WO2005042753A1

WO2005042753A1 - Production of human glycosylated proteins in transgenic insects

Info

Publication number: WO2005042753A1
Application number: PCT/US2004/035553
Authority: WO
Inventors: Donald Jarvis; Nikolai Van Beek; Malcolm Fraser
Original assignee: Chesapeake Perl, Inc.
Priority date: 2003-10-28
Filing date: 2004-10-28
Publication date: 2005-05-12
Also published as: US20070067855A1

Abstract

This invention relates, e.g., to transgenic insects, or progeny thereof, whose cells contain at least one genomically integrated, expressible, nucleic acid encoding two or more of a set of Nglycosylation enzymes that can glycosylate a heterologous protein with a mammalianized (e.g., humanized) glycosylation pattern. The glycosylation genes are preferably expressed in the insect cells in catalytic amounts. Also described are methods to use such a transgenic insect to produce heterologous, mammalianized polypeptides of interest.

Description

PRODUCTION OF HUMAN GLYCOSYLATED PROTEINS TN TRANSGENIC INSECTS

FIELD OF THE INVENTION This invention relates, e.g., to N-glycosylation of proteins in insects, and provides methods, vectors, and transgenic insects.

BACKGROUND INFORMATION The biotechnology revolution has created vast new potential for pharmaceuticals, yet that potential remains unrealized due largely to problems in manufacturing. Biopharmaceuticals, which have greatly expanded targets for therapeutic intervention, now represent about 30% of the drugs in the development pipeline. However, the biopharmaceutical industry does not have the manufacturing infrastructure required to meet patient needs; in other words, discovery has far outpaced production. A series of difficulties that cascade throughout the drug development cycle — process changes, scale-up problems, and capacity shortages, all of which cause repeated clinical trials — exhaust developers' money before drugs can be approved for use. Methods have been developed for producing biopharmaceuticals, particularly recombinant proteins such as enzymes and antibodies, in a variety of hosts, including bacteria, yeast, mammalian cell culture, and transgenic mammals and plants. However, each of these systems suffers from shortcomings. Bacterial fermentation is unable to modify proteins. Mammalian cell culture cannot easily be scaled up. Transgenic mammals are expensive and time consuming to produce and raise problems of public acceptance. To be fully functional, most proteins require "post-translational modification," or further changes to overall structure and composition. The most common change involves a process called glycosylation, an enzyme-mediated addition of specific sugars to the protein backbone. Glycosylation is important for protein use in humans, as it can affect the efficacy, stability and often safety of a potential drug. The best known biotherapeutics are treatments for diabetes, sclerosis, Hodgkin's lymphoma, Crohn's disease, and various promising therapies for AIDS and cancer. Seven of the current top ten biopharmaceuticals (Procrit, Epogen, Intron A/Rebetron, Neupogen, Humulin, Avonex, Rituxan, Enbrel, Remicade, and Cerezyme) require glycosylation. It would be desirable to produce recombinant proteins that have proper mammalian (e.g., human) glycosylation patterns, in insect cells. Such a process could provide the industry a flexible, low-capital-intensive, fast-turnaround, linearly scalable process for manufacturing authentic human-type glycoproteins for, e.g., therapeutic applications.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 shows protein N-glycosylation pathways. (Jarvis et al. (1998) Current Opinion in Biotechnology 9, 528-533 and Jarvis, D.L. (2003) Virology 310, 1-7.)

Figure 2 shows N-glycosylation pathways by which GlcNAc-transferase I to VI incorporate GlcNAc residues into a Man(αl-6)[Man(αl-3)] Manβ-RN-glycan core. (Montreuil et al. (eds.), Glycoproteins, Vol. 29a. Elsevier, Amsterdam, 1995)

Figure 3 shows a typical piggyBac vector. The sizes of the promoters, enzyme pairs, piggyBac and GFP marker are as follows:

Promoter sizes: piggyBac size:

(2X) iel promoter 2.4Kb, hr5 fragment 0.5Kbp, total 2.9Kb 5' TR 0.1Kb, 3'TR 0.3Kb, total 0.4Kb

(2X) hsp70 0.94Kb, hrS fragment 0.5Kb, total 1.44Kb (2X) CMV 0.13Kb, (7X) TetO 0.3Kb, total

Enzyme pair size: GFP marker gene size: 3XP3/GFP gene 1.29Kb

2.6 Kb human GlcNAc-TI, 1.34 Kb human GlcNAc-TII, total

3.94Kb

1.65 Kb rat alpha 2,6-sialyltransferase, 1.00 Kb mouse alpha 2,3-sialyltransferase, total 2.6Kb

1.3 Kb mouse SAS, 1.7 Kb mouse CMP-SAS, total 3Kb

Largest size for an individual piggyBac transposon construct will be 8.13Kb, well within the limits of demonstrated mobility

Figure 4 shows three constructs. Figure 4A shows pDIEl-GnTII/GalT-DsRedl-TOPO.4; Figure 4B shows pDIEl-ST6.1/ST3.4-ECFP-TOPO.4; Figure 4C shows pDIE-SAS/CMP.SAS-EYFP- TOPO.4. Abbreviations: DIE1, dual immediate early 1; GnTII, N-acetylglucosaminyltransferase II; GalT, β4-galactosyltransferase, ST6.1, alpha 2,6-sialyltransferase; ST3.4, alpha 2,3-sialyltransferase; ECFP, enhanced cyano fluorescent protein; SAS, sialic acid synthase; CMP.SAS, CMP-sialic acid synthetase; EYFP, enhanced yellow fluorescent protein. DESCRIPTION OF THE INVENTION This invention relates, e.g., to insects (such as insect larvae) which contain, in at least some of their cells, expressible nucleic acid sequences encoding one or more (e.g., two or more) of a set of glycosylation enzymes noted below, such that expression of the glycosylation enzyme(s) allows for the production of partially or completely mammalianized (e.g., humanized) glycosylation of a polypeptide of interest that is introduced into, or that is present endogenously in, the insect. The introduced polypeptide is generally a recombinant polypeptide (which may comprise coding sequences that are endogenous to, or heterologous to, the insect). Preferably, the recombinant polypeptide of interest is heterologous to the insect. In some embodiments, the glycosylation enzymes are produced in catalytic amounts. That is, the expression of the glycosylation enzyme(s) is effective and sufficient to glycosylate, in the insect, a polypeptide of interest (e.g., a heterologous polypeptide) in a mammalianized glycosylation pattern, yet is not so great that it significantly inhibits viability of the insect, or compromises the ability of the insect to produce high yield of the mammalianized polypeptide of interest. In other embodiments, one or more of the glycosylation enzymes are produced in greater amounts (e.g., at the same level as a heterologous polypeptide that is to be glycosylated). An "effective amount" of a glycosylation protein is an amount that results in partial or completely mammalianized glycosylation of a heterologous polypeptide that is introduced into, or is endogenously present in, the insect. In some embodiments, the glycosylation enzymes are produced in a coordinate fashion. The expressible nucleic acid sequences can be stably integrated into the somatic and germ line cells of the insect (in a transgenic insect); or they can be integrated in the somatic cells (e.g., following introduction into the insect with, for example, a suitable transposon-based vector or retrovirus vector); or they can be transiently produced (e.g., following introduction into the insect with, for example, a baculovirus-based vector). The invention also relates to methods using an insect as above for producing a polypeptide of interest, such as a heterologous polypeptide, such that the polypeptide of interest exhibits a partially or completely mammalianized glycosylation pattern. For example, an expressible nucleic acid encoding the polypeptide of interest can be introduced an insect which is transgenic for the mentioned glycosylation enzyme(s) (e.g., the expressible nucleic acid is fed to the transgenic insect) in e.g., either a baculovirus-based vector, a transposon-based vector, or a retrovirus vector, such that the introduced nucleic acid becomes either transiently or stably introduced into a somatic cell of the insect, and the protein of interest is expressed and glycosylated in that somatic cell. Alternatively, a multiply transgenic insect can be generated, in which expressible nucleic acid encoding the polypeptide of interest and expressible nucleic acid encoding the glycosylation enzyme(s) are both stably integrated in the somatic and germ line cells of the insect. The polypeptide of interest can then be produced and glycosylated in the multiply transgenic insect cells. In another embodiment, a nucleic acid comprising expressible nucleic acid sequences encoding the glycosylation enzyme(s) and a nucleic acid comprising expressible nucleic acid sequences encoding the polypeptide of interest are co-introduced (either on the same vector or on different vectors) into somatic cells of a non-transgenic insect. The vector may be, e.g., a baculovirus-based vector, a transposon-based vector, or a retrovirus vector. The polypeptide of interest is then produced and glycosylated in somatic cells that contain both nucleic acids. One embodiment of the invention is an insect comprising in at least some of its cells at least two of the glycosylation enzymes noted below (e.g., in catalytic amounts) and a heterologous polypeptide of interest, wherein the heterologous polypeptide is glycosylated by the glycosylation enzymes in a mammalian (e.g., human) glycosylation pattern. Advantages of the insects and methods of the invention include that the insects are simple and economical to cultivate (for example, insects have fewer requirements for special growth conditions than do cells in culture, and can be cultivated at low cost, in a controlled environment); high yields of the glycosylated polypeptide can be produced rapidly, for large scale production; polypeptides produced in insect cells by the methods of the invention are unlikely to be contaminated by mammalian viruses or prions; insect cultures (e.g., larval cultures) can be grown under space-efficient conditions and can be synchronized to reach the same level of maturity at the same time; and one can control toxicity to the insect, thereby achieving high survivability, in spite of the complexities of heterogeneity of cells in the insect, a complex physiological environment, and the variety of life phases during insect development. Each larva (caterpillar) is effectively a self-contained mini-bioreactor consisting of millions of host cells. Mass rearing, infecting, and harvesting proteins from these larval bioreactors allows one to capitalize on the low cost and great scalability of the insect as a protein production system. In some embodiments, expression of the glycosylation enzyme(s) is regulatable (e.g., inducible). The ability to avoid constitutive production of glycosylation enzymes, which might be toxic to the insect, or might reduce the yield of a glycosylated protein of interest, is an advantage of this embodiment of the invention. Glycosylation enzymes involved in the present invention include the following: N-glycoproteins are one subclass of eukaryotic glycoproteins that are particularly important in biotechnology. Many pharmaceutically relevant products, such as immunoglobulins, cytokines, blood clotting factors, and anticoagulants are N-glycosylated. The glycans on these molecules play important roles in their functions and influence their therapeutic potential. For example, terminal sialic acids influence the pharmacokinetics of N-glycoproteins because nonsialylated N-glycoproteins are rapidly cleared from the circulatory system. The mammalian N-glycosylation pathway. Important enzymatic functions involved in the mammalian protein N-glycosylation are well defined (see, e.g., Kornfeld et al. (1985) Ann. Rev. Biochem. 54, 631-664; Montreuil et al. (1995) "Glycoproteins". New Comprehensive Biochemistry (A. Neuberger, and L. L. M. Van Deenen, Eds.), 29a Elsevier, Amsterdam; Varki et al. (1999). "Essentials of Glycobiology." Cold Spring Harbor Press, Cold Spring Harbor, New York). The products of this processing pathway are termed "N-glycoproteins" because their carbohydrate side chains are linked to the polypeptide backbone by an N-glycosidic bond to the asparagine residue. This pathway begins with the transfer of a pre-assembled glycan, Glc₃Man₉GlcNAc₂, from a lipid carrier to an asparagine residue within a specific recognition site in a nascent polypeptide (see Fig. 1, Step 1). Standard monosaccharide abbreviations used in this application include: Glc (glucose), Man (mannose), GlcNAc (N-acetylglucosamine), Gal (galactose), GalNAc (N-acetylgalactosamine), Fuc (fucose), Sia (sialic acid), ManNAc (N- acetylmannosamine). Transfer occurs as the nascent polypeptide enters the lumen of the rough endoplasmic reticulum (RER) and is followed by trimming of the glucose residues (step 2) to produce Man₉GlcNAc₂, which is generally termed a "high-mannose" N-glycan. In some cases, there is no further processing and the high mannose N-glycan is the end product. In other cases, the high mannose glycan serves as an intermediate that is further processed by a sequential series of enzymatic reactions catalyzed by glycosidases and glycosyltransferases localized along the secretory pathway. Four of the nine mannose residues are trimmed by class I alpha-mannosidases (Man I's) in the ER and Golgi apparatus (step 3), yielding Man₅GlcNAc₂. One GlcNAc residue is then added by N-acetylglucosaminyltransferase I (GlcNAc-TI; step 4), which permits alpha-mannosidase II (Man II; step 5) to remove two more mannose residues. This leads to elongation of the trimmed structures and the production of "complex" N-glycans by various Golgi glycosyltransferases, including N-acetylglucosaminyltransferases (GlcNAc-Ts), fucosyltransferases (Fuc-Ts), galactosyltransferases (Gal-Ts), N-acetylgalactosaminyltransferases (GalNAc-T's), and sialyltransferases (Sial-Ts), as shown in steps 5-7. The complex N-glycans shown on the bottom right of Fig. 1 are common "biantennary" structures. Mammalian cells also can produce more highly branched complex N-glycans with up to five antennae. In addition to the glycosyltransferases shown in Fig. 1, N-glycan elongation requires various nucleotide sugars, including UDP-GlcNAc, UDP-Gal, and CMP-sialic acid. These compounds are the donor substrates for the glycosyltransferases catalyzing the elongation reactions. The nucleotide sugars are synthesized in the cytoplasm or nucleus of the cell and are imported into the lumen of the Golgi apparatus, where the elongation reactions occur, by specific nucleotide sugar transporters. The insect N-glycosylation pathway. The initial steps in the insect N-glycosylation pathway are identical to those in the mammalian pathway, producing the common intermediate, GlcNAcMan₃GlcNAc (+/-Fuc). While mammalian cells have sufficient levels of glycosyltransferases to elongate this common intermediate and produce complex N-glycans, insect cells generally appear to have low or undetectable levels of these activities and no detectable CMP-sialic acid. In addition, some insect cells have a processing N-acetylglucosaminidase (GlcNAcase) that trims this intermediate to produce simple "paucimannose" N-glycans. Accordingly, the major processed N-glycans found on recombinant glycoproteins produced by baculovirus infected insect cell lines or larvae are usually paucimannose structures (Fig. 1). This conclusion is supported by data from, e.g., structural studies on the N-glycans isolated from insect or insect cell-derived glycoproteins, the use of specific N-glycan processing inhibitors, enzyme activity assays, analyses of endogenous nucleotide sugar levels, and the isolation and characterization of insect genes encoding various N-glycan processing enzymes. Baculovirus-expressed recombinant glycoproteins almost never have terminally sialylated N-glycans. The inability to routinely produce complex, terminally sialylated N-glycans is a major technical barrier associated with the use of the baculovirus expression system for recombinant glycoprotein production, at least because baculovirus produced unsialylated glycoproteins have very short half-lives in vivo. The present inventors have created transgenic lepidopteran insect larvae that can support the production of humanized recombinant glycoproteins by baculovirus expression vectors. The inventive larvae express levels of relevant enzymes that are effective to produce complex, terminally sialylated N-glycans in high quantity and consistent quality.

