAALPGTPGTPGGGGNQGL), was found both singly phosphorylated at Thr 303 and doubly phosphorylated at Thr300 and Thr303. 

Potential Kinases

To identify which kinases may be involved in the phosphorylation events we detected in Cx44 and Cx49, we submitted the sequences of both proteins to the NetPhosK 1.0 Web server ⁴⁷ (Tables 3 4) . The server reported several possible sites on the each protein, but only the sites that we have confirmed to be phosphorylated are reported. Several of the phosphorylation sites identified in Cx49 are consistent with sites phosphorylated by PKC. In addition, potential MAP kinase (MAPK) sites were identified in both connexins, as well as casein kinase I sites (CKI). Both MAPK and CKI are known to phosphorylate connexins. ^{29 48} Several sites in Cx44 are also potential glycogen synthase kinase 3 sites. This kinase has not been shown to phosphorylate connexins. 

Discussion

In this study, we identified several posttranslational modifications in the cytoplasmic portions of the bovine lens fiber connexins, Cx44 and Cx49 (Fig. 4) . Our experimental approach involved digestion of the whole membrane fraction followed by identification of peptides from the entire mixture. The advantage of this shotgun approach is that it is unbiased, it avoids loss of material during multiple purification steps, and data can be obtained simultaneously on several different proteins. 

We detected several different posttranslational modifications of the proteins, including phosphorylation, acetylation, truncation, oxidation, and deamidation. Although the acetylation and phosphorylation almost certainly occur in vivo, the oxidation of Met and the deamidation of Asn are less clear-cut phenomena. Some laboratories have shown that insoluble crystallin fractions from the lens have an increase in oxidized residues, ⁴⁹ however Met is readily oxidized, and it is impossible to determine how much of the oxidation occurred during sample preparation, as even spectra taken from the same MALDI target will show an increase in the amount of oxidized Met over time. Similarly, deamidation of Asn occurs readily when it is followed by a Gly residue, as is the case for Asn121 on Cx49, and up to 80% can be deamidated during an overnight digestion. ⁵⁰  

Both Cx44 and Cx49 are cleaved during lens fiber maturation in vivo. Cx49 has been reported to be cleaved by calpain at Glu290 and Ser300. ²⁴ We were able to confirm the cleavage at Glu290, but did not detect a peptide corresponding to a cleavage at Ser300. The two previously reported cleavage sites were identified by N-terminal sequencing of the soluble fragments produced by treatment of isolated ovine lens cortical membranes with rabbit muscle calpain II. It is possible that there is an initial cleavage at Ser300, which is then followed by a second cleavage at the Glu290 site. Consequently, after cleavage most of the membrane-bound Cx49 would terminate at the Glu290. There were two other truncated peptides identified: one truncated at His284 and the other at Phe286. These sites are close to the Glu290 cleavage site and could be nonspecific cleavage products of the calpain. The location of the in vivo cleavage site in the C terminus in Cx44 has not been reported previously. Our data indicate that Leu255 is the Cx44 cleavage site in bovine lens. The trypsin we used is treated with both reductive methylation to inhibit autodigestion and prevent the formation of pseudotrypsin, and with TPCK, which inactivates chymotrypsin. Furthermore, we did not observe chymotryptic activity at other sites. It therefore seems unlikely that this cleavage was due to residual chymotryptic activity. However, unlike the Cx49 Glu290 cleavage site, the sequence surrounding Leu255 in Cx44 is not conserved in the orthologous proteins in other species, which also undergo C-terminal cleavage. In an attempt to confirm the Leu255 cleavage site in Cx44, we digested lens membrane samples with GluC-protein, but we were unable to detect any peptides with the predicted mass. 

