Metabolic engineering of Escherichia coli BL21 for biosynthesis of heparosan, a bioengineered heparin precursor Chunyu Zhang a , Long Liu a,n , Liping Teng b , Jinghua Chen b , Jian Liu c , Jianghua Li a , Guocheng Du a , Jian Chen d,nn a Key Laboratory of Carbohydrate Chemistry and Biotechnology, Ministry of Education, Jiangnan University, Wuxi 214122, China b School of Medicine and Pharmaceutics, Jiangnan University, Wuxi 214122, China c Division of Chemical Biology and Medical Chemistry, Eshelman School of Pharmacy, University of North Carolina, NC 27599, USA d National Engineering Laboratory for Cereal Fermentation and Technology, Jiangnan University, Wuxi 214122, China article info Article history: Received 27 April 2012 Received in revised form 6 June 2012 Accepted 28 June 2012 Available online 8 July 2012 Keywords: Bioengineered heparin Heparosan Escherichia coli K5 Metabolic engineering Molecular weight Chemical structure abstract As a precursor of bioengineered heparin, heparosan is currently produced from Escherichia coli K5, which is pathogenic bacteria potentially causing urinary tract infection. Thus, it would be advantageous to develop an alternative source of heparosan from a non-pathogeneic strain. In this work we reported the biosynthesis of heparosan via the metabolic engineering of non-pathogenic E. coli BL21 as a production host. Four genes, KfiA, KfiB, KfiC and KfiD, encoding enzymes for the biosynthesis of heparosan in E. coli K5, were cloned into inducible plasmids pETDuet-1 and pRSFDuet-1 and further transformed into E. coli BL21, yielding six recombinant strains as follows: sA, sC, sAC, sABC, sACD and sABCD. The single expression of KfiA (sA) or KfiC (sC) in E. coli BL21 did not produce heparosan, while the co-expression of KfiA and KfiC (sAC) could produce 63 mg/L heparosan in shake flask. The strain sABC and sACD could produce 100 and 120 mg/L heparosan, respectively, indicating that the expression of KfiB or KfiD was beneficial for heparosan production. The strain sABCD could produce 334 mg/L heparosan in shake flask and 652 mg/L heparosan in 3-L batch bioreactor. The heparosan yield was further increased to 1.88 g/L in a dissolved oxygen-stat fed-batch culture in 3-L bioreactor. As revealed by the nuclear magnetic resonance analysis, the chemical structure of heparosan from recombinant E. coli BL21 and E. coli K5 was identical. The weight average molecular weight of heparosan from E. coli K5, sAC, sABC, sACD, and sABCD was 51.67, 39.63, 91.47, 64.51, and 118.30 kDa, respectively. This work provides a viable process for the production of heparosan as a precursor of bioengineered heparin from a safer bacteria strain. & 2012 Elsevier Inc. All rights reserved. 1. Introduction Heparin and heparan sulfate (HS) are important glycosamino- glycans that involves a large number of biological processes such as blood coagulation, virus infection, cell differentiation, tumor metastasis and angiogenesis (Lidholt et al., 1988; Folkman et al., 1989; Lin et al., 2002; Marino et al., 2002; Sasisekharan et al., 2002; Tiwari et al., 2004; Tyrell et al., 1995). Heparin has been discovered as a drug to prevent blood coagulation since 1916 and has become the most popular anticoagulant (Baik et al., 2012; Bhaskar et al., 2012; Linhardt, 1991; Liu et al., 2009). Heparin is currently extracted from animal tissues such as porcine intestine and bovine lung. The heparin supply chain was reportedly contaminated by over sulfated chondroitin sulfate, causing nearly 100 deaths alone in the USA in 2008 (Kishimoto et al., 2008; Laurencin and Nair, 2008). The accident raised the concerns over the vulnerability of animal sourced heparin. The US FDA then started inspection of foreign suppliers and upgraded the pharmacopeial monographs to reduce the likelihood of similar crisis (Linhardt and Liu, 2012). However, these efforts may lead to an insufficient supply of the critical drug as the animal sources for heparin preparation were very limited. With the increasing demand of the drug, the cost of heparin active pharmaceutical ingredient has increased 10-fold (Linhardt and Liu, 2012). Many efforts have been made to synthesize heparin to overcome the side effects and insufficient supply of heparin. Among them, chemical synthesis and chemo-enzymatic synthesis are representa- tive. Chemical synthesis is complicated and merely amenable to the synthesis of oligosaccharides less than hexasaccharide. An antic- oagulant pentasaccharide, namely fondaparinux, was chemically synthesized by more than 60 steps, with yield as low as 0.5% (Liu and Liu, 2010). Although it was marketed and had good pharmaco- kinetic/pharmacodynamics properties, fondaparinux is unable to Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/ymben Metabolic Engineering 1096-7176/$ - see front matter & 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ymben.2012.06.005 n Corresponding author. Fax: þ86 510 85918309. nn Corresponding author. Fax: þ86 510 85910799. E-mail addresses: longliu@jiangnan.edu.cn (L. Liu), jchen@jiangnan.edu.cn (J. Chen). Metabolic Engineering 14 (2012) 521–527