In one aspect, this invention relates to a transgenic insect, or progeny thereof, whose somatic and germ cells contain recombinant nucleic acid: A. two or more of the glycosylation enzymes: a beta-l,2-N- acetylglucosaminyltransferase (e.g., beta-l,2-N-acetylglucosaminyltransferase I and/or beta- 1,2- N-acetylglucosaminyltransferase II); a βl, 4-galactosyltransferase (e.g., beta 4- galactosyltransferase I); and/or a sialyltransferase [e.g., one of the many suitable alpha 2,6- sialyltransferases and/or one of the many suitable alpha 2,3-sialyltransferases (such as alpha 2,3- sialyltransferase III and/or alpha 2,3-sialyltransferase IV)]; or B. one or more of the glycosylation enzymes: a beta-l,2-N-acetylglucosaminyltransferase (e.g. , beta- 1 ,2-N-acetylglucosaminyltransferase I and/or beta- 1 ,2-N- acetylglucosaminyltransferase II); and/or a sialyltransferase [e.g., one of the many suitable alpha 2,6-sialyltransferases and/or one of the many suitable alpha 2,3-sialyltransferases (such as alpha 2,3-sialyltransferase III and/or alpha 2,3-sialyltransferase IV)], wherein each recombinant nucleic acid encoding a glycosylation enzyme is integrated in the insect genome, and is present in one or more copies, wherein each recombinant nucleic acid encoding a glycosylation enzyme is operably linked to an expression control sequence, and wherein expression of said glycosylation enzyme(s) (e.g., in a catalytic amount) allows for production of a partially or completely mammalianized glycosylated protein in the insect. In one embodiment, the somatic and germ cells contain recombinant nucleic acid encoding: A. two or more of the glycosylation enzymes: a) beta-l,2-N-acetylglucosaminyltransferase I, b) beta-l,2-N-acetylglucosaminyltransferase II, c) a βl, 4-galactosyltransferase (e.g., beta 4-galactosyltransferase I), and/or d) a sialyltransferase [e.g., an alpha 2,6-sialyltransferase and/or an alpha 2,3- sialyltransferase (such as alpha 2,3-sialyltransferase III and/or alpha 2,3-sialyltransferase IV)], or B. one or more of the glycosylation enzymes: a) beta-l,2-N-acetylglucosaminyltransferase I, b) beta- 1 ,2-N-acetylglucosaminyltransferase II, and/or d) a sialyltransferase [e.g., an alpha 2,6-sialyltransferase and/or an alpha 2,3- sialyltransferase (such as alpha 2,3-sialyltransferase III and/or alpha 2,3-sialyltransferase IV)]- In another embodiment, the somatic and germ cells contain recombinant nucleic acid encoding: A. two or more of the glycosylation enzymes: b) beta-l,2-N-acetylglucosaminyltransferase II, c) a βl, 4-galactosyltransferase (e.g., beta4-galactosyltransferase I), d-1) an alpha 2,6-sialyltransferase, and/or d-2) an alpha 2,3-sialyltransferase (such as alpha 2,3-sialyltransferase III and/or alpha 2,3-sialyltransferase IV)], or B. one or more of the glycosylation enzymes: b) beta-l,2-N-acetylglucosaminyltransferase II, d-1) an alpha 2,6-sialyltransferase, and/or d-2) an alpha 2,3-sialyltransferase (such as alpha 2,3-sialyltransferase III and/or alpha 2,3-sialyltransferase IV). The expression control sequences to which each recombinant nucleic acid encoding a glycosylation enzyme is operably linked may be the same or different. In all of the embodiments discussed herein in which expression control sequences regulate the expression of more than one nucleic acid sequence, the expression control sequences may be the same or different. The integrated copies may be tandemly integrated, integrated into different regions of the same chromosome, or integrated into different chromosomes. As used herein, the term "recombinant" nucleic acid refers to a nucleic acid that encodes a polypeptide which is heterologous to the insect, and or a nucleic acid which has been genetically engineered (e.g., cloned into a vector) before being introduced into the insect. Thus, a nucleic acid encoding a protein originating from a particular type of insect (endogenous to that type of insect), but engineered so as to be produced at increased levels, and then introduced back into that type of insect, is considered to be recombinant. As used herein, the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise. For example, "an" alpha sialyltransferase, as used above, means one or more alpha sialyltransferases, which can encompass two different types of alpha sialyltransferase, such as an alpha 2,6-sialyltransferase and an alpha 2,3-sialyltransferase. A coding sequence that is operably linked to an expression control sequence is sometimes referred to herein as an "expressible" nucleic acid sequence. In embodiments of the invention, the somatic and germ cells of the transgenic insect comprise genomically integrated recombinant nucleic acid encoding enzyme a); enzyme a) and enzyme b); enzyme a), enzyme b) and enzyme c); or, preferably, enzyme a), enzyme b), enzyme c) and enzyme d). When more than one of these glycosylation enzymes are present in the transgenic insect, they may be integrated into different regions of the same chromosome, or integrated into different chromosomes. In one embodiment, if nucleic acid encoding enzyme c) is present, nucleic acid encoding at least one of enzymes a), b) or d) is also present. Insect cells generally do not comprise enzymes a) through d) above, or comprise such low amounts of these enzymes that little if any enzymatic activity is detectable. Therefore, N- glycosylated glycoproteins that are produced in insect cells generally exhibit structures similar to the "paucimannose" structure shown in Figure 1. By contrast, N-glycosylated glycoproteins that are produced in mammalian cells exhibit structures similar to the "complex" structures shown in Figure 1. These complex structures are generated by the sequential action of proteins a) through c) above, followed by the action of enzyme(s) d), which introduce sialic acid moieties onto the termini of the arms of the biantennary carbohydrate side chains. For example, an alpha 2,6- sialyltransferase can sialylate the lower (alpha-3) branch of a biantennary glycan; an alpha 2,3- sialyltransferase can sialylate the upper (alpha-6) branch and/or lower (alpha-3) branch of a biantennary glycan; and various other combinations can occur. Either partially or fully sialylated structures are suitable for various uses. Sialic acid residues also may be alpha 3- or alpha 6- linked to additional branches, if those branches are produced by the actions of N- acetylglucosaminyltransferases IV, V, and VI. A polypeptide that is acted upon by, for example, enzyme a), is referred to herein as a partially mammalianized (e.g., humanized) glycopolypeptide. It differs from most naturally produced polypeptides in the insect by virtue of the presence of the carbohydrate residue provided by enzyme a). Similarly, any polypeptide glycosylated by fewer than the full set of enzymes a) through d) above is also referred to herein as a "partially mammalianized (e.g., humanized)" glycopolypeptide. A glycopolypeptide that exhibits a "complex" glycoprotein structure (e.g., a mammalian (preferably, human) glycan profile) is said to be "completely mammalianized (humanized)", or to exhibit a glycosylation pattern characteristic of mammals (e.g., humans). Partially and completely mammalianized glycosylation structures are found in many types of mammalian cells, such as bovine or human cells. The term, a "mammalianized" glycopolypeptide, as used herein, refers to a glycopolypeptide that exhibits a glycan profile characteristic of a mammalian glycoprotein, as discussed above. A "mammalianized" glycopolypeptide, as used herein, encompasses both partially and completely mammalianized glycopolypeptides. The terms "mammalianized glycopolypeptide," "mammalianized glycoprotein," "mammalianized polypeptide" and "mammalianized protein" are sometimes used interchangeably herein. Partially or completely mammalianized polypeptides exhibit a number of advantages compared to polypeptides produced by an insect that lacks the glycosylation enzymes of the invention. These advantages include, e.g., enhanced stability when introduced into a mammal, altered activities, or the like. An insect that expresses fewer than a full set of enzymes a) through d) has a variety of utilities, which will be evident to the skilled worker. For example, such an insect can be used to generate a protein of interest that exhibits a partially mammalianized glycosylation pattern, and that consequently exhibits improved properties compared to a polypeptide produced by an insect that is not so modified. If an insect naturally produces small amounts of, for example, one or more enzymes which lie upstream in the glycosylation pathway, expression of an enzyme that lies further downstream in the pathway can cap and stabilize the glycosylation product resulting from the small amounts of the upstream enzyme(s). Therefore, an insect that naturally makes one or more of the upstream enzymes may be transgenically modified to express one or more recombinant downstream enzymes, provided that the transgenic insect produces sufficient amounts of a sialylization enzyme to produce a sialic acid cap. Another embodiment of the invention is a transgenic insect as above whose somatic and germ cells further comprise recombinant nucleic acid encoding one or more of the following glycosylation enzymes: e) a sialic acid synthase and/or f) CMP-sialic acid synthetase, wherein each recombinant nucleic acid encoding a glycosylation enzyme is genomically integrated in the insect genome, and is present in one or more copies, and wherein each recombinant nucleic acid encoding a glycosylation enzyme is operably linked to an expression control sequence. Preferably, both e) and f) are present. Some insects may generate sufficient sialic acid, themselves, to sialylate heterologous proteins in the methods of the invention. However, many insects lack such an endogenous source of sialic acid, or produce insufficient quantities. Therefore, for those insects, the needed sialic acid can be introduced into the insects with their diet. Alternatively, and preferably, the sialic acid can be provided by introducing into the cells of the insects enzymes e) and/or f), preferably both e) and f). For example, nucleic acids expressing the enzymes can be integrated into the cells of the insect. These two enzymes together, when presented with the substrate ManNAc (N-acetylmannosamine) will generate the needed CMP-sialic acid. The ManNAc can be presented to the insect by conventional means, e.g., orally, in its diet. In a preferred embodiment, a transgenic insect of the invention expresses in its somatic and germ line cells all of enzymes a) through f). Optionally, the somatic and germ cells of any of the transgenic insects described above further comprise recombinant nucleic acid encoding one or more of the following auxiliary glycosylation proteins: g) UDP-N-acetylglucosamine 2 epimerase/N-acetylmannosamine kinase; h) beta- 1,4-N-acetylglucosaminyltransferase III; i) beta-l,4-N-acetylglucosaminyltransferase IV; j) beta-l,6-N-acetylglucosaminyltransferase V; k) beta-l,4-N-acetylglucosaminyltransferase VI; 1) a beta 1,4-N-acetylgalactosaminyltransferase; m) CMP-sialic acid transporter; n) UDP-galactose transporter, wherein each recombinant nucleic acid encoding an auxiliary glycosylation protein is genomically integrated in the insect genome, in one or more copies, and wherein each recombinant nucleic acid is operably linked to an expression control sequence. Enzyme g) converts N-acetylglucosamine to N-acetylmannosamine-phosphate, which allows one to feed larvae N-acetylglucosamine, rather than N-acetylmannosamine, to support sialoglycoprotein biosynthesis. N-acetylglucosamine is considerably less expensive than N- acetylmannosamine. Enzymes h) through k) allow insect cells to produce tri, tetra, or pentaantennary N- glycans. See Figure 2 for a diagram of the reactions carried out by some of these enzymes. Enzyme h) adds "bisecting" GlcNAc in β 1,4 linkage to the core. Enzyme i) adds GlcNAc in β 1,4 linkage to the alpha 3 branch mannose. Enzyme j) adds GlcNAc in β 1,6 linkage to the alpha 6 branch mannose. Enzyme k) adds GlcNAc in β 1,4 linkage to the alpha 6 branch mannose. Enzyme 1) transfers N-acetylgalactosamine in beta 1,4 linkage to terminal N- acetylglucosamine residues in N-glycans. It can serve as an alternative to βl,4- galactosyltransferase, transferring GalNAc, instead of Gal to outer chain positions of some N- glycoproteins. Protein m) transports CMP-sialic acid into Golgi apparatus. (Although it was unexpected that insect cells would have this transporter, cell culture studies performed by the present inventors indicate that insect cells can somehow move CMP-sialic acid into Golgi, even in the absence of added transporting enzyme. Added CMP-sialic acid transporter can enhance this transport.) Protein n) transports UDP-galactose into Golgi apparatus. (Some insect cells express low levels of this transporter. Engineering insect cells to express a mammalian UDP-galactose transporter can improve the efficiency of the transport.) These auxiliary enzymes are listed above in the approximate order of preference. The nucleic acids encoding glycosylation enzymes that are expressed in the insects of the invention can be obtained from any suitable source, examples of which will be evident to skilled workers. For example, the enzyme can be one that is naturally produced in the insect, but at ineffectively low levels. An insect of the invention can be designed to produce increased amounts of the enzyme, which are effective for producing a partially or completely mammalianized glycosylation pattern in a polypeptide of interest. In another embodiment, the glycosylation enzyme is obtained from an insect of a different insect species. In another embodiment, the glycosylation enzyme is obtained from an invertebrate other than an insect (e.g. C. elegans) or from a vertebrate (such as a chicken or a mammal). Suitable mammalian sources include, e.g., mouse, rat, cow or human. Enzymes obtained from different sources can be used in conjunction with one another. Methods for cloning and expressing such enzymes are conventional. A sequence "obtained" from a particular source does not necessarily encode a polypeptide sequence identical to that of the wild type enzyme from that source. Any glycosylation enzyme that retains the enzymatic function of the wild type enzyme, including naturally occurring allelic variants or mutations that are introduced artificially into the protein, can be used. Enzymatically active fragments of the enzyme can also be used. As used herein, the term "insect" includes any stage of development of an insect, including a one-celled germ line cell, a fertilized egg, an early embryo, a larva, including any of a first through a fifth instar larva, a pupa, or an adult insect. For the production of mammalianized polypeptides of interest, a large larva, such as a fourth or fifth instar larva is preferred. It will be evident to a skilled worker which insect stage is suitable for a particular purpose, such as for direct production of a glycosylated polypeptide of interest, for storage or transport of an insect to a different location, for generation of progeny, for further genetic crosses, or the like. Any of a variety of insects are suitable. Among suitable insects are, e.g., Lepidoptera (e.g., Bombyx mori, Manduca sexta, Hyalophora cecropia, Spodoptera exigua, Spodoptera frugiperda, Spodoptera litoralis, Spodoptera litura, Heliothis virescens, Helicoverpa zea, Helicoverpa armigera, Trichoplusia ni, Plutella xylostella, Anagrapha falcifera, Cydia pomonella, Cryptophlebia leucotreta, and Estigmene acred), and insect species from the orders Coleoptera, Hymenoptera, Orthoptera, and Diptera. Preferably, the insect is from the order Lepidoptera, most preferably Trichoplusia ni (T. ni). The term "expression control sequence," as used herein, refers to a polynucleotide sequence that regulates expression of a polypeptide coded for by a polynucleotide to which it is functionally ("operably") linked. Expression can be regulated at the level of the mRNA or polypeptide. Thus, the term expression control sequence includes mRNA-related elements and protein-related elements. Such elements include promoters, domains within promoters, upstream elements, enhancers, elements that confer tissue or cell specificity, response elements, ribosome binding sequences, transcriptional terminators, etc. An expression control sequence is "operably linked" to a nucleotide coding sequence when the expression control sequence is positioned in such a manner to effect or achieve expression of the coding sequence. For example, when a promoter is operably linked 5' to a coding sequence, expression of the coding sequence is driven by the promoter. Suitable expression control sequences that can function in insect cells will be evident to the skilled worker. In some embodiments, it is desirable that the expression control sequence comprises a constitutive promoter. Among the many suitable "strong" promoters which can be used are the baculovirus promoters for the plO, polyhedrin (polh), p 6.9, capsid, and cathepsin- like genes. Among the many "weak" promoters which are suitable are the baculovirus promoters for the iel, ie2, ieO, etl, 39K (aka pp31), and gp64 genes. Other suitable strong constitutive promoters include the B. mori actin gene promoter; Drosophila melanogaster hsp70, actin, α-1- tubulin or ubiquitin gene promoters; RSV or MMTV promoters; copia promoter; gypsy promoter; and the cytomegalovirus IE gene promoter. If it is desired to increase the amount of gene expression from a weak promoter, enhancer elements, such as the baculovirus enhancer element, hr5, may be used in conjunction with the promoter. In some embodiments, the expression control sequence comprises a tissue- or organ- specific promoter. Many such expression control sequences will be evident to the skilled worker. For example, suitable promoters that direct expression in insect silk glands include the Bombyx mori p25 promoter, which directs organ-specific expression in the posterior silk gland, and the silk fibroin Heavy chain gene promoter, which directs specific expression of genes in the median silk gland. Example XVI describes the generation and use of transgenic insects of the invention that express glycosylation enzymes specifically in their silk glands. In general, the glycosylating enzymes of the invention are required in catalytic amounts. Therefore, in one embodiment of the invention, much lower amounts of these enzymes are present than of the heterologous polypeptides of interest, which are generated in massive, large amounts, glycosylated, and harvested for further use. For example, a suitable molar ratio of heterologous protein produced to a glycosylating enzyme may be greater than about 100:1. Alternatively, the glycosylating enzymes may be in comparable (e.g., approximately stochiometric) amounts to the heterologous protein to be glycosylated. A skilled worker can readily select suitable promoters and/or conditions to express suitable amounts of the glycosylating enzymes (e.g., amounts which are sufficient to (effective to) glycosylate relatively high amounts of a protein of interest). Furthermore, a skilled worker can readily ensure that the glycosylation enzymes are present in sufficient local concentrations, and at an optimal time during insect propagation. In some embodiments of the invention, as is discussed in more detail elsewhere herein, it is desirable that an expression control sequence is regulatable (e.g., comprises an inducible promoter and/or enhancer element). Suitable regulatable promoters include, e.g., Drosophila or other hsp70 promoters, the Drosophila metallothionein promoter, an ecdysone-regulated promoter, the Saccharomyces cerevisciae Gal4/UAS system, and other well-known inducible systems. A Tet-regulatable molecular switch may be used in conjunction with any constitutive promoter, such as those described elsewhere herein (e.g., in conjunction with the CMV-IE promoter, or baculovirus promoters). Another type of inducible promoter is a baculovirus late or very late promoter that is only activated following infection by a baculovirus. Methods for designing and preparing constructs suitable for generating transgenic insects (or vectors for infection of an insect) are conventional. For these methods, as well as other molecular biology procedures related to the invention, see, e.g., Sambrook et al, Molecular Cloning: A Laboratory Manual, Second Edition, Cold Spring Harbor, N.Y., (1989); Wu et al, Methods in Gene Biotechnology (CRC Press, New York, NY, 1997), Recombinant Gene Expression Protocols, in Methods in Molecular Biology, Vol. 62, (Tuan, ed., Humana Press, Totowa, NJ, 1997); and Current Protocols in Molecular Biology, (Ausabel et al, Eds.,), John Wiley & Sons, NY (1994-1999). Some suitable methods are described elsewhere herein. A variety of immortalized lepidopteran insect cell lines are suitable for infection by the vectors/constructs of the invention. Among these are Sf9 (Vaughn et al. (1977) In Vitro 13, 213- 217) and Tn 5B1-4 (Hive Five^®; Wickham et al. (1992) Biotech. Progr. 8, 391-6). Methods for generating transgenic insects are conventional. For example, in one embodiment, one or more genes to be introduced are placed under the control of a suitable expression control sequence, and are cloned into a vector, such as a viral vector (e.g., an attenuated baculovirus vector, or a non-permissive viral vector that is not infective for the particular insect of interest). The sequences to be introduced into the insect are flanked by genomic sequences from the insect. The construct is then introduced into an insect egg (e.g., by microinjection), and the transgene(s) then integrate by homologous recombination of the flanking sequences into comparable sequences in the insect genome. One method according to the invention employs an approach adapted from the techniques presented in Yamao et al. (1999) Genes and Development 13, 511-516. In that publication, a non-permissive insect host (B. mori) was infected with a recombinant AcMNPV carrying a gene of interest flanked by sequences derived from the host genome. The virus delivered its DNA, but could not consummate its infection cycle. The viral DNA recombined with the host genome via an extremely low frequency homologous recombination event between the host sequences in the viral DNA and the same sequences in the B. mori genome. In another embodiment, the vector is a transposase-based vector. One form of such transposase-based vectors is a viral vector (such as those described above) that further comprises inverted terminal repeats of a suitable transposon, between which the transgene of interest is cloned. One or more genes of interest, under the control of a suitable expression control sequence(s), are cloned into the transposon-based vector. In some systems, the transposon based vector carries its own transposase. However, generally, the transposon based vector does not encode a suitable transposase. In this case, the vector is co-infected into an insect (e.g., an insect larva) with a helper virus or plasmid that provides a transposase. The recombinant vector (along with, generally, a helper) is introduced by conventional methods (such as microinjection) into an egg or early embryo; and the transgene(s) become integrated at a transposon site (such as sequences corresponding the inverted terminal repeat of the transposon) in the insect genome. Suitable types of transposon-based vectors will be evident to the skilled worker. These include, e.g., Minos, mariner, Hermes, sleeping beauty, and piggyBac. In a preferred embodiment, the vector is a "piggyBac" vector. A typical piggyBac vector is shown in Figure 3. The TTAA-specific, short repeat elements are a group of transposons (Class II mobile elements) that have similar structures and movement properties. A typical piggyBac vector (formerly IFP2) is the most extensively studied of these insertion elements. piggyBac is 2.4 kb long and terminates in 13 bp perfect inverted repeats, with additional internal 19 bp inverted repeats located asymmetrically with respect to the ends (Cary et al. (1989) Virology. 172, 156-69). A. piggyBac vector may encode a trans-acting transposase that facilitates its own movement; alternatively, these sequences can be deleted and this function can be supplied on a helper plasmid or virus. piggyBac has been deleted for non-essential genes, into which large inserts can be cloned. Inserts as large as about 15 kB can be cloned into certain piggyBac vectors. This allows, for example, for the insertion of about six or seven genes with their expression control sequences. Thus, a collection of glycosylation enzymes, marker proteins, or the like, can be introduced together via a single transposon vector, into a single site in an insect genome. Several piggyBac vectors have been developed for insect transgenesis. Two particularly useful constructs, defined as minimal constructs for the movement of piggyBac vectored sequences, were developed by analysis of deletion mutations both within and outside of the boundaries of the transposon (Li et al. (2001) Mol. Genet. Genomics. 266, 190-8). Using constructs such as these it is possible to increase the amount of genetic material mobilized by the piggyBac transposase by minimizing the size of the vector. The minimal requirements for movement include the 5' and 3' terminal repeat domains and attendant TTAA target sequences. Nearly all of the internal domain may be removed, although more recent data indicates that some of this region may be required for efficient translocation of the mobilized sequences into the genome of the insect. In addition, a minimum of 50 bases separating the TTAA target sites of the element is required for efficient mobilization (Li et al. (2001), supra). piggyBac can transpose in insect cells while carrying a marker gene, and movement of the piggyBac element can occur in cells from lepidopteran species distantly related to the species from which it originated. piggyBac has been shown to transform D. melanogaster, the Carribean fruit fly, Anastrepha suspena, the oriental fruit fly, Bactrocera dorsalis, Bombyx mori, Pectinophora glossypiella, Tribolium castellani, and several mosquito species. At least three lepidopteran species, P. gossypiella, T. ni and B. mori, have been successfully transformed by the piggyBac element. Generally, a helper virus or plasmid that expresses a transposase is co-infected with the transposon-based vector as above. Expression of the transposase is determined by the choice of promoter for the insect system being tested. Toward that end, the present inventors have constructed several promoter-driven helper constructs that are useful for lepidopteran transformation, including the Drosophila hsp70, baculovirus iel promoter, and Drosophila Actin 5C promoter. Of these helper constructs, the hsp70 promoted helper, is particularly useful and serves as the primary helper for the transgenesis experiments in the Examples. One method according to the invention employs an approach adapted from the techniques h presented in Yamao et al., Abstract for poster presentation at the 6 International Conference on the Molecular Biology and Genetics of the Lepidoptera, in Kolympari, Crete Greece, Aug. 25- 30, 2003. In this publication, a nonpermissive host, B. mori, was infected with two recombinant AcMNPVs. One encoded the piggyBac transposase under the control of Drosophila heat shock protein 70 promoter and the other encoded the gene of interest (the one to be inserted into the B. mori genome) under the control of the B. mori actin A3 promoter and flanked by the piggyBac inverted terminal repeats. The design was that the transposase expressed by one virus mobilized the DNA in-between the inverted terminal repeats in the other and integrated that DNA into the host genome. The presence of resident copies of the piggyBac transposon in certain populations of T. ni does not appear to interfere with transposition of the transposon. Furthermore, the inventors have isolated a strain of T. ni which lacks resident copies of the piggyBac transposon. T. ni embryos have been injected with piggyBac vectors, and transformants have been successfully recovered and characterized to confirm piggyBac mobilization into the genome. For further guidance on the use of baculovirus-based vectors, see, e.g., WO01/29204 and