We also detected a peptide truncated at Asp124 in the cytoplasmic loop portion of Cx44. This cleavage occurred between Asp124 and Pro125. Asp-Pro bonds are typically very labile in mass spectrometers and disassociate readily in the instrument. However, in this case, the resultant peptide would have a mass of 1857.8 Da and not 1875.8 Da, as we observed. Furthermore, the full-length tryptic peptide eluted two fractions later than the truncated peptide, indicating that the cleavage occurred before the separation was performed. Either the peptide bond is extremely vulnerable, causing cleavage during sample preparation, or the cleavage occurs in the lens in vivo. Yu et al. ⁵¹ reported a cleavage in the cytoplasmic loop of chick Cx45.6 that interferes with its interaction with aquaporin0 (AQP0). However, Cx45.6 is orthologous to Cx49, and not Cx44. This is consistent with the results from a study by Dunia et al. ⁵² suggesting that in mice it is Cx50 that interacts with AQP0, but not Cx46. The phosphorylation sites we identified in the loop portion of Cx49 could regulate this interaction with AQP0, and perhaps the cleavage in the loop of Cx44 has a different function. 

N-terminal acetylation is a common modification in eukaryotic proteins, with some researchers estimating that up to 85% of all proteins in the cell have this modification. ^{53 54} The functional significance of N-terminal acetylation is in most cases unknown. The N termini of connexins play a role in the regulation of channel gating and permeability, ^{55 56 57 58 59 60 61 62 63 64} and for Cx26 and Cx32 it has been suggested that the first 10 residues lie in the channel pore. ⁵⁸ Acetylation may therefore have a role in altering the gating and permeability properties of connexins. It is interesting to note in this context that our data indicate that only a portion of each of the two connexins was acetylated. This result is in contrast to the common observation that acetylation tends to be an all or none phenomenon. For example, when we looked at the N termini of αA-, αB-, βA2-, βB2-, and βB3-crystallin we observed only acetylated ions. This suggests that the acetylation of the two connexins may be posttranslational, rather than cotranslational, and supports the idea that acetylation of the connexins may have a regulatory role. 

Phosphorylation is known to regulate many facets of the connexin life cycle, as well as channel biophysical properties. ^{35 36 37} Zampighi et al. ³⁸ showed that rat Cx50 is normally phosphorylated at low levels on serine residues, but that activation of PKC with TPA induces additional serine phosphorylation, as well as de novo threonine phosphorylation. We similarly detected phosphorylation only on serine residues in Cx49, and most of these were predicted PKC phosphorylation sites. Zampighi et al. ³⁸ also showed that Cx46 is normally heavily phosphorylated on serine residues and that activation of PKC causes additional phosphorylation of threonine residues. One of the interesting observations regarding Cx44 from our study is that some of the peaks from the phosphorylated peptides are more intense than the peaks from the nonphosphorylated ones. Phosphorylation is thought to reduce ionization efficiency, suggesting that in these cases most of the protein in the cell is phosphorylated on these residues, namely Thr238, Ser241, and Ser245. 

None of the phosphorylation sites we have identified corresponds to known cataract-causing human mutations. The closest is a mutation in Cx50 identified in a Russian family with early-onset zonular pulverulent cataract that results in an Ile247Met substitution. ⁶⁵ This position corresponds to a conserved Ile at position 254 in bovine Cx49, which is only four amino acids away from a phosphoserine at position 258. Perhaps the substitution of Met for Ile causes a conformational change in the protein that interferes with this phosphorylation event. 

In summary, we have identified several phosphorylation sites in the lens fiber connexins Cx44 and Cx49. In addition, we have also identified potential in vivo cleavage sites in both connexins, the N-terminal acetylation of both Cx44 and Cx49, and a deamidation in Cx49. Further work should allow us to determine which kinases are responsible for the phosphorylations of specific residues and the functional effects of the phosphorylation of specific residues. 

 Supported by grants from the Natural Sciences and Engineering Research Council of Canada and Genome Canada.

 Submitted for publication September 12, 2007; revised November 7 and December 12, 2007; accepted February 20, 2008.

 Disclosure: D. Shearer, None; W. Ens, None; K. Standing, None; G. Valdimarsson, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Portions of MALDI-MS spectra of trypsin-digested crude membrane preparations from cow lens. (A) A portion of a mass spectrum from one spot of a MALDI target. The spectrum shown is from spot 18, representing the fraction spotted between 18 and 19 minutes after HPLC was initiated. The peak at 1633 is from a Cx44 peptide (residues 228-242), and the two peaks at 1713 and 1793 are from the mono- and diphosphorylated versions of the same peptide. The peaks at 1696 and 1776 are from the same peptide with the initial glutamine residue modified to pyroglutamic acid. (B) A portion of a mass spectrum showing a peak at 2472 from the N terminus of Cx49 (residues 2-22). The peak at 2514 from the subsequent fraction (C) is the acetylated peptide of the peak at 2472 in (B).

Table 1.

 

The Peptides from the Tryptic Digests Assigned to Cx44

Residues	Sequence	Predicted M+H Mass	Measured Mass	Log GPM e-Value
2-9	GDWSFLGR	937.45	937.45, 979.50*	−2.0
106-115	RKEREEEPPK	1297.69	1297.67	−1.2
106-132	RKEREEEPPKAAGPEGHQDPAPVRDDR	3066.51	3066.66	−2.7
107-115	KEREEEPPK	1141.59	1141.57	−1.1
108-115	EREEEPPK	1013.49	1013.47	−2.3
108-124	EREEEPPKAAGPEGHQD	1875.85	1875.84	−9.2
108-129	EREEEPPKAAGPEGHQDPAPVR	2396.16	2396.17	−8.1
108-132	EREEEPPKAAGPEGHQDPAPVRDDR	2782.31	2782.38	−1.9
110-115	EEEPPK	728.35	728.34	>−1
110-129	EEEPPKAAGPEGHQDPAPVR	2111.02	2111.01	−7.0
110-132	EEEPPKAAGPEGHQDPAPVRDDR	2497.17	2497.18	>−1
116-129	AAGPEGHQDPAPVR	1401.69	1401.69	−3.6
116-132	AAGPEGHQDPAPVRDDR	1787.84	1787.84	−1.8
116-134	AAGPEGHQDPAPVRDDRGK	1972.96	1972.95	−3.6
144-152	TYVFNIIFK	1144.64	1144.63	−3.0
226-242	LKQGMTSPFRPDTPGSR	1874.96	1874.94, 1954.91, †, 2034.89, ‡, 1890.94, §, 1970.92, † , §, 2050.89, ‡ , §	−2.3
228-242	QGMTSPFRPDTPGSR	1633.775	1633.79, 1713.76, †, 1793.736, ‡, 1649.78, §, 1729.74, † , §, 1809.71, ‡ , §, 1616.74, ∥, 1776.68, † , ∥	>−1
243-255	AGSVKPVGGSPLL	1181.689	1181.70, 1261.66, †	−4.4
316-331	AQNWANREAEPQTSSR	1845.847	1845.85, 1925.85, †	−4.3

 The GMP log e-values reported are for the unmodified peptides. All MS/MS data indicating modified peptides were manually confirmed by visual inspection of the data.

*  Acetylated.

†  Singly phosphorylated.

‡  Doubly phosphorylated.

§  Oxidized Met.

∥  Glu-to-pyroglutamine conversion.

Table 2.

 

Peptides from the Tryptic Digests Assigned to Cx49

Residues	Sequence	Predicted M+H Mass	Measured Masses	Log GPM e-Value
2-22	GDWSFLGNILEEVEHSTVIGR	2472.28	2472.28, 2514.23*	−6.1
108-125	EREAEELSQQSPGNGGER	1972.88	1973.89, †, 2053.85, † , ‡	−4.1
110-125	EAEELSQQSPGNGGER	1687.74	1688.75, †, 1768.70, † , ‡, 1848.683, † , §	−5.7
110-137	EAEELSQQSPGNGGERAPLAADQGSVKK	2853.40	2854.38&, 2934.34, † , ‡, 3014.34, † , ‡	−2.0
126-137	APLAADQGSVKK	1184.66	1184.67, 1264.63, ‡	−5.9
148-156	FRLEGTLLR	1104.65	1104.66	−1.1
241-257	SAFKRPVEQPLGEIPEK	1925.05	1925.07	−2.0
241-269	SAFKRPVEQPLGEIPEKSLHSIAVSSIQK	3175.75	3175.75, 3255.62, ‡	−4.8
245-157	RPVEQPLGEIPEK	1491.82	1491.86, 1492.84	−1.9
245-269	RPVEQPLGEIPEKSLHSIAVSSIQK	2742.82	2742.61, 2822.50, ‡, 2902.45, §, 2982.41, ∥, 3062.44, ¶	−3.0
258-269	SLHSIAVSSIQK	1269.29	1269.72	−6.5
258-271	SLHSIAVSSIQKAK	1468.85	1468.91	−3.8
258-276	SLHSIAVSSIQKAKGYQLL	2043.16	2043.16	−2.9
258-284	SLHSIAVSSIQKAKGYQLLEEEKIVSH	2994.63	2994.63	−8.7
258-290	SLHSIAVSSIQKAKGYQLLEEEKIVSHYFPLTE	3744.99	3744.98	−4.3
270-290	AKGYQLLEEEKIVSHYFPLTE	2494.29	2494.28	−7.3
272-280	GYQLLEEEK	1108.55	1108.55	−3.5
272-284	GYQLLEEEKIVSH	1544.80	1544.79	−5.6
272-286	GYQLLEEEKIVSHYF	1854.93	1854.93	−4.6
272-290	GYQLLEEEKIVSHYFPLTE	2295.15	2295.16	−6.8
281-290	IVSHYFPLTE	1205.62	1205.62	−1.3
281-310	IVSHYFPLTEVGMVEASPLSAKPFSQFEEK	3367.69	3367.69, 3383.70, #, 3447.67, ‡, 3527.65, §	−8.5
311-321	VGPGPLGDLSR	1067.59	1067.55	−3.5
322-356	AYQETLPSYAQVGAQEGVEEEQPVEAAAEPEVGEK	3732.74	3732.72	−8.2
363-378	VSTEGQETAVLEVEK	1731.90	1731.88	−2.5
379-397	VEPPEVEKEVEKEEPPPEK	2218.11	2218.11	−8.8
387-397	EVEKEEPPPEK	1310.65	1310.65	−2.7

 The GMP log e-values reported are for the unmodified peptides. All MS/MS data indicating modified peptides were manually confirmed by visual inspection of the data.

*  Acetylated.

†  Mass of deamidated ions.

‡  Singly phosphorylated.

§  Doubly phosphorylated.

∥  Triply phosphorylated.

¶  Quadruply phosphorylated.

#  Oxidized Met.

Figure 2.

 

MS/MS spectra of the lens fiber connexin N-terminal peptides. (A) A spectrum of the N-terminal peptide of Cx44. (B) A spectrum of the acetylated N-terminal peptide from Cx44. The b-ions were shifted up by 42 Da, whereas the y-ions were the same mass as the nonacetylated peptide. M+H denotes the parent ions.

Figure 3.

 

MS/MS spectra of the peptide of amino acids 243-255 from Cx44. (A) Spectrum from the unmodified peptide; (B) phosphorylated peptide. The spectrum of the phosphorylated peptide shows the strong loss of phosphoric acid at 98 Da that indicates phosphorylation. The stronger intensity of the b-98 ions while the Y8 ion remains unchanged indicates that the first serine is the phosphorylated residue.

Table 3.

 

Kinases Predicted to Phosphorylate Sites Detected in Cx44

Residue in Protein	Predicted Kinase
Thr238	GSK3, MAPK
Ser241	GSK3
Ser245	PKA, PKC
Thr300	GSK3, MAPK
Thr303	GSK, MAPK, CK1
Thr328 or Ser(329,330)*	PKC

*  Only one of the residues from 328 to 300 was phosphorylated, and it was not possible to determine the specific residue that was phosphorylated.

Table 4.

 

Kinases Predicted to Phosphorylate Sites Detected in Cx49

Residue in Protein	Predicted Kinase
Ser115	CK1, ATM, DNAPK
Ser118	CK1
Ser134	PKC
Ser258	PKC
Ser261	PKC
Ser265	PKC
Ser266	PKC
Ser297	MAPK
Ser300	PKC

Figure 4.

 

Diagram showing in vivo posttranslational modifications of Cx44 and Cx49. Shaded areas represent areas of the proteins that were undetectable. In vivo cleavage sites (shaded diamonds) and phosphorylation sites (P) are indicated.

The authors thank Vic Spicer and Oleg Krokhin (University of Manitoba) for assistance during the course of the work.