After proinsulin is translocated into the lumen of the ER, it is then delivered in transport “COP-coated vesicles” to the cis-Golgi apparatus  (Figure 6.4). Up until relatively recently, it was thought that newly synthesized proinsulin was passed from the cis-Golgi network “stack” via the medial- to the trans-stack of the Golgi apparatus stacks in “COP”-coated vesicles, but a landmark study conducted in β cells using electron microscope tomography showed that the Golgi apparatus is actually one continuous organellar compartment, and not a series of stacks . As such, newly synthesized proinsulin traverses through the lumen of the β cell’s Golgi apparatus to the trans-Golgi network (TGN) , where it accumulates in clathrin-coated regions . This is the site of secretory granule biogenesis (Figure 6.4). The means by which newly synthesized proinsulin (and other select proteins destined to the β granule), is specifically targeted to sites of β-granule biogenesis in the TGN remains a matter of debate . However, it is known to be a highly efficient process, with >99% efficient of newly synthesized proinsulin sorted to the β granule and regulated secretory pathway under normal conditions [91,95].
Generally, analogous to other neuroendocrine cells, the process of β-granule biogenesis should also require other factors including intraluminal acidic pH 6,5, Ca2+, ATP, GTP-hydrolysis cytosolic proteins and perhaps protein tyrosine phosphorylation [88,91,93,95]. Although β-granule biogenesis occurs in limited clathrin-coated regions of the TGN , the role that clathrin itself plays is unclear although is likely involved in the process of a newly formed immature β granule “budding off” the TGN. An immature β granule then undergoes a maturation process [93,95]. Maturation of β granules involves proinsulin conversion, progressive intragranular acidification, loss of the clathrin-coated regions, and formation of hexameric insulin crystals [91,93]. Acidification provides the correct intragranular pH (pH 5.0 – 5.5) for proinsulin processing to proceed , and optimal insulin crystal formation around insulin’s isoelectric point (pKi 5.3) . Delivery of newly synthesized proinsulin to an immature β granule occurs around 30 – 40 min post-translation where proinsulin processing begins and is >90% completed ∼3 h later [91,93] (Figure 6.4).
Proteolytic enzymes of proinsulin conversion
The major site for processing of proinsulin to biologically active insulin is the immature secretory granule compartment of the β cell [91,93] (Figure 6.4). Production of insulin (and C-peptide) occurs via limited proteolysis of the proinsulin precursor molecule, which is catalyzed by two Ca2+-dependent endopeptidases, PC2 and PC1/3 and a Ni2+-dependent exopeptidase, CP-H [88,91]. There are two dibasic sites on the human proinsulin molecule: Arg3, Arg32 and Lys64, Arg65, that signal limited endoproteolytic cleavage of proinsulin to excise the C-peptide moiety and to generate insulin with its disulphide-linked A- and B-chains correctly aligned (Figure 6.3). Endoproteolytic peptide bond cleavage of proinsulin occurs on the carboxylic side of the Arg31, Arg32 or Lys64, Arg65, followed by rapid and specific exopeptidic removal of the newly exposed basic amino acids by CP-H [88,91]. The two distinct β-granule proinsulin-processing endopeptidase activities were originally discovered as Ca2+-dependent with an acidic pH optimum and were later identified as the PC1/3 and PC2 endopeptidase genes [88,91].
A scheme of proinsulin conversion is illustrated in Figure 6.5. Proinsulin conversion could occur by two possible routes. Either PC2 first cleaves on the carboxylic side of Lys64-Arg65 to yield a split 65,66 proinsulin intermediate, followed by CP-H trimming of the newly exposed lysine and arginine residues to yield des 64,65 proinsulin. Then PC1/3 can then cleave des 64,65 proinsulin at Arg31, Arg32, which together with CP-H trimming of the exposed arginine residues, yields insulin and C-peptide (Figure 6.5). Alternatively, PC1/3 first could cleave at the carboxylic side of Arg31, Arg32 to yield a split 32,33 proinsulin intermediate, followed by CP-H trimming of the revealed arginine residues to yield des 31,32 proinsulin. PC2 can then cleave des 32,33 proinsulin at Lys64-Arg65, which together with CP-H trimming of the lysine and arginine residues, yields insulin and C-peptide (Figure 6.5). However, PC2 has a much stronger preference for the des 31,32 proinsulin substrate than proinsulin, whereas PC1/3 has an equivalent preference for proinsulin or des 64,65 proinsulin substrates [88,91]. As such, in humans, the sequential processing of proinsulin via des 31,32 proinsulin is the predominant route, where PC1/3 cleaves intact proinsulin first, followed by PC2 cleavage of des 31,32 proinsulin (Figure 6.5). This is consistent with the presence of the des 31,32 proinsulin conversion intermediates in the human circulation but negligible levels of des 64,65 proinsulin .
PC2, PC1/3 and CP-H are expressed in most neuroendocrine cells where they are involved in posttranslational processing of other prohormone precursors . The role of these proteolytic enzymes in proinsulin conversion has been substantiated in various gene-deletion studies. PC2, PC1/3 or CP-H deficiencies render multiple endocrine deficiencies. PC2 knockout mice have defective proinsulin processing with increased levels of the split proinsulin conversion intermediate des 31,32 proinsulin, consistent with PC2 preferentially cleaving at the Lys64, Arg65 site on proinsulin, and the preferred sequential proinsulin processing route  (Figure 6.3). Indeed, PC2 null mice have a polyendocrine phenotype the most obvious being fasting hypoglycemia and glucose intolerance due to a deficiency of circulating glucagon levels rather than increased insulin levels . A human mutation of both PC1/3 alleles exists, which results in negligible PC1/3 activity . This generates a complicated phenotype of multiple endocrine disorders due to general abnormal prohormone processing, one of which is very low insulin levels and high proinsulin levels, together with abnormal glucose homeostasis, consistent with defective proinsulin processing . A very similar phenotype is found in the PC1/3 null transgenic mouse model . Finally, the obese Fat/Fat mice have been found to have a mutation in the CP-H gene resulting in negligible CP-H activity . These CP-H null animals are hyperproinsulinemic, suggesting that CP-H trimming off of basic amino acids after PC2 and PC1/3 endopeptidic cleavage accelerates proinsulin proteolytic maturation through to insulin . Moreover di-arginyl insulin (that has >50% reduced biological activity) rather than insulin is produced indicating the role that CP-H plays in trimming basic amino acids during the proinsulin conversion process .