USP 6,551,825 and USP 6,18,064. Other recent references that discuss piggyBac vectors and methods for generating transgenic insects using them include, e.g., Handler et al. (1998) Proc Natl Acad Sci 95, 7520-7525; Fraser, M. J (2001) The TTAA-specific family of transposable elements. In: Insect transgenesis: Methods and Applications. A. A. James and A.H.Handler, eds. CRC Press, Orlando, FL; Lobo et al. (1999) Mol. Gen. Genetics 261, 803-810; Grossman et al. (2000) Insect Biochem. Mol. Biol. 30, 909-914; Lobo et al. (2001) Mol Gen. Genom. 265, 66-71; Lorenzen et al. (2003) Insect Mol Biol.12,433-40; Hacker et al. (2003) Proc Natl Acad Sci U S A. 100, 7720-5; Sumitani et al. (2003) Insect Biochem Mol Biol. 33, 449-58; Horn et al. (2003) Genetics 163 647-61; and Tomita et al. (2003) Nat Biotechnol. 21, 52-6. Methods for introducing constructs into an embryo to generate a transgenic insect (e.g., by microinjection) are conventional. Survivorship is usually quite high (up to 75%) for microinjected embryos. In general, preblastoderm eggs are stuck with a fine glass capillary holding a solution of the plasmid DNA and/or the recombinant virus. GO larvae hatched from the virus-injected eggs are then screened for expression of the gene of interest. Breeding transgenic Gl's with normal insects results in Mendelian inheritance. The inventors have succeeded in generating transformants using a piggyBac transposon. See the Examples herein for a further discussion of such microinjection procedures. Once a transgene(s) is stably integrated into the genome of an insect egg or early embryo, conventional methods can be used to generate a transgenic insect, in which the transgene(s) is present in all of the insect somatic and germ cells. When a subset of the complete set of glycosylation enzymes is present in a transgenic insect, other transposon-based vectors, which express different subsets of the glycosylation genes, can be introduced sequentially into the insect genome, and transgenic insects can then be generated. In another embodiment, when different subsets of the complete set of glycosylation enzymes are present in two or more individual transgenic insects, these insects can be genetically crossed to produce a transgenic insect that expresses a larger subset, or a complete set, of the glycosylation enzyme genes. In some embodiments, the transgenic insects are heterozygous for the glycosylation enzyme genes. For example, when potentially toxic glycosylation enzymes are produced constitutively, it may be advantageous for the insects to be heterozygous, to limit the amount of the enzyme that is produced. In other embodiments, the insects are homozygous for the transgenes. Methods for producing homozygous transgenic insects (e.g., using suitable back- crosses) are conventional. Another embodiment of the invention is an isolated cell, or progeny thereof, derived from a transgenic insect of the invention. Suitable cells include isolated germ line cells, and cells that can be used for the in vitro production of a polypeptide exhibiting a partial or complete pattern of mammalian glycosylation. Methods for obtaining and propagating cells from a transgenic insect, and using them (e.g. to generate more insects, or to generate glycosylated proteins) are conventional. The transgenic insects discussed above can be used to produce polypeptides of interest that exhibit partial or complete patterns of mammalian glycosylation. For example, the insects can be used in methods for glycosylating polypeptides in a mammalian (human) glycosylation pattern. One embodiment of the invention is a method for producing, in an insect, a mammalianized (e.g., humanized) glycosylated form of a polypeptide of interest that is endogenous to the insect. The method comprises cultivating (culturing, rearing) a transgenic insect as discussed above (preferably in the form of a larva) under conditions effective to produce a mammalianized glycosylated form of said polypeptide of interest. Conditions for cultivating insects, such as insect larvae, are conventional. For example, insects expressing enzymes a), b), c), d), e) (a sialic acid synthase) and f) (CMP-sialic acid synthetase) are generally grown in the presence of the substrate (food), N-acetylmannosamine. If enzyme g) is also being produced by the insect, the substrate N-acetylglucosamine can be supplied, instead of N- acetylmannosamine. Another embodiment of the invention is a method for producing, in an insect (preferably an insect larva), a mammalianized (e.g., humanized) glycosylated recombinant polypeptide. In embodiments of the invention, the recombinant polypeptide is an endogenous insect protein or, preferably, it is a heterologous protein. In one embodiment, this method comprises introducing into a transgenic insect as above (preferably in the form of a larva) a construct comprising nucleic acid encoding said recombinant protein, operably linked to an expression control sequence. In a preferred embodiment, these sequences are cloned into a suitable viral vector (such as a baculovirus-based vector, entomopox-based vector, or others). The coding sequences may be operably linked to an expression control sequence from the virus, itself, or to another suitable expression control sequence. Suitable virus-based vectors include, e.g., baculovirus vectors (such as vectors based on Autographa californica NPV, Orgyia pseudotsugata NPV, Lymantria dispar NPV, Bombyx mori NPV, Rachoplusia ou NPV, Spodoptera exigua NPV, Heliothis zea NPV, Galleria mellonella NPV, Anagrapha falcifera nucleopolyhedrovirus (AfNPV), Trichoplusia ni singlenuclepolyhedrovirus (TnSNPV)); retroviral vectors; and viral vectors that comprise transposon recognition sequences (e.g., piggyBac vectors); etc. As discussed above, baculovirus-based vectors have been generated (or can be generated without undue experimentation) that allow the cloning of large numbers of inserts, at any of a variety of cloning sites in the viral vector. Thus, more than one heterologous polypeptide may be introduced together into a transgenic insect of the invention. The viral vector can be introduced into an insect (e.g., an insect larva) by conventional methods, such as by oral ingestion. In one embodiment, the baculovirus replicates until the host insect is killed. The insect lives long enough to produce large amounts of the glycosylated polypeptide of interest. In another embodiment, a baculovirus is used that is attenuated or non-permissive for the host. In this case, the host is not killed by replication of the baculovirus, itself (although the host may be damaged by the expression of the glycosylation enzymes and/or the heterologous protein of interest). In another embodiment, sequences encoding one or more recombinant proteins of interest, operably linked to an expression control sequence, are cloned into a suitable transposon- based vector (such as a piggyBac vector). Like the baculovirus vectors discussed above, transposon-based vectors can carry large inserts, so more than one heterologous polypeptide may be introduced together into a transgenic insect of the invention. Transposon-based vectors may on occasion insert into the DNA of somatic cells, and thus be stably expressed for relatively long periods of time. In another embodiment, sequences encoding one or more recombinant proteins of interest, operably linked to an expression control sequence, are cloned into a retrovirus vector, or any other suitable virus vector. Such a construct may insert into the DNA of somatic cells, and thus be stably expressed for relatively long periods of time. Any heterologous polypeptide of interest may be expressed (and glycosylated) in an insect of the invention. A "heterologous" polypeptide, as used herein, refers to a polypeptide that is not naturally produced by the insect. The polypeptide may be of any suitable size, ranging from a small peptide (e.g., a peptide that contains an epitope that could be useful as a vaccine, or for generating an antibody of interest) to a full-length protein. The terms peptide, polypeptide and protein are used interchangeably herein. Preferably, the polypeptides expressed in this system are glycosylated in their natural mammalian (e.g., human) host. Suitable polypeptides include, e.g., marker proteins and therapeutic proteins. Among the wide variety of heterologous proteins that can be produced are antibodies, cytokines, blood clotting factors, anticoagulants, viral antigens, enzymes, receptors, pharmaceuticals, vaccines (e.g., for viral or parasite infections), enzymes, hormones, viral insecticides, etc. More specifically, some representative examples of suitable heterologous proteins are human genes, including growth hormone (hGH), macrophage colony-stimulating factor (hM-CSF), beta-interferon (HuIFN-beta), alpha-interferon, interleukins, growth factors, including fibroblast growth factors, and CD4. Other suitable proteins include a surface polypeptide from a pathogen, such as a parasite or virus, which can be useful in a vaccine, e.g., a surface antigen of Plasmodium, a prolylendopeptidase from Flavobacterium, the fusion glycoprotein (F) from Newcastle disease virus (NDV), hepatitis B and C virus antigens, proteins from human T-cell leukemia virus type I, human papillomavirus type 6b E2 DNA binding gene product, influenza virus haemagglutinin, etc. Other suitable proteins include therapeutic proteins which are currently produced recombinantly by other methods, and sold commercially, including antibodies and antibody fusion proteins [e.g., Campath (BCLL); Enbrel-RA (TNF inhibitor); Remicade-RA (TNF inhibitor); ReoPro (angioplasty); Rituxan (NHL); Synagis (RSV); Zenapax (transplant rejection); Zevalin (NHL); Herceptin (breast cancer); Humira (RA); MRA (RA); anti IL6 receptor (MAB); Xolair (asthma); Amevive (psoriasis); Bexxar (NHL); Antegren (Crohn's disease)]; lysosomal storage proteins [e.g., Cerezyme (Gaucher's disease); Aldurazyme-MPS-1 (Hurlers syndrome); Fabrazyme (Fabry disease)]; therapeutic enzymes [e.g., Epogen (anemia); activase (tissue plasminogen activator, thromobolysis)]; and others [including ABX-EGF (colorectal cancer); LymphoCIDE (NHL)]. See also USP 5,041,379 and 6,485,937. The heterologous protein can also be a marker protein. The marker may be introduced by itself, or in conjunction with one or more other heterologous polypeptides. Such a marker may be used, e.g., to confirm that a construct is functioning as desired, to identify those larvae in which the heterologous construct is being expressed, etc. Suitable markers will be evident to the skilled worker and include, e.g., green fluorescent protein (GFP), DsRed, EYFP, ECFP, EVFP and derivatives of EGFP. See also the markers listed at the web site of BD Biosciences (Clontech). A heterologous polypeptide can be expressed as an unfused polypeptide, a fusion polypeptide, a recombinant occlusion body, etc. If it is desirable to secrete a heterologous protein, a mammalian (e.g., human) signal peptide can be replaced with an insect signal sequence, e.g., an insect signal peptide from the insect cuticle gene or adipokinetic hormone, or prepromellitin protein, from baculovirus gp64 or egt proteins, or others. Methods for introducing constructs of the invention into insects, such as a transgenic insect of the invention, are conventional. See, e.g., USP 5,593,669 and Example XIV for some typical methods. A skilled worker will recognize appropriate times (a time window) during insect propagation in which such super-infection is possible. In some embodiments, the super- infection results in transient expression of the recombinant gene. In other embodiments, the recombinant gene is stably introduced into a somatic cell of the insect. The method for producing a mammalianized heterologous polypeptide of interest may further comprise culturing the insect under conditions effective for expressing the heterologous protein and for glycosylating it in a mammalianized (humanized) fashion. The method may further comprise harvesting the mammalianized (humanized) glycosylated heterologous polypeptide. Methods for cultivating and/or breeding the insects are conventional. In some cases, for example when detrimental products, such as certain glycosylating enzymes, are being produced in an insect, specialized cultivating methods may be employed. Some methods for cultivating insects are discussed in USP 6,153,409 and in the Examples. Methods for harvesting and, if desired, purifying the heterologous protein, are conventional. One embodiment of the invention is a transgenic insect of the invention that is infected with a vector (such as a baculovirus-based vector, a transposon-based vector, or a retrovirus vector) that encodes a heterologous polypeptide of interest, operably linked to an expression control sequence. Another embodiment is a transgenic insect of the invention that expresses one or more glycosylation enzymes as discussed herein that allow for the production of a partially or completely mammalianized glycosylated polypeptide in the insect. Another embodiment is a transgenic insect of the invention that expresses such glycosylation enzymes, and that is infected with a vector that encodes a heterologous polypeptide of interest, operably linked to an expression control sequence. Another method for producing, in an insect, one or more heterologous mammalianized (e.g., humanized) glycosylated polypeptides of interest, comprises using a multiply transgenic insect, which is a transgenic insect as above (whose somatic and germ cells contain genomically integrated nucleic acids encoding glycosylation enzymes), whose somatic and germ cells further comprise genomically integrated recombinant nucleic acid encoding said heterologous polypeptide(s) of interest, operably linked to an expression control sequence. Although the polypeptide of interest may be expressed in a multiply transgenic insect as above, it is still considered to be "heterologous" to the insect. Methods to generate such multiple transgenic insects are conventional. For example, one can start with an insect that is transgenic for a set of glycosylation enzymes, and then insert into the host genome a transgene that expresses a heterologous polypeptide of interest. Alternatively, one can begin with an insect that is transgenic for a polypeptide of interest (such as collagen, IFN, etc), and then introduce into the host genome DNA encoding a set of glycosylating enzymes. Genetic crosses and/or sequential introduction of suitable constructs may be employed to generate a multiply transgenic insect. A multiply transgenic insect as above can be cultivated, and the glycosylated heterologous polypeptides made therein can be harvested, using conventional procedures. This aspect of the invention thus relates both to multiple transgenic insects as above, and to methods of using the insects to produce heterologous glycosylated polypeptides.

