The sequence of human betaB1-crystallin cDNA allows mass spectrometric detection of betaB1 protein missing portions of its N-terminal extension.

The sequence of human betaB1-crystallin cDNA encoded a protein of 251 amino acids in length. Mass spectrometric analysis of intact betaB1 from young human lens confirmed the deduced amino acid sequence. Lenses of human donors newborn to 27 years of age also contained partially degraded forms of betaB1 missing 15, 33, 34, 35, 36, 39, 40, and 41 amino acid residues from their N-terminal extensions. The similarity of the cleavage site between residues 15 and 16 in human betaB1 to the cleavage occurring in bovine betaB1 suggested that lenses of both species may contain a similar proteolytic activity. The remaining cleavage sites occurring in human betaB1 did not closely match those occurring in other species, possibly due to the widely divergent amino acid sequence of the N-terminal extension of betaB1 amoung species. Results from animal models suggest that cleavage of the N-terminal extension of betaB1-crystallin could enhance protein insolubilization and cataract in lens. However, the presence of partially degraded betaB1-crystallins in both water-soluble and water-insoluble fractions of lenses of young donors suggested that the rate that proteolyzed betaB1-crystallins become water-insoluble is relatively slow in humans.

The lens is a unique organ in that 20 -60% of its wet weight is composed of proteins called crystallins (1). Mammalian crystallins can be divided into two broad classes, the ␣-crystallins, which share homology to heat shock proteins and exhibit chaperone-like properties (2), and the ␤/␥-crystallin family, which like ␣-crystallins, also share a common ancestor but perform primarily a structural role. The individual proteins of ␤/␥crystallin family contain four homologous motifs, each folded in a "Greek key" pattern (3). These motifs are organized into two equivalent domains, which are connected by approximately 4 amino acid residues, the entire structure thus forming the core of the protein. There are seven different ␤-crystallin proteins in bovine lens. These are named ␤B1, ␤B2, ␤B3, ␤A1, ␤A2, ␤A3, and ␤A4 (4). Each crystallin transcript is translated to yield a single protein species, except the mRNA coding for ␤A3, which also yields ␤A1, owing to an alternate downstream initiation codon (5). Orthologous ␤-crystallins are also found in other vertebrate species. Except for ␤A2, orthologs of all known ␤-crystallin proteins have been demonstrated in rat (6). While the presence of all protein products has not been established in the chicken, the cDNAs of all six ␤-crystallin transcripts have been sequenced (7). In human lens, the presence of ␤B1 (8), ␤B2 (9), ␤A3 (10), and ␤A4 1 have been documented. However, to date, there is no evidence for the significant accumulation of ␤A1, ␤A2, or ␤B3 proteins in postnatal human lens.
All ␤-crystallins contain N-terminal extensions ranging from 10 to 58 amino acids in length as well C-terminal extensions of approximately 15 amino acids in the ␤B-crystallin subgroup (4). These N-and C-terminal extensions are missing in ␥-crystallins. Since ␤-crystallins form complex mixtures of homodimers, heterodimers, and higher order structures, while ␥-crystallins do not, it has been hypothesized that the extensions function to stabilize the intermolecular associations between ␤-crystallins (11). It is likely not a coincidence that ␤B1, which contains the longest N-terminal extension, is also selectively found in the largest ␤-crystallin aggregate (12). Thus, recent work has explored the possible function of ␤-crystallin N-terminal extensions in stabilizing the structure of ␤-crystallin oligomers (13,14,16,17). We hypothesize that the presence of ␤-crystallin N-terminal extensions is important, because removal of from 2 to 49 residues from the N-terminal extensions of various ␤-crystallin subunits in lenses of young rodents was associated with protein insolubilization and cataract formation (6,18). The mechanism of protein insolubilization following shortening of the N-terminal extensions is unknown. However, it presumably involves an alteration in the manner the ␤-crystallin subunits oligomerize. The specific cleavages in the N-terminal extensions may result from activation of the protease calpain II (18). Similar cleavage of ␤-crystallin Nterminal extensions and insolubilization of ␤-crystallins also occurs during maturation of normal rat lens (19). However, these older lenses may remain transparent due to the relatively slower rate that the process ensues during normal development.
A major focus of these laboratories is to determine if similar partial degradation and insolubilization of ␤-crystallins also occurs in human lenses. In human lenses, the amount of watersoluble ␤-crystallin decreases dramatically with aging, and partial degradation of ␤B2 has been reported (20). Similarly, a 29-kDa ␤-crystallin identified as ␤B1, was prominent in fetal human lenses, but it was nearly gone by age 5 (21). Thus, evidence exists for the partial degradation of ␤-crystallins during aging of human lenses. However, few detailed studies exist that examine cleavage sites in human ␤-crystallins. These studies are limited by the lack of sequence data for human * This work was supported by National Institutes of Health Grants EY07755 (to L. L. D.) and EY07609 (to J. B. S.) and by a Research to Prevent Blindness Award (to L. L. D.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM  ␤-crystallins. Only the complete sequences of human ␤B2 and ␤A3 are known (22,23). This study reports the sequence of human ␤B1 cDNA and confirms the deduced amino acid sequence by mass spectrometry. With the knowledge of the sequence, we then found that partially degraded forms of ␤B1 missing various portions of its N-terminal extension were present in both water-soluble and water-insoluble fractions of human lens.

MATERIALS AND METHODS
Cloning and Sequencing of Human ␤B1 cDNA-Total RNA was isolated from both freshly enucleated or Ϫ70°C frozen lenses of human donors of 10 years of age or less (Lions Eye Bank, Portland, OR) (24). cDNA was produced by reverse transcription of total lens RNA using an oligo(dT) containing adapter primer (3Ј-RACE 2 system, Life Technologies, Inc.). Following degradation of the RNA template by RNase H, the total lens cDNA was purified by GlassMAX spin cartridges, and the cDNA was then homopolymer tailed using dCTP and terminal deoxynucleotidyl transferase (5Ј-RACE system, Life Technologies, Inc.). Cycling conditions for all PCR used 94°C for 5 min, (94°C for 45 s, 64°C for 45 s, 72°C for 60 s) ϫ 35 cycles, ending with 72°C for 10 min. Conditions recommended by the 5Ј-RACE system were used (Life Technologies, Inc.), except AmpliWax beads were employed (Perkin Elmer Corp.). The lens cDNA was subjected to 5Ј-RACE PCR using the 5Ј-RACE anchor primer, and the gene-specific antisense primer TT-GAAGTTGGCCCCTTCAAACAGGCAGA (primer 1) complementary to residues 293-320 of rat ␤B1 cDNA (25) and residues 471-498 bovine ␤B1 cDNA (26). The resulting approximately 500-base pair product was cut out from a 1.5% agarose gel and purified using Prep-A-Gene matrix (Bio-Rad). A second 5Ј-RACE PCR reaction was carried out using the 5Ј-RACE anchor primer, and a nested gene-specific antisense primer TGATGGGCCGGAAGGACATGAGCCGGTC (primer 2) complementary to residues 241-268 of rat lens ␤B1 cDNA (25) and residues 419 -446 of bovine lens ␤B1 cDNA (26). The resulting product was directly sequenced using the AutoRead sequencing kit and automated laser fluorescence DNA analysis system (Pharmacia Biotech, Inc.). The sequencing reaction was primed using a 5Ј fluorescein-labeled primer 2. The resulting sequence was used to choose two human ␤B1-specific primers, a sense primer CUACUACUACUAAGTGCTCAAATCTG-GCAGAC (primer 3) corresponding to residues 308 -327 of human ␤B1 cDNA, and an antisense primer CUACUACUACUAGCGAGGG-TACTCGCCCTTCT (primer 4) complementary to residues 422-441 of human ␤B1 cDNA (note: the four CUA repeats on the 5Ј-ends were added for subsequent cloning of the PCR product). The human ␤B1 cDNA 5Ј-RACE PCR product produced above was subjected to a further 5Ј-RACE PCR reaction using the nested human ␤B1-specific antisense primer 4 and the 5Ј-RACE anchor primer. 3Ј-RACE PCR of human ␤B1 c-DNA was carried out using the RACE ready human lens cDNA described above, the 3Ј-RACE universal amplification primer, and the human ␤B1 sense primer 3. The protocol utilized the cycling conditions described above and the 3Ј-RACE system from Life Technologies, Inc. The 5Ј-and 3Ј-RACE PCR products were then cloned by treatment with uracil DNA glycosylase, ligation into plasmid pAMP10, and transformation of DH5␣-competent cells using the CloneAmp pAMP10 system (Life Technologies, Inc.). Transformants were selected by ampicillin resistance, and plasmid DNA were isolated using a FlexiPrep Kit (Pharmacia). Both sense and antisense strands of three different 5Ј-RACE clones and three different 3Ј-RACE clones were sequenced using fluorolabled M13 forward and reverse primers as described above. The complete sequence of human ␤B1 cDNA was assembled by aligning the overlapping sequences of the 5Ј-and 3Ј-RACE clones.
Isolation of ␤B1 from Newborn Human Lenses-The water-soluble protein from whole decapsulated lenses from a newborn human donor was separated by two-dimensional electrophoresis and transferred to polyvinylidine difluoride membranes as described previously (18), except 250-g samples were applied to each gel. Intact ␤B1 and its degradation product were individually excised from the membranes of 12 gels, and digested with trypsin in situ on the membrane surface as described previously (6).
Confirmation of the Deduced Protein Sequence of ␤B1-The tryptic digests of the proteins on the two-dimensional gel corresponding to ␤B1 and its degradation product from newborn human lens were each ana-lyzed using on-line reversed-phase HPLC coupled to an electrospray ionization mass spectrometer (ESIMS) (Fisons VG-platform quadrupole mass spectrometer, Manchester, United Kingdom). The peptides were fractionated on a C-18 microbore column (50 ϫ 1 mm), with a 40-min linear gradient of 0 -60% CH 3 CN in water, both solutions with 0.1% trifluoroacetic acid. The flow of 50 l/min was divided by a postcolumn splitter, 90% to a UV detector and fraction collector, and 10% to the mass spectrometer. Spectra were obtained over a mass range of 300-2000 Da. The instrument was calibrated using the peaks of NaI. Mass accuracy was within 0.3 Da. Confirmation of the N-terminal cleavage site in the partially degraded ␤B1 was carried out by Edman sequencing of the protein blotted onto polyvinylidine difluoride membrane (18).
Isolation of the ␤-Crystallins from Young Adult Lenses-Three young adult lenses were examined to determine the presence of ␤B1-crystallin in the water-soluble and water-insoluble portions of the lens crystallins and to ascertain the N-terminal cleavage sites. The lenses, obtained from the National Disease Research Interchange, were from healthy accident victims, ages 16, 20, and 27, with no known drug usage. Soluble protein was obtained from each lens, and ␣-, ␤and ␥-crystallins were isolated as described previously (27). The water-insoluble portion of the lens homogenate was treated with 6 M guanidine hydrochloride, 50 mM Tris, 1 mM EDTA, 0.02% NaN 3 , pH 7.4. Solubilization was assisted by grinding with a pellet pestle (Kontes Scientific Glassware/ Instruments, Vineland, NJ). Following centrifugation at 15,000 ϫ g for 15 min, the guanidine-soluble portion was separated into portions with molecular weights corresponding to monomers, dimers, and higher molecular weight crystallins by gel filtration chromatography on a 1.5 ϫ 112-cm column of Sephacryl S-300 (Sigma) with a flow rate of 6.8 ml/h using the solubilization buffer as the eluent. The column was calibrated with cyctochrome c (12.4 kDa), carbonic anhydrase (29 kDa), and bovine serum albumin (66 kDa). Only the monomeric portion of the water-insoluble crystallins was examined.
Detection of ␤B1 in the Water-soluble and Water-insoluble Fractions by Analysis of Tryptic Fragments-Total ␤-crystallins (40 g) from the 2 The abbreviations used are: RACE, rapid amplification of cDNA ends; PCR, polymerase chain reaction; HPLC, high performance liquid chromatography; ESIMS, electrospray ionization mass spectrometer. water-soluble portion of the 27-year-old lens were digested with trypsin (50:1, substrate/enzyme) at pH 8.2 for 4 h. The resulting peptides were separated by reversed-phase HPLC using a Vydac C-18 column and a linear gradient of 1-40% CH 3 CN over 60 min and then to 65% CH 3 CN at 70 min. The fractions were collected, and the molecular masses of the peptides present in each fraction were determined by fast atom bombardment mass spectrometry (FABMS) (Kratos MS-50, Kratos Analytical, Manchester, United Kingdom) or by ESIMS. The water-insoluble guanidine-soluble portions of all three young adult lenses were examined for the presence of intact and partially degraded ␤B1. The monomeric portion isolated by gel filtration was fractionated by reversephase HPLC, resulting in five poorly resolved peaks (28). Each peak, except for the first for which there was insufficient material, was digested with trypsin as described previously for the water-soluble portion, and the molecular masses of the resulting peptides were determined by fast atom bombardment mass spectrometry or ESIMS. Accuracy of the molecular mass determinations was within 0.3 Da. Identification of peptides was facilitated by computer programs written in our laboratory (29). When a peptide identification based on molecular mass was ambiguous, the portion (90%) of the sample that had been diverted before the ESIMS analysis was further analyzed 1) by peptide sequencing (Purdue Laboratory for Macromolecular Structure), 2) by further digestion with carboxypeptidase or pepsin followed by mass spectrometric analysis of the products, or 3) by analysis of the MS/MS fragmentation pattern of the peptide. MS/MS analyses were performed on a Fisons VG-Autospec Mass Spectrometer equipped with an orthogonal acceleration time-of-flight analyzer.
ESIMS Analysis of ␤-Crystallins without Tryptic Digestion-Undigested ␤ H -crystallins, isolated from a 20-year-old lens by gel filtration, were analyzed by reverse-phase HPLC/ESIMS to obtain the molecular masses of the proteins eluting at 43-45% CH 3 CN where ␤B1 was expected to elute. Because this sample was a mixture, the molecular masses of only the major components are reported. Mass accuracy for the ESIMS determinations was within 0.025%.

RESULTS
The Nucleotide and Deduced Amino Acid Sequence of Human ␤B1 cDNA and Protein-The cDNA of human ␤B1-crystallin was amplified by 5Ј-and 3Ј-RACE PCR, ligated into plasmid vector pAMP10, and both strands of each insert were sequenced. Excluding the dC homopolymer region introduced for 5Ј-RACE PCR and the poly(A) tail on the 3Ј-end, the overlapping sequences of the 5Ј-and 3Ј-RACE PCR products resulted in the 923-base pair sequence shown in Fig. 1 (GenBank accession number U35340). The initiation codon began at nucleotide 73, resulted in an open reading frame of 756 nucleotides, and ended with a stop codon at nucleotides 829 -831. The 3Ј-end of the cDNA contained a 92-base pair noncoding region, and a poly(A) addition site at nucleotides 905-910. The 5Ј-end of the cDNA contained a 72-base pair noncoding region, suggesting that the 5Ј-end of the cDNA may be complete. Excluding the N-terminal methionine, the human ␤B1 cDNA coded for a protein 251 amino acids in length.
There are evolutionary constraints upon the sequence of the core protein of ␤-crystallin subunits that are required for proper folding of the four Greek key motifs, and maintenance of proper interactions at the subunit interfaces of ␤-crystallin dimers (25,30,31). Because of this, amino acid residues 58 -234, which contained motifs 1-4 and the connecting peptide of human ␤B1, exhibited 89, 88, and 77% sequence identity, respectively, with the reported sequences of orthologous rat, bovine, and chicken ␤B1 (25,26,32). As previously noted during comparison of rat, bovine, and chicken ␤B1 sequences (30), the N-terminal extension of human ␤B1 exhibited the greatest divergence in sequence between species (Fig. 2). The human ␤B1 N-terminal extension exhibited only 61, 56, and 18% sequence identity with bovine, rat, and chicken ␤B1 N-terminal extensions, respectively. Most notable was the loss of the Pro-Ala repeats found between residues 35 and 44 in bovine ␤B1, and the poor conservation of the Pro-X repeats found between residues 29 -38 in rat ␤B1 (Fig. 2). Similarly, the bovine, rat, and chicken ␤B1 C-terminal extensions shared only 78, 72, and 33% sequence identity with the human ␤B1 C-terminal extension, respectively. Confirmation of the Deduced ␤B1 Amino Acid Sequence by Mass Analysis of ␤B1 from Newborn Human Lenses-Watersoluble protein from newborn human lenses was separated by two-dimensional electrophoresis, and individual protein species were identified by either direct Edman sequence analysis, or Edman sequence analysis of tryptic fragments followed by comparison to deduced amino acid sequences from reported cDNA sequences. 1 The identities of the major protein species are indicated in Fig. 3. The partially degraded ␤B1-crystallin labeled ␤B1 (Ϫ15) was identified based on the N-terminal sequence PGPDT, which indicated it was missing 15 residues from its N terminus (Fig. 2).
The regions containing intact ␤B1 and ␤B1 (Ϫ15) were digested with trypsin, and the masses of the resulting fragments were analyzed. The peptides from these digests gave masses corresponding to all expected tryptic peptides in the mass range analyzed, confirming the deduced amino acid sequence (Fig. 4). The mass of 545 Da for the N-terminal peptide indicated that the N-terminal methionine was removed, and that the N-terminal Ser was acetylated. The mass of the peptide containing Cys-79 appeared 71 Da higher than expected due to the formation of an acrylamide adduct during the two-dimensional gel separation (33). The deduced protein sequence in Fig.  1 has a calculated molecular mass of 27,935 Da, agreeing very well with the ESIMS-determined molecular mass of 27,933 Da reported by He et al. (10). Mass spectral analysis of ␤B1 (Ϫ15) also gave all the expected tryptic peptides, except those corresponding to residues 1-5 and 6 -21. The presence of a peptide with a molecular mass corresponding to peptide 16 -21 (613 Da) confirmed the Edman sequencing data, indicating that the protein had been cleaved between Asn-15 and Pro-16 (Fig. 4). A peptide with a mass corresponding to the C-terminal tryptic fragment indicated the presence of an intact C terminus in the young lens.
Detection of Additional Truncated Forms of ␤B1 in the Water-soluble Crystallins of Lenses of Increasing Age-The HPLC/ ESIMS analysis of the first eluting ␤-crystallin peak from gel filtration (also called ␤ H ) from the 20-year-old lens indicated the presence of many components. The masses of the major components suggested that several proteins were degraded forms of ␤B1-crystallin. Masses consistent with intact ␤B1, as well as degradation products due to loss of 15, 34, 39 and 40 residues from the N terminus, were present (Table I) Mass spectral analysis of the tryptic peptides produced from the water-soluble lens crystallins of young adult (age 27) suggested that additional forms of partially degraded ␤B1-crystallin were present. Even though this digest was a mixture of peptides from several proteins, peptides encompassing over 75% of ␤B1-crystallin could be identified from comparisons of their molecular masses and HPLC elution times with those found in the digest of ␤B1 from the newborn lens. Expected peptides derived from the N terminus of intact ␤B1 and ␤B1 (16 -251) were not evident, probably because the N-terminal peptides from these species were present at very low levels, and the mass spectral response was masked by the presence of other more responsive peptides. However, the ESIMS data summarized above indicated that cleavage at Lys-49 by trypsin should produce peptides corresponding to peptides 35-49, 40 -49, and 41-49. Ions with molecular masses corresponding to these three peptides were found. In addition, the digest contained ions with molecular masses consistent with the presence of peptides 34 -49, 36 -49, 37-49, and 42-49, indicating that there are numerous other N-terminal degradation sites of ␤B1crystallin evident by age 27.
Detection of ␤B1 in the Water-insoluble Crystallins-The water-insoluble portions of all three young adult lenses were examined for the presence of ␤B1-crystallin. The results were very similar to the water-soluble ␤-crystallin fraction of adult lenses. Ions with masses corresponding to the majority of peptides from ␤B1 were identified. However, the data suggested  4. Masses of peptides found in the tryptic digests of ␤B1-crystallins. The masses in boldface were found in the digest of the intact ␤B1 from a newborn lens, confiming the deduced amino acid sequence of ␤B1 shown in Fig. 1. a, the 613-Da tryptic peptide was found in the digest of the protein marked ␤B1 (Ϫ15) in Fig. 3, thus confirming the cleavage site in this protein between residues 15 and 16. b, masses of 2 out of 7 other tryptic fragments, indicating further cleavage of ␤B1 during maturation. The identities of these tryptic fragments and others with asterisks were confirmed by other experiments discussed in the text. that most of the N-terminal extension of insoluble ␤B1 was missing. Neither peptide 1-5, peptide 6 -21, nor peptide 16 -21 were detected, and the response for peptide 24 -49 was very weak. Also, the presence of peptides corresponding to residues 41-49 (916 Da) and 42-49 (815 Da) indicated N-terminal cleavage at these sites.
Because the tryptic digests of water-insoluble proteins were a complex mixture, the identities of many of the peptides, particularly those indicating the presence of N-terminal degradation, were confirmed. One of three methods was used for confirmation: 1) further digestion with another enzyme, such as pepsin or carboxypeptidase, followed by a second mass spectral analysis of these products; 2) Edman sequencing, if the peptide was pure; or 3) MS/MS analysis of the collision-induced fragmentation pattern of the peptide. Analysis of the MS/MS fragmentation patterns of the peptides at 916 Da and 815 Da, isolated from the 16-year-old lens, showed several fragments that confirmed the identities of the peptides as residues 41-49 and 42-49, respectively. The MS/MS spectrum for peptide 41-49 is shown in Fig. 6.

DISCUSSION
The major finding of this study was that ␤B1-crystallin of human lens undergoes extensive cleavage at its N-terminal extension during lens maturation. The first cleavage product of human ␤B1, ␤B1 (16 -251), was already abundantly evident in lenses of a newborn donor. This suggested that ␤B1 may be the most protease susceptible crystallin in the human lens.
Alcala et al. (21) showed that from 8 months of gestation to 5 years of age, the percentage of intact ␤B1 in the water-soluble protein of human lens dropped from 10 to 0.5%. During the same interval, the percentage of a 27-kDa protein, corresponding to ␤B1 (16 -251) of the present study, increased from 3.5 to 7%. Thereafter, the 27-kDa protein decreased, so that by 87 years of age it composed only 1.2% of the water-soluble lens protein. The loss of intact ␤B1 and transient accumulation of the 27-kDa protein also correlated with the age of the lens fiber. The conversion from intact ␤B1 to 27-kDa protein and the subsequent loss of 27-kDa protein was more pronounced in the deeper lens cortical and nuclear fibers than in the superficial cortex. These results are consistent with the present findings.
␤B1 and ␤B1 (16 -251) were most abundant in newborn lenses. However, by early adulthood, tryptic fragments corresponding to the N terminus of intact ␤B1 and ␤B1 (16 -251) were not detected, while tryptic fragments from forms of ␤B1 with more extensively degraded N termini were abundant.
Bovine ␤B1 also undergoes a similar cleavage as that producing human ␤B1 (16 -251). Bovine ␤B1 is cleaved between Asn-14 and Pro-15 (34). The 6 amino acid residues surrounding this cleavage site are identical between bovine and human ␤B1 (Fig. 2). This suggested that both species contain a similar protease capable of cleaving at this site. Rat ␤B1 does not undergo cleavage at this site, possible due to the loss of the Pro residue found at positions 18 and 17 in human and bovine ␤B1, respectively (Fig. 2). Bovine ␤B1 also contained a cleavage site between Ala-11 and Ala-12 (34). Neither mass spectrometric analysis or Edman sequencing in the present study detected a similar cleavage in human ␤B1, possibly due to the lack of sequence identity in this region between bovine and human ␤B1.
Further cleavage of human ␤B1 occurred between residues 33 and 41. ESIMS analysis of ␤ H -crystallins from a 2-year-old bovine lens did not yield molecular masses indicative of similar N-terminal degradation in this region. Instead, the bovine ␤B1 from a 2-year-old lens appears to have undergone even more extensive degradation, since the molecular masses of the proteins in ␤ H -crystallin were all less than 25,000 Da (35). Bovine lens ␤B1 may not have undergone cleavage in regions similar to human ␤B1, because residues 35-44 in bovine ␤B1 contained Pro-Ala repeats not found in human ␤B1. Pro-Ala repeats in bovine ␤B1 and Pro-X repeats in rat ␤B1 may be more resistant to proteolytic attack than the corresponding region in human ␤B1.
Cleavage at five sites within the N-terminal extension of ␤B1 has also been reported during maturation of rat lens (Fig. 2). At least two of these sites in rat ␤B1 are confirmed calpain II cleavage sites (19). However, none of the cleavage sites within the N-terminal extension of rat ␤B1 corresponded in relative position to cleavage sites found in human ␤B1 (Fig. 2). The difference in the manner rat ␤B1 and human ␤B1 are degraded could be due to a lack of sequence identity at the respective cleavage sites. Sequences within the N-terminal extensions of ␤B1 from various species may have evolved to exhibit specific susceptibilities to proteolytic attack. Alternatively, the lenses of each species may contain proteolytic activities with different specificities. The predominate proteolytic activity responsible for partial degradation of ␤-crystallin N-terminal extensions in rat lens may be calpain II (19). However, the proteases responsible for degradation of the N-terminal extension of bovine and human ␤B1 remain unknown.
The extensive cleavage of the N-terminal extension of ␤B1 occurs quite early in life. This suggested that the cleavage plays an important role in the maturation process. In rat, the partially degraded forms of ␤B1 were found in only the waterinsoluble fraction of the lens (19). Therefore, the cleavage of the rat ␤B1 N-terminal extension, as well as the N-terminal extensions of other ␤-crystallins, may rapidly induce protein insolubilization (6). Such insolubilization may be important in initiating dehydration and hardening of the rat lens (36). However, the results in the present study indicated that human ␤-crystallins respond quite differently following partial proteolysis. Partially degraded ␤B1 was found in both water-soluble and water-insoluble fractions of young human lens. Also, the majority of human ␤B1-crystallin is partially degraded before adulthood, but most crystallin insolubilization occurs in human lenses after the third decade of life (15). Future studies will determine if the water-insoluble fraction of aged lenses contains a greater proportion of partially degraded ␤-crystallins than does the water-soluble fraction. Also, the relationship between the extent of proteolysis and lens opacification requires closer examination.