In some embodiments of the invention, the glycosylation genes in a transgenic insect are under the control of (operably linked to) a regulatable control system. Suitable regulatable control systems, which will be evident to the skilled worker, include the inducible expression promoters/enhancers discussed elsewhere herein, such as hsp70, or a Tet-based inducible system, used in conjunction with any suitable constitutive promoter (e.g., the Tet-CMV IE or the Tet- baculovirus Iel systems). The use of regulatable control sequences can allow for the glycosylation enzymes to be expressed at low levels, or not to be expressed, until the polypeptide of interest begins to be expressed. By "low levels" is meant, e.g., levels that are too low to achieve partially or fully mammalianized (e.g., humanized) polypeptides, and/or levels that are not toxic to the host. In one embodiment, the inducible promoter is a baculovirus-specific promoter. For example, a transgenic insect (preferably a larva) of the invention may comprise a set of glycosylation genes that are under the control of one or more late or very late baculovirus promoters. When the insect is propagated, little if any expression of the glycosylation genes occurs. However, following infection of the insect with a baculovirus vector containing a heterologous gene of interest, the baculovirus infection induces expression of the glycosylation genes, so that the heterologous polypeptide of interest which is expressed from the baculovirus vector is glycosylated as it is produced. This insures that potentially toxic glycosylation enzymes are expressed only, at a significant level, or primarily, during the period during which the enzymatic activity is required. Similarly, a multiply transgenic insect that comprises genomically integrated copies of both glycosylation enzymes and heterologous polypeptides of interest can be designed such that the polypeptide of interest and the glycosylation enzymes are expressed at suitable levels, at the desired time during insect growth, by selecting appropriate expression control sequences for each of the genes. A skilled worker can readily design suitable constructs, using, e.g., suitable combinations of inducible promoters, constitutive promoters, promoters expressed at different times (temporally regulated) during baculovirus infection, etc.

Another method for producing, in an insect, one or more heterologous mammalianized (e.g., humanized) glycosylated polypeptides of interest, does not involve using transgenic insects. Rather, in this aspect of the invention, an insect (preferably an insect larva) is infected with one or more vectors (preferably viral vectors) that comprise nucleic acid sequences encoding a recombinant polypeptide of interest and/or one or more glycosylation enzymes. The sequences encoding both the polypeptide(s) of interest and the glycosylation enzyme(s) are operably linked to expression control sequences. Any of the combinations of glycosylation enzymes discussed above may be introduced into the insect; and any of the expression control sequences, including regulatable promoters, may be used. A skilled worker will recognize what types of expression control sequences and what combinations of glycosylation enzymes are suitable. Any of a variety of vectors may be used. Preferably, the vector is a baculovirus-based vector, such as those described elsewhere herein. As noted, such vectors can carry large numbers of large inserts. Thus, a partial or complete set of glycosylating enzymes can be introduced into the insect on a single vector, insuring that the entire set of enzymes will be expressed in a given cell. In some embodiments, the heterologous polypeptide of interest is encoded on the same vector as the glycosylation enzymes; in other embodiments, it is carried on a separate vector. One, two, or even more baculovirus-based vectors may be introduced into an insect. The vectors may be introduced simultaneously, or sequentially, provided that they are introduced within the allotted time window. In another embodiment, the glycosylating enzyme and polypeptide of interest sequences are cloned into one of the transposon-based vectors described elsewhere herein, such as a piggyback vector, or into a retrovirus vector, and used to infect an insect. One embodiment of the invention is an insect comprising, in at least some of its cells, glycosylation enzymes as described above that allow the production of partially or completely mammalianized glycoproteins of interest in the insect, and a heterologous polypeptide. Another embodiment is an insect comprising, in at least some of its cells, an expressible recombinant nucleic acid encoding a polypeptide of interest, and expressible nucleic acid encoding glycosylation enzymes as described above that allow the production of partially or completely mammalianized glycoproteins of interest in the insect. Another embodiment is a method for producing, in an insect larva, a partially or completely mammalianized glycosylated polypeptide of interest that is heterologous to the insect, comprising introducing a vector comprising nucleic acid encoding said heterologous polypeptide, operably linked to an expression control sequence, into a transgenic insect larva, or progeny thereof, whose somatic and germ cells contain recombinant nucleic acid encoding one or more (e.g., two or more) of the glycosylation enzymes: a) beta-l,2-N-acetylglucosaminyltransferase I, b) beta- 1 ,2-N-acetylglucosaminyltransferase II, c) a βl, 4-galactosyltransferase, and/or d) a sialyltransferase, wherein each recombinant nucleic acid encoding a glycosylation enzyme is integrated in the insect genome, and is present in one or more copies, wherein each recombinant nucleic acid encoding a glycosylation enzyme is operably linked to an expression control sequence, wherein expression of said glycosylation enzymes allows for production of a partially or completely mammalianized glycosylated protein in the insect, and wherein if the insect (particularly if it is B. mori) contains genomically integrated nucleic acid encoding enzyme c), then the insect also contains genomically integrated nucleic acid encoding at least one of enzymes a), b) or d). Another embodiment is a method for producing, in an insect larva, a partially or completely mammalianized glycosylated polypeptide of interest that is heterologous to the insect, comprising introducing a vector comprising nucleic acid encoding said heterologous polypeptide, operably linked to an expression control sequence, into a transgenic insect larva, or progeny thereof, whose somatic and germ cells contain recombinant nucleic acid encoding one or more (e.g., two or more) of the glycosylation enzymes: a) beta-l,2-N-acetylglucosaminyltransferase I, b) beta-l,2-N-acetylglucosaminyltransferase II, c) a βl, 4-galactosyltransferase, and/or d) a sialyltransferase, wherein each recombinant nucleic acid encoding a glycosylation enzyme is integrated in the insect genome, and is present in one or more copies, wherein each recombinant nucleic acid encoding a glycosylation enzyme is operably linked to an expression control sequence, wherein expression of said glycosylation enzymes allows for production of a partially or completely mammalianized glycosylated protein in the insect, and wherein, if the insect is B. mori, the glycosylated polypeptide is not expressed in a tissue- specific manner (e.g., is not expressed specifically in the silk glands). Another embodiment is a library of transgenic insects of the invention (TRANSPILLAR larvae or other forms of the insect) expressing a variety (e.g., more than one, preferably at least about 50 different glycosylated proteins. Preferably, each member of such a library comprises, in its somatic and germ cells, expressible sequences encoding both a suite of glycosylation enzymes and one or polypeptides of interest (which are designated to become glycosylated in a mammalianized fashion). In a preferred embodiment, the sequences encoding the glycosylation enzymes are under the control of a regulatable expression control sequence, so the insect can be maintained without expressing the glycosylation enzymes (which are potentially toxic to the cells), and the glycosylation enzymes are not turned on until they are needed in order to glycosylate the polypeptide of interest. Another embodiment is a library of transgenic insects of the invention (TRANSPILLAR larvae or other forms of the insect) that can be used to glycosylate proteins in a variety of partial or complete glycosylation patterns. Any of the suites of glycosylation enzymes discussed elsewhere herein can be used. The number of suitable permutations of glycosylation enzymes can range between about one and abut 400. Preferably, at least one of the insects expresses a full complement of glycosylation enzymes, including, e.g., beta-l,2-N-acetylglucosaminyl- transferase II; a βl, 4-galactosyltransferase; an alpha 2,6-sialyltransferase; an alpha 2,3- sialyltransferase; a sialic acid synthase; and CMP-sialic acid synthetase (and, optionally, beta- 1,2-N-acetylglucosaminyltransferase I). As was the case for the library discussed above, the sequences encoding the glycosylation enzymes are preferably under the control of a regulatable expression confrol sequence, so the insect can be maintained without expressing the glycosylation enzymes (which are potentially toxic to the cells), and the glycosylation enzymes are not turned on until they are needed in order to glycosylate a polypeptide of interest. For example, the glycosylation enzymes can be placed under the control of one or more late baculovirus promoters, and expression of the glycosylation enzymes can be turned on by infecting such an insect larva with a baculovirus that encodes an expressible polypeptide of interest, which is destined to become glycosylated in a mammalianized fashion. Another embodiment is a method for producing, in an insect larva, a partially or completely mammalianized glycosylated polypeptide of interest that is endogenous or heterologous to an insect as described herein, or an insect as described herein, wherein the insect is not Bombyx mori In the foregoing and in the following examples, all temperatures are set forth in uncorrected degrees Celsius; and, unless otherwise indicated, all parts and percentages are by weight.

EXAMPLES Example I. General overview of one aspect of the invention A colony of lepidopteran insect larvae (Trichoplusia ni) is stably transformed with a set of genes important for mammalianizing (e.g., humanizing) their protein N-glycosylation pathways. The piggyBac system is used in a series of consecutive transpositional events to translocate a set of about 2-8 or more glycosylation genes (preferably a set of about 6-8 glycosylation genes) into the germline of insect embryos. Stable incorporation of these genes results in rnammalianization (humanization) of all endogenous glycoproteins. One indication that these genetic modifications are not lethal to these insects is that that the N-glycosylation pathway has been humanized in cultured insect cell lines with no obvious deleterious effects. The risk of such detrimental effects occurring is further assessed by transforming Drosophila melanogaster. This model system is amenable to more rapid experiments than is the T. ni system. In some experiments, a molecular regulator of expression, the tefracycline repressor, is incorporated into the design for lepidopteran fransformations. This design precludes fransgene expression until the insects are infected with the baculovirus vector. Transgene expression is switched off until the late phase of infection, when the insects have already been effectively converted to bioreactors for recombinant glycoprotein production and are doomed to die as a result of the viral infection, anyway. Modular piggyBac expression vector cassettes encoding various mammalian enzymes involved in glycoprotein processing are constructed. These constructs are tested for their ability to induce enzymatic activity during transient transfection of cultured insect cells. Subsequently, these piggyBac vectors are used to transform D. melanogaster and the overall physiological influence of mammalian glycoprotein processing enzyme expression is examined in these insects. If there are no adverse effects, the piggyBac vectors are used to fransform the lepidopteran host, T. ni. Alternatively, new constructs designed for regulated expression of the mammalian genes are constructed, tested, and used to transform T ni, as described above. After the transgenic insect lines are established, their N-glycosylation capabilities are examined using a model recombinant glycoprotein expressed during baculovirus infection. Subsequently, glycosylation of a biotechnologically relevant recombinant glycoprotein is examined using this virus-host system.

Example II. Experiments in insect cell lines Aspects of the invention can be carried out by adapting methods used in insect cell culture. See, e.g., U.S. Patent 6,461,863. Insect cell lines were genetically transformed to create improved hosts for the production of humanized recombinant glycoproteins by baculovirus vectors. Sf9 cells were transformed with an expression plasmid encoding the cDNA for a mammalian β4Gal-TI to create a transgenic insect cell line called SfB4GalT (Hollister et al. (1998) Glycobiology 8, 473-80). The β4Gal-TI cDNA was placed under the control of the promoter from a baculovirus immediate early gene called iel, which provides constitutive foreign gene expression in lepidopteran insect cells. SfB4GalT cells grew normally, supported baculovirus replication, and constitutively expressed the mammalian β4Gal-TI gene. In addition, unlike the parental Sf9 cells, SfB4GalT cells were able to produce terminally galactosylated recombinant glycoproteins, such as human tissue plasminogen activator, when infected with baculovirus expression vectors. An iel expression plasmid encoding a mammalian alpha 2,6- Sial-T (ST6GalI) was used to super-transform SfB4GalT cells and produce another transgenic cell line, Sfl34GalT/ST6. This new cell line encoded and expressed both β4Gal-TI and ST6GalI, grew normally, and supported baculovirus replication (Hollister et al. (2001) Glycobiology J , 1- 9). In addition, this cell line could produce terminally sialylated recombinant N-glycoproteins during baculovirus infection. Two analogous transgenic High Five® derivatives, Tnβ4GalT and Tnβ4GalT/ST6, also had the same capabilities as the corresponding Sf9 derivatives (Breitbach et al. (2001) Biotech. Bioengr. 74, 230-9). The major processed N-glycans produced by these cells are monoantennary structures in which only the lower branch, not the upper, is elongated. These results suggested that these cell lines lacked sufficient levels of endogenous GlcNAc-TII activity to initiate elongation of the upper branch, which is necessary to produce conventional biantennary N-glycans (Fig. 1). A new transgenic cell line, designated SfSWT-1, was prepared by transforming Sf9 cells with five different mammalian glycosyltransferase genes, including GlcNAc-TI, GlcNAc-TII, β4Gal-TI, ST6GalI, and alpha 2,3-Sial-T (ST3GalIV). SfSWT-1 cells encode and express all five transgenes under iel control, have normal growth properties, and support baculovirus replication. In addition, these cells can produce biantennary, terminally sialylated N-glycans identical to those produced by mammalian cells. See, e.g., Hollister et al. (2002) Biochemistry 41, 15093. SfB4GalT/ST6 and SfSWT-1 cells can also produce sialylated N-glycans even though these cells have no detectable CMP-sialic acid, which is required as the donor substrate for ST6GalI and ST3GalIV. Subsequent experiments showed that both transgenic cell lines require either fetal bovine serum or a purified sialylated glycoprotein in order to produce sialylated glycoproteins (Hollister et al. (2003) Glycobiology 13, 487-495). Without wishing to be bound be any particular mechanism, it is suggested that terminal sialic acids from these exogenous sources are probably recycled for incorporation into newly synthesized glycoproteins, an interpretation that is consistent with known mechanisms for sialic acid uptake and reutilization in mammalian cells. However, insect cells were further engineered for de novo CMP-sialic acid production to circumvent the need for an exogenous sialic acid donor (Aumiller et al. (2003) Glycobiology 13, 497-507). Example III. Selecting mammalian processing genes Modifying the results of comparative analysis of the mammalian and insect protein N-glycosylation pathways, we incorporate mammalian glycosylation enzyme genes, including GlcNAc-TII, β4Gal-TI, ST6GalI, ST3GalIV, sialic acid synthase (SAS), and/or CMP-sialic acid synthetase (CMP-SAS) genes, into an insect genome to compensate for the lack of these enzymes in insect larvae. GlcNAc-TII initiates elongation of the upper branch, which is necessary to convert N-glycan intermediates to conventional biantennary structures. β4Gal-TI, ST6GalI, ST3GalIV complete the elongation and terminal sialylation of N-glycans. Both sialyltransferase genes are incorporated because STδGall and ST3GalIV transfer sialic acids in alpha 2,6- or alpha 2,3-linkages, respectively, and some human N-glycoproteins have one linkage, some have the other, and some have both. Since transgenic larvae may not be able to scavenge sialic acid, the SAS and CMP-SAS genes are included to ensure a conventional source of CMP-sialic acid. SAS and CMP-SAS convert N-acetylmannosamine, a monosaccharide precursor that can be incorporated into the larval diet, to CMP-sialic acid. Addition of these transgenically engineered mammalian genes enables transgenic insect larvae to produce complex, terminally sialylated N-glycans. To counteract the possibility that the insects used have too little GlcNAc-TI or too much GlcNAcase activity to efficiently elongate the lower branch of N-glycan intermediates (see Fig. 1), or that the insects lack the transporter needed to move CMP-sialic acid into the Golgi apparatus, additional mammalian genes encoding GlcNAc-TI or a CMP-sialic acid transporter into the transgenic insects are incorporated as necessary. Increasing the level of GlcNAc-TI activity effectively is expected to counteract the negative effect of the GlcNAcase on N-glycan processing, as previously demonstrated in insect cell lines. Down-regulation of GlcNAcase gene expression is also used. Additional genes are incorporated into transgenic insects by either super-transformation or cross-breeding. Example IV. Selecting expression control sequences The baculovirus iel promoter/7zr5 enhancer (iel/hr5) combination is chosen for constitutive foreign gene expression. An advantage of using this combination is that baculovirus infection induces the expression of integrated transgenes under iellhrS control, which increases the levels of the enzymes needed for glycoprotein processing prior to the time the glycoprotein of interest is expressed. The Tet-mediated expression system provides regulatable gene expression when linked to the Cytomegalovirus minimal promoter (CMV). This system works effectively in insect systems. In addition, using the appropriate Tet repressor mutation, either repression or induction of gene expression, may be achieved upon exposure to tetracycline or doxycycline. We utilize the TetO and CMV promoter sequences to achieve controlled expression of the mammalian glycoprotein processing enzymes in the insect larvae, and test the utility of the Tet expression system for controlled expression from the ie-l/hr5 baculovirus immediate early promoter. Example V. Selecting a model recombinant glycoprotein The transgenic insect's ability to process recombinant glycoproteins during baculovirus infection is determined using GST-SfManl as a model. GST-SfManl is a glutathione-S- transferase (GST)-tagged, secreted form of an endogenous class I Sf9 cell alpha-mannosidase. This hybrid protein is well characterized and has been used as a model in previous studies of N-glycan processing in native and transformed insect cell lines. GST-SfManl allows us to progress relatively quickly through an analysis of the glycoprotein processing capabilities of our transgenic insects and to produce products, such as tissue plasminogen activator, transferrin, β- trace protein, and/or other N-glycosylated proteins of interest. Example VI. Preparation and testing of constructs for transformation of insects A. piggyBac vectors. The piggyBac element has a demonstrated capacity of at least 9.5 kb of inserted DNA, with an overall transposon size of 9.9 kb. Insertions up to 10 kb, with an overall size of 10.5 kb for the element, can be mobilized at normal frequencies. Gene expression vectors for transformation of D. melanogaster and T. ni are constructed using a cassette approach that allows us to insert different promoter regions between pairs of genes for analysis of expression in our insect systems. Each gene is individually PCR amplified to allow positioning of appropriate restriction enzyme sites on either side of the gene. The amplified products are cloned and sequenced to insure integrity. Each gene pair is then assembled from the individual amplified genes in a plasmid clone. The use of different restriction sites at the termini of each gene insures directional cloning of that gene in the plasmid. For example, gene pairs as indicated below can be designed to progressively extend the insect N-glycosylation pathway (Fig. 1). Other gene pairs can also be used, examples of which will be evident to the skilled worker. Each gene pair is tagged with a different fluorescent reporter gene for transformation. For this purpose we utilize the 3XP3 promoter driving expression of the DsRed, ECFP, and EYFP genes. The 3XP3 promoter is active in nerve tissues, principally the eye of the insect. Visualization of the GFP markers is possible not only in white-eye mutants, but also in pigmented eye wild type insects. Since there is no available white-eye mutant strain in the target insect, T. ni, this promoter is very useful in screening our transgenic lepidopterans. The three fluorescent protein markers chosen are distinguishable from each other using the appropriate wavelength filter, permitting the monitoring of multiple transformations in a single insect. The following scheme was employed to engineer the plasmids shown in Figure 4. Steps for assembling the intermediate elements of these constructs, such as gene pair cassettes, cassettes with the marker protein, etc. were conventional. Primers used to amplify sub-portions of the constructs were generated based on known sequences, which are readily available to the skilled worker. Convenient restriction enzyme recognition sites were added during PCR amplification and used to insert the PCR products into recipient plasmids. Some of these restriction sites are indicated in the structures shown in Figure 4. 1. Amplified HR5-IE1 element and cloned into TOPO to make pHr5IElR.TOPO.l. 2. Amplified IE1 promoter and cloned into TOPO to make pIElL.TOPO.l. 3. Excised IE1L from pIEL.TOPO.l, subcloned into pHr5IElR.TOPO.l to make pDIEl.TOPO. l. 4. Deleted Xbal site in pDIEl.TOPO.l to make pDIEl.TOPO.2. 5. Amplified BGH poly A signal, cloned into TOPO to make pBGHPolyA.TOPO.l. 6. Excised BGH poly A, cloned into pDIEl.TOPO.2 to create pDIEl.TOPO.3. 7. Amplified 3XP3 promoter, cloned into TOPO to make p3xP3.TOP0.1. 8. Subcloned BGH poly A signal from pBGHPolyA.TOPO.l into p3xP3.TOPO.l to make p3xP3.TOPO.2. 9. Amplified DSRed marker, cloned into TOPO to make pDSRed.TOPO.l. 10. Excised DSRed from pDSRed.TOPO.1, subcloned into p3xP3.TOPO.2 to make p3xP3DSRed.TOPO.2. 11. Amplified ECFP marker, cloned into TOPO to make pECFP.TOPO.l. 12. Excised ECFP marker from pECFP.TOPO.l, subcloned into ρ3xP3.TOPO.2 to make p3xP3ECFP.TOPO.2. 13. Amplified EYFP marker, cloned into TOPO to make pEYFP.TOPO.l. 14. Excised EYFP marker from pEYFP.TOPO.1, subcloned into ρ3xP3.TOPO.2 to make p3xP3EYTP.TOPO.2. 15. Excised 3xP3DSRed, 3xP3ECFP, and 3xP3EYFP cassettes from p3xP3DSRed.TOPO.2, p3xP3ECFP.TOPO.2, and p3xP3EYFP.TOPO.2, respectively. Subcloned each into pDIEl-TOPO.3 to create pDIE.DSRed.TOPO.3, pDIE.ECFP.TOPO.3, and pDIE.EYFP.TOPO.3, respectively. 16.Excised BGH Poly A from pBGH.PolyA.TOPO.l, subcloned into pDIE.DSRed.TOPO.3, pDIE.ECFP.TOPO.3, and pDIE.EYFP.TOPO.3 to create pDIE.DSRed.TOPO.4, pDIE.ECFP.TOPO.4, and pDIE.EYFP.TOPO.4, respectively. 17. Amplified human GlcNAc-TII, bovine β4GalT, rat ST6GalI, mouse ST3GalIII, mouse SAS, and mouse CMP-SAS, cloned each individual amplimer into TOPO (yielded 6 individual TOPO subclones). 18. Excised human GlcNAc-TII and bovine β4GalT from TOPO clones, subcloned into pDIE.DSRed.TOPO.4 to create pDIE.GnTII/GalT.DSRed.TOPO.4. 19. Excised rat ST6GalI and mouse ST3GalIII from TOPO clones, subcloned into pDIE.ECFP.TOPO.4 to create pDIE.ST6.1/ST3.4.ECFP.TOPO.4. 20. Excised mouse SAS, and mouse CMP-SAS from TOPO clones, subcloned into pDIE.EYFP.TOPO.4 to create pDIE.SAS/CMP.SAS.EYFP.TOPO.4. 21. Excised each DIE. enzyme l/enzyme2. eye marker cassette from the TOPO.4 clones listed in item #20 and subcloned into the piggybac vector, pXLBac-2, in-between the transposition elements in that vector. This set of steps resulted in the creation of the three plasmids shown in Figure 4, each encoding two "glycosylation enzymes" under hr5IEl control and a marker gene under 3XP3 control. In a variation of the above method, the bivalent promoter cassettes are excised and replaced with similar cassettes containing alternate confrol elements, examples of which will be evident to the skilled worker. For example, the hr5IEl promoter cassette noted above can be replaced with cassettes such as the following (bounded by appropriate restriction enzyme sites): <=hsp70-hr5-hsp70=> ^CMV -7xTetO-CMV => <^iel/hr5-7xTetO-iel/hr5=> The three plasmids shown in Figure 4 are used to create transgenic larvae in conjunction with a plasmid encoding the piggybac transposase. B. Testing piggybac vectors in vitro. Each constructed piggyBac vector is rapidly tested for its ability to express the relevant mammalian genes under control of the iel/hr5 promoter by transient fransfection assays in insect cell lines. Briefly, Sf9 cell cultures are individually transfected with various piggyBac vectors encoding the glycosyltransferases or with the empty promoter cassette vectors as negative controls. Immediate early expression plasmids encoding GlcNAc-TII, β4Gal-TI, ST6GalI, or ST3GalIV are available and are used as positive controls. The cells are lysed at 24 h post-transfection and lysates are used for conventional glycosyltransferase assays. Three of these assays have been previously described in detail (Hollister et al. (2001) Glycobiology H, 1-9; Hollister et al. (2002) Biochemistry 41, 15093- 15104). A different type of transient expression assay is needed to test the piggyBac vectors encoding SAS and CMP-SAS, which are the enzymes involved in sialic acid biosynthesis. In these assays, SfB4GalT/ST6 cells are transiently transfected with the construct encoding SAS and CMP-SAS, then, 24 h later, the cells are stained with a fluorochrome-conjugated lectin, Sambucus nigra agglutinin (SNA), which is specific for terminal alpha 2,6-linked sialic acids. SfB4GalT/ST6 cells cannot produce sialylated N-glycoproteins when cultured in serum-free media. If the piggyBac vector encoding SAS and CMP-SAS is functional, it induces Sfβ4GalT/ST6 cells to produce sialylated N-glycoproteins even when cultured in serum-free medium containing N-acetylmannosamine, and the transfected cells stain with SNA. One negative control for this assay is SfB4GalT/ST6 cells transfected with the empty promoter cassette vector and cultured in serum-free medium containing N-acetylmannosamine. Another negative control is to transform these cells with the piggyBac vector encoding SAS and CMP- SAS, but cultured in serum-free medium lacking N-acetylmannosamine. The positive controls are SfB4GalT/ST6 cells transfected with the empty promoter cassette and cultured in serum-free medium supplemented with fetuin, which supports N-glycoprotein sialylation by these cells. piggyBac vectors comprising the constructs shown in Figure 4 were tested for transient fransfection in Sf9 cells in culture. In the final step of constructing piggy Bac-based vectors for inserting glycosyltransferase genes into insects or insect cells, the restriction fragments carrying the glycosyltransferase genes (two genes per fragment) and the fluorescent protein marker gene were inserted into the piggyBac plasmid pXLBacII in two different orientations with respect to the piggyBac terminal repeat sequences (TR-L/IR-L and TR-R/IR-R). Hence, for each set of glycosyltransferase genes, two different piggyBac vectors were constructed, with the glycosyltransferase genes and the fluorescent protein marker in opposite orientations. The piggyBac vectors were tested to measure the activity of the glycosyltransferase genes. Unexpectedly, it was found that, for each individual glycosyltransferase gene, the vector in which the gene was oriented so that it pointed towards the left-hand piggyBac terminal repeat (TR-L/IR-L) produced significantly more glycosyltransferase activity for that particular gene than the piggyBac vector in which the same gene was pointing towards the right-hand terminal repeat (TR-R/IR-R). The glycosyltransferase activity levels for the piggyBac plasmids were also higher than those found with non-piggyBac plasmids carrying the same genes. Example VII. Testing transformation efficiency and the effect of transgene expression in the model insect system, Drosophila melanogaster The addition of mammalian processing enzymes extensively modifies the

N-glycosylation profile of endogenous proteins in the insect. N-glycans can directly or indirectly influence protein functions in many different ways. To assess whether alterations of endogenous N-glycoproteins resulting from our genetic manipulations are phenotypically acceptable, the expression of the mammalian enzymes is studied in a model insect system. D. melanogaster is used as the model insect system for fransformation experiments, since it can be efficiently transformed with piggyBac, easily handled, rapidly manipulated, and easily screened for transformation. An experimental protocol is used to determine whether or not the hr5-IEl promoted, constitutively expressed glycosylation enzyme transposon vectors cause detrimental effects upon expression in transgenic insects. Since severe detrimental effects may result in difficulties detecting fransformations at all, we use a co-transformation strategy that is more likely to produce interpretable results than single plasmid transformation attempts. The injections are performed simultaneously with and without a control piggybac vector expressing a complementary fluorescent eye color gene. This allows us to determine whether the glycosylation plasmids are capable of generating viable transformants,. If viable transformants are not found, then we can at least be assured that our fransformation experiments are performed correctly and that the glycosylation plasmids themselves are detrimental. In this case, other procedures, such as the use of regulated expression control sequences, are used. A variety of types of regulatable expression of glycosylation genes are employed. For example, insects transformed with a construct under the control of a regulatable expression control sequence, such as a TetO/CMV-IE construct, are directly compared under repressed and induced conditions in this system, allowing a well-controlled assessment of the effect of gene expression on the insect. The TetO/CMV promoter construct system is useful, at least because this system is already developed in Drosophila, and appropriate repressor strains are available. A rtTetR-MT strain is available for these transformations, which produces a mutant version of the Tet repressor protein that acts as an inducer of TetO/CMV expression when flies are fed on media containing tetracycline or doxycycline. Alternatively, a native Tet repressor fransformed Drosophila strain is used for the transformations; this permits suppression of the CMV promoter activity in the presence of tetracycline or doxycycline. In either case, induction or de-repression of gene expression is examined at various times throughout the life cycle of the transformed insects to determine what effect expression of mammalian glycosylation enzymes has on the insect. RT-PCR assays are performed on extracts following induction of expression to confirm expression and determine rates of accumulation of transcripts for the fransgenes. Glycosylation is assessed using glycosyltransferase assays as described elsewhere herein. If there is no noticeable effect on the transgenic insects with each of the individual constructs, mating and selection are performed to produce lines having two, and then three constructs, and similar analyses of toxicity and expression levels are performed. Alternatively, manipulations in lepidopteran insects include inducible promoters that can be activated upon infection with a baculovirus vector. Several outcomes of the introduction of mammalian glycosylation pathways are evaluated in this tractable model system. Possible undesirable outcomes that are tested for include, e.g., developmental abnormalities, sterility, incomplete or abnormal embryonic development. In other tests, lethality at any stage is evaluated following heat shock, or through crosses with appropriate Drosophila rTA repressor/activator strains, respectively. The pXLBacII-SAS/CMP.SAS-EYFP plasmid (the clone#42-3 plasmid) was tested by co-injecting 1052 Drosophila embryos with pXLBacII-SAS/CMP.SAS-EYFP and the pCaSpeR- hs-orf helper plasmid. A total of 396 hatched larvae (37.6%) and 100 Adults (51 males and 49 males) were recovered to establish crosses with wild type individuals. Of these, 1 family expressed the yellow fluorescence expected for this construct, verifying that these two enzymes are not toxic when expressed in transgenic insects. In another experiment to confirm these findings, 2038 embryos were injected, again using pXLBacII-SAS/CMP.SAS-EYFP, pBSII ITR1.1 K ECFP and pCaSpeR-hs-of helper plasmid. These embryos are studied as above. A co-injection experiment was also performed on 982 embryos using the control plasmid pBSII ITRl.lK ECFP in addition to the pXLBacII-S AS/CMP. SAS-EYFP and the helper plasmid. We recovered 195 hatched larvae (19.8%) and 54 Adults (18 males and 36 males), with one family expressing the cyan fluorescence marker of the pBSII ITR 1.1K ECFP control plasmid. To evaluate two other sets of constructs: A) The pXLBacII-ST6.1/ST3.3M-CFP plasmid (e.g., the clone#21-l plasmid) is tested by co-injecting the plasmid into Drosophila embryos along with the pCaSpeR-hs-orf helper plasmid. From these injected embryos, larvae are hatched, with some surviving to adulthood. Each of these surviving adults is mated to a wild type individual to produce crosses which are screened for fluorescent eye transformants. Further embryos are also injected and studied as above, to confirm the findings from the first set of injections. For example, injections may include drosophila embryos co-injected with the pXLBacII-ST6.1/ST3.3M-CFP, the internal control plasmid pBSII ITR1.1K-EYFP, and the pCaSpeR-hs-orf helper plasmid. B.) The pXLBacII-GnTII/GalT-DsRed plasmid (e.g., clone#57) is injected along with the helper plasmid into embryos. In one experiment, 1184 embryos, with 272 larvae (22.9%) hatched and 62Adults (28males and 34 males) recovered for mating with wild type individuals. Further analysis is performed as above. For each set of injections, control injections are performed using the pBSII ITR 1.1K ECFP plasmid and the pCaSpeR-hs-orf helper. For example, in one experiment 562 drosophila embryos were injected, 199 larvae (35.4%) hatched and 53 Adults (all males) were recovered and mated. The lack of any female survivors was rather unusual, but attributed to chance. In this case three families expressing the yellow fluorescence were obtained. Example VIII. Producing transgenic T. ni T ni is transfected with the piggyBac element. In brief, the protocol involves the timed harvesting of eggs from wax paper. T. ni prefer to lay their eggs when the lights go off. Timing the light cycle for 12 hours on and 12 hours off such that the moths begin laying eggs at 8:00 AM allows harvesting of eggs for the next two hours. The eggs are easily released from the wax paper by brushing, and collected into a glass petri dish. They are then washed in 2% formalin, rinsed with water, air dried for 15 minutes, and then picked up from the filter paper with a fine brush. The eggs are secured with double-sided tape to a slide for microinjection. T. ni eggs have a top and bottom symmetry, but the embryo develops horizontally around the egg. It is therefore impossible to determine where germ line nuclei are developing. Instead, we use the rapid diffusion of the injected DNA throughout the embryo, coupled with slow cellularization (up to 4 hours), to permit the injected DNA to make its way into germ line nuclei. A protocol for establishment of transgenic T. ni is outlined below: A. Establishing transgenic T ni expressing mammalian glycosylation enzymes: Surviving insects from microinjections with the constructs discussed above are individually mated with wild-type T. ni. These matings are performed by combining five female wild type moths with each surviving microinjected GO male. All GO females are mass mated to wild type males. Expression of the fluorescent marker in these GO insects is not necessarily a prerequisite for their selection for mating, since establishment of the fransgene in germ line tissue is not necessarily reflected as an expressed fluorescence. The progeny FI insects from these matings are screened for expression of the fluorescent marker in the eyes of adults, or in any other tissues. Some position effects can generate fluorescence in tissues other than the eye as well. The screening of these adults is performed immediately upon emergence from the puparium and prior to mating. These adults are anesthetized by exposure to ether, CO₂, or cold. Positive insects are selected for individual mating to wild type insects of the other sex. The proportion for mating is one FI positive to five wild type of the opposite sex. F2 progeny is screened from each FI line and fluorescent positive males and females from each line are mated to establish a homozygous lineage for each line. Once homozygous lines are established for each enzyme, they are examined for expression of the fransgene. In each case, RT-PCR are used to measure expression of the fransgene. Glycosylation is assessed as described elsewhere herein. B. Establishing a baculovirus-induced Tet-responsive system for expression of mammalian glycosylation enzymes in T. ni: It may be advantageous, or even necessary, to have these mammalian glycosylation enzymes expressed only during a baculovirus infection. We adapt the Tet-inducible strategy already shown to be effective in Drosophila (Stebbins et al. (2001) Proc. Natl. Acad. Sci. 98, 10775-10780) to the baculovirus infected lepidopteran system by generating a transgenic T. ni line that expresses the rtTA-M2 mutation of the Tet repressor protein (TetON) under the control of the baculovirus p6.9 late promoter gene (Hill-Perkins et al. (1990) J. Gen. Virol. 71, 971-976). A similar strategy has been employed to effect controlled expression of genes from the baculovirus very late plO promoter during baculovirus infections of cell cultures (Wu et al. (2000) J. Biotech. 80, 75-83). In our case, the p6.9 promoter, which is only active during baculovirus infection and is silent in the absence of baculovirus early gene expression, is used to ensure that expression of the N-glycan processing enzymes (in this case, under the control of a TetON inducible promoter) occurs before the recombinant glycoprotein of interest is expressed under polyhedrin control by the baculovirus vector. The TetON protein gene is linked to this promoter and assembled within a piggyBac transposon with a 3XP3-GFP marker gene for transfer into the genome of T. ni. The inducible expression of this protein is assessed once transgenic strains are established by RT-PCR assays after baculovirus infection. Since there are only three GFP derivatives that can be used simultaneously in a given insect (Example V), this TetON strain must be constructed independently of the mammalian glycosylation strains. Matings and screening by southern hybridization and baculovirus-inducible expression of glycosyltransferases establish the final combined homozygous strains. In these strains, the mammalian N-glycan processing enzymes are only expressed during baculovirus infection in the presence of tetracycline or doxycycline. Example IX. Expressing and purifying GST-SfManl from normal and transgenic insects GST-SfManl is the recombinant model glycoprotein that is used to evaluate the N-glycan processing capabilities of our fransgenic insects, as discussed above. A recombinant baculovirus encoding a secreted form of this product under the control of the strong polyhedrin promoter is available from a previous study (Kawar et al. (2000) Glycobiology K), 347-55). This virus is , used to produce GST-SfManl for structural analyses of the N-glycans produced by parental and transgenic insect larvae. To avoid wound-induced stress from injection, viral inoculations are done orally. Inoculum stocks suitable for oral infection (consisting of the pre-occluded virus) are prepared according to conventional protocols and the potency is determined by conventional bioassay procedures. For experimental infections, groups of synchronized early fifth instar T. ni larvae are given a small (50μl) plug of diet with the desired dose of viral inoculum. The insect is allowed to feed for a defined time interval and only larvae that have consumed the entire diet plug are included in the experiment. In experiments determining optimal times of harvest, larvae are harvested at preset time intervals (e.g. 84, 96, and 108 h post infection), and about 25 μl haemolymph is collected from each larva in a tube with buffer containing l-phenyl-2-thiourea to inhibit melanization. For production experiments, recombinant GST-SfManl is harvested at the optimal time post infection and purified by glutathione affinity chromatography, using a slight modification of a previously described method (Hollister et al. (2001) Glycobiology ϋ, 1-9). Briefly, the hemolymph is harvested from infected larvae in the presence of l-phenyl-2-thiourea, the samples are clarified by low speed centrifugation, and budded virus is removed by ultracentrifugation. The resulting supernatant is concentrated with polyethylene glycol and the precipitate harvested by centrifugation. The pellet is dissolved in glutathione column binding buffer [25 mM Tris-HCl pH 8.0, 250 mM NaCl and 1.5% (v/v) Triton X-100] and extensively dialyzed against this same buffer. The dialyzed material is then applied at room temperature to an immobilized glutathione-agarose column prepared from a commercial affinity matrix and equilibrated with column binding buffer. The column is then washed with excess column binding buffer, washed again with excess glycosidase buffer (5 mM Na₂HPO₄, pH 7.5), and the GST-SfManl is eluted with a small volume of glycosidase buffer supplemented with 10 mM reduced glutathione. Affinity-purified GST-SfManl preparations are re-dialyzed against glycosidase buffer (5 mM Na₂HPO₄, pH 7.5) to remove the glutathione and the total protein concentration is determined using a commercial Bradford assay. Samples of the starting material, flow-through, washes, and eluants are analyzed by SDS-PAGE with Coomassie blue staining or immunoblotting to monitor the purification procedure. To address a potential problem in the purification of GST-SfManl from baculovirus- infected insect larvae - that lipids in the larval hemolymph may interfere with binding of the recombinant glycoprotein to the glutathione affinity column - we significantly dilute the material to be applied to the column, then circulate it over the affinity column for an extended time period in a cold room. Alternatively, a different affinity purification method is used. We produce a recombinant baculovirus that encodes a 6X HIS-tagged version of GST-SfManl. The 6X-HIS tag allows us to use metal affinity column chromatography as an alternative approach to purify essentially the same model glycoprotein. The properties of this protein, including expression levels, secretion efficiencies, glycosylation, and N-glycan processing, are evaluated to ensure that it has the same desirable features as the GST-tagged form of SfManl.

Example X. Characterizing N-glycans produced by normal and transgenic insects Lectin blotting assays, together with stringent specificity controls, are a simple and effective way to analyze N-glycans on recombinant glycoproteins. See, e.g., Hollister et al. (2001) Glycobiology VL, 1-9; Breitbach et al. (2001) Biotech. Bioengr. 74, 230-9; Jarvis et al. (1995) Virology 212, 500-11; Jarvis et al (1996) Nature/Biotech. 14, 1288-92. The advantages of the lectin blotting method include simplicity and rapidity. Although lectin binding is an indirect method, when properly controlled, lectin blotting experiments are uniformly confirmed using more direct and sophisticated analytical methods. Lectin blotting assays are coupled with competing sugar and glycosidase controls, as previously described (Hollister et al. (2001) Glycobiology H, 1-9), to examine the compositions of the N-glycans on the GST-SfManl produced by normal or transgenic insect larvae. These analyses provide an initial view of the N-glycan processing capabilities in our transgenic insects and provide a justification for performing more labor-intensive and expensive, but more definitive and comprehensive, structural analyses. The N-glycans from GST-SfManl or other model glycoproteins produced by the normal or transgenic insect larvae are removed in preparation for the latter structural analyses. We have previously shown that GST-SfManl can be quantitatively deglycosylated using an endoglycosidase called peptide-N-glycosidase-F (PNGase-F). The behavior of GST-SfManl produced by baculovirus-infected Trichoplusia ni larvae is examined. If the latter protein is core-fucosylated, it is not completely deglycosylated with PNGase-F. This problem is addressed by using a mixture of PNGase-F and another endoglycosidase, PNGase-A (Tretter et al. (1991) Eur. J. Biochem. 199, 647-652). About 1 mg of purified GST-SfManl is required from each source for comprehensive N-glycan structural analyses. The N-glycans are released from 1 mg samples of the recombinant protein from each source by exhaustive endoglycosidase digestion, as described previously (Hollister et al. (2001) Glycobiology 1_L 1-9). The released N-glycans in the spent reactions are bound to graphitized carbon cartridges. The protein and salts are washed out with water, then total N-glycans eluted with acetonitrile. Alternatively, trifluoroacetic acid is used to separately elute neutral and charged (sialylated) N-glycan species for independent structural analyses (Handler et al. (2001) Biotechniques 31_, 820, 824-8). After elution from these cartridges, the N-glycans are analyzed by various chromatographic and mass spectroscopic methods, as described below. In addition, one can couple the PNGase-F-mediated release of N-glycans with various exoglycosidase treatments (Packer et al. (1998) Glycoconj J 15. 737-47). A comparison of the chromatographic or spectroscopic profiles of the N-glycans released with PNGase-F alone and those released and partially degraded by combined digestions with PNGase- F and an exoglycosidase are used to identify the terminal monosaccharides on N-glycans. For example, if one couples PNGase-F and sialidase treatments and the profile changes in the predicted fashion, then this provides direct evidence that the original N-glycan was sialylated. Many specific exoglycosidases are commercially available for this purpose, including β- galactosidases, alpha-fucosidases, β-N-acetylhexosaminidases, and alpha-mannosidases, and these reagents can be applied to effectively "sequence" N-glycans. While each specific endo- and exoglycosidase reaction requires specific buffers and other conditions, these are readily available from the literature and manufacturer's recommendations. There are many conventional ways to analyze N-glycan structures. We use one common chromatographic method known as high pH anion exchange chromatography with pulsed amperometric detection (HPAEC-PAD). See, e.g., Hollister et al. (2001) Glycobiology ϋ, 1-9; Hollister et al. (2002) Biochemistry 41, 15093-15104. N-glycans are isolated from the GST- SfManl produced by various larvae, as described above, then injected into an HPAEC-PAD system equipped with a Carbo-Pac PA100 column equilibrated with 50 mM NaOH. This column is specifically designed for oligosaccharide separations. After being injected, the column is washed with 50 mM NaOH, then N-glycans are eluted with a linear gradient of 0 to 125 mM sodium acetate over 45 minutes at a flow rate of 1 ml/min. Commercial N-glycans and/or N-glycans from the GST-SfManl produced by our normal and fransgenic insect cell lines are used as standards. The latter structures have been unequivocally determined using mass spectroscopic and tandem mass spectroscopic methods. In addition, some commercial monosaccharide standards, particularly sialic acid, can be useful for these experiments. Using these standards, we are able to identify any co-eluting N-glycan structures isolated from the GST-SfManl or other recombinant glycoproteins produced by normal or fransgenic insect larvae. Together with the data obtained from exoglycosidase sequencing experiments, we are able to identify N-glycan structures with great confidence. The unequivocal, comprehensive determination of N-glycan structures are carried out using mass spectroscopic and tandem mass spectroscopic analysis of N-glycan samples. The presence of contaminants in the N-glycan preparations that can interfere with pulsed amperometric detection can be circumvented by using established methods to label N-glycans with various fluorochromes, such as 2-aminobenzamide, which enables their specific detection if a fluorescence detector is added to the HPAEC-PAD system (Kotani et al. (1998) Anal Biochem 264, 66-73). Example XI. Introducing multiple glycosylating enzymes into T. ni The iel/hr5 subset of the piggyBac vectors described above is used to introduce the mammalian processing enzymes directly into the baculovirus genome. A baculovirus encoding each of the mammalian processing enzymes discussed above is used as the parental virus for the production of baculovirus expression vectors encoding recombinant glycoproteins with mammalian glycan profiles. The piggyBac vectors encoding GlcNAc-TII, β4Gal-TI, ST6GalI, ST3GalIV, SAS, and CMP-SAS under the confrol of the iel promoter are used to introduce these genes into the genome of the baculovirus, Autographa californica multicapsid nucleopolyhedrovirus (AcMNPV). Briefly, viral genomic DNA isolated by a conventional method is mixed with the appropriate piggyBac vector DNA in the presence of a helper plasmid encoding the fransposase. Expression of the transposase helper is driven by polh, and the mixture is used to transfect Sf9 cells by conventional transfection procedures. Medium from the transfected cells is harvested four days later and budded virus progeny resolved, using conventional baculovirus plaque assays (see, e.g., O'Reilly et al. (1992) "Baculovirus expression vectors." W.H. Freeman and Company, New York). Recombinants are identified by the presence of fluorescent protein markers, which can be visualized directly in the infected cells. This is done in a step wise fashion, inserting each vector construct independently and sequentially, or by simultaneous insertion of all three constructs. Recombinant baculovirus clones that have all three fluorescent protein markers are amplified and checked for each of the transgenes of interest by dot blot assays. Virus clones that have all of the mammalian genes of interest are examined for their ability to express those genes during infection of Sf9 or Sf 4GalT/ST6 cells, as described above. This yields a novel baculovirus expression vector with the mammalian glycoprotein processing genes that are needed to extend the insect N-glycosylation pathway. An existing transfer plasmid is used to introduce an E. coli LacZ marker and three Bsu36I sites into the polyhedrin locus of this virus, by analogy to a previously described baculovirus vector (Kitts et al. (1993) Biotechniques M, 810-7). This greatly facilitates the subsequent use of the resulting recombinant virus as a parental strain for the isolation of secondary recombinant baculovirus expression vectors encoding glycoproteins of interest under the confrol of the strong polyhedrin promoter. We introduce the polyhedrin-driven GST-SfManl gene into this parental virus using conventional protocols, and use the resulting virus to express GST-SfManl in normal insect larvae. The existing virus, AcGST-SfManl, which lacks the mammalian glycoprotein processing genes, is used as a negative control. Subsequently, both forms of GST-SfManl are affinity- purified and analyzed by lectin blotting assays, as described above. In addition, the N-glycans are removed from each protein preparation and used for more comprehensive structural analyses, as described above. Example XII. Negative regulation of endogenous GlcNAcase activity A recent study suggests that it might be possible to enhance the efficiency of N-glycan processing in insect cells by inhibiting the endogenous GlcNAcase activity (Watanabe et al. (2001) J. Biol. Chem. 277, 5090-5093). While this approach is economically is not feasible for large scale protein production using insect larvae, an attractive alternative approach uses RNA- interference (RNAi) to reduce or eliminate GlcNAcase activity. This approach requires at least a partial GlcNAcase gene sequence. This inhibition can enhance the efficiency of producing mammalianized proteins in the transgenic insects of the invention. Inhibiting the GlcNAcase in normal insects, which contain no mammalian glycosyltransferase genes, may be desirable with regard to on N-glycan processing by these organisms. Methods to design and generate RNAi specific for a nucleic acid sequence are conventional. In one embodiment, short selected double stranded sequences are synthesized chemically and annealed. In another embodiment, the two strands of the double strand siRNA are transcribed from a suitable expression vector in vitro, annealed, and transfected as dsRNA into the cells. In another embodiment, the GlcNAcase cDNA is used to construct a piggyBac vector encoding an inverted repeat corresponding to all or part of the GlcNAcase coding sequence, with a short spacer sequence in-between. This sequence is placed under the confrol of the iel or iel-tef" promoter for constitutive or regulated production of a dsRNA molecule with a stem-loop structure, which mediates post-transcriptional gene silencing (Kennerdell et al. (2000). Nat Biotechnol 18, 896-8). The GlcNAcase stem-loop construct is assembled from two PCR products encoding the entire open reading frame flanked by unique restriction sites, essentially as described (Kennerdell et al. (20O0), supra). One PCR product begins with a unique Bglll site and the other begins with a unique Spel site. Each has a slightly different Sfil site on its 3' end, which, when digested and religated, produces dimers with a nonpalindromic, central 5 bp sequence. This sequence serves as the spacer between the inverted repeats and will create the loop in the RNA stem-loop structure. The PCR amplimers are digested with Sfil, ligated, and dimers are gel-purified, digested with Bglll and Spel, and subcloned downstream of the iel or iel-tef" promoter in the piggyBac vectors described above. The resulting piggyBac vector is used to transform or super ransform T. ni larvae, as described above. Ultimately, GST-SfManl is produced in larvae known to be expressing the RNAi construct, affinity-purified, and its N- glycans are isolated and analyzed, as described above. siRNAs specific for the cloned GlcNAcase, or for a portion thereof, are expected to reduce or eliminate this enzyme activity in cultured Sf9 cells, and thus possibly to increase the efficiency of glycoprotein sialylation. Expression or introduction of an interfering RNA is also expected to reduce or eliminate

GlcNAcase activity in transgenic insects expressing mammalian glycosyltransferases. See, e.g., Kramer and Bentley (2003) Metabolic Engineering 5, 183-190, which reports that an siRNA against GFP (green fluorescent protein) is effective to inhibit expression of that protein in T. ni larvae. Example XIII. Overcoming potential immunogenicity problems Some insects (e.g., T. ni) have an alpha- 1,3-fucosyltransferase (FT3) that can add alpha 1,3-linked fucose residues to the linkage sugar of N-linked glycans (Marz et al. (1995). Protein glycosylation in insects. In "Glycoproteins" (J. Montreuil, J. F. G. Vliegenthart, and H. Schachter, Eds.), Vol. 29a, pp. 543-563. Elsevier, Amsterdam; Kubelka et al. (1994) Arch. Biochem. Biophys. 308, 148-157; Staudacher et al. (1992) Eur. J. Biochem. 207, 987-993). This activity has been observed, and genes encoding this enzyme have been cloned and sequenced from, e.g., Arabadidopsis, Drosophila and C. elegans. The presence of this enzyme is a potential problem because the addition of this fucose residue generates an immunogenic carbohydrate epitope related or identical to the horseradish peroxidase (HRP) epitope found on some plant glycoproteins (Fabini et al. (2001) J Biol Chem 276, 28058-67). One method to address this problem is to first identify whether the transgenic insects have alpha 1,3-linked core fucose residues by structural analyses of the N-glycans isolated from recombinant glycoproteins they produced. If they have this moiety, the problem is addressed in one or more of the following ways. A. Post production enzyme treatment: The simplest solution is to treat the purified recombinant glycoprotein with alpha-fucosidase. This enzyme is absolutely specific for terminal, alpha-linked fucose residues and is widely used to remove fucose residues from N-glycans (Jacob et al. (1994) Meth. Enzymol. 230, 280-99). Samples of the purified recombinant glycoprotein taken before and after treatment are analyzed by western blotting with a commercially available anti-HRP antibody, which only binds to the glycoprotein if it has alpha 1,3-linked fucose (Fabini et al. (2001) J Biol Chem 206, 28058-67). If alpha-fucosidase treatment is effective, the recombinant glycoprotein is separated from the enzyme and the preparation is complete. The completed preparation is deglycosylated and the structures of the released N-glycans directly determined, as described above, to confirm defucosylation at a higher level of sensitivity. If the western blots or direct structural analyses indicate that alpha- fucosidase treatment did not effectively defucosylate the recombinant glycoprotein, an alternative solution is undertaken. B. Characterization of alternative lepidopteran fucosylation properties: Another method is to identify an AcMNPV-permissive insect species with no FT3 activity by analyzing the FT3 status of different lepidopteran insect species, including T. ni, Spodoptera frugiperda, Estigmene acrea, Heliothis virescens, and Spodoptera exigua. For these assays, a BEV is used to express a recombinant glycoprotein of interest in each insect, then the product is isolated and probed for alpha 1,3-fucose using the anti-HRP antibody. Any glycoprotein preparation that fails to react with this antibody is deglycosylated with a mixture of PNGase-F and PNGase-A and the N-glycans are recovered and their structures directly analyzed using HPLC or mass spectroscopy, which provide a higher level of sensitivity, as described above. An AcMNPV- permissive host that lacks FT3 is used in place of T. ni as the parental insect for the transgenesis experiments described above. Alternatively T ni can be used because this is the insect used in mass larval rearing and infection for recombinant protein expression. C. RNAi suppression ofFT3 expression: Another method is to prepare an insect by using the RNAi approach, by analogy to the experiments described above for knocking out the GlcNAcase gene. This solution to the immunogenicity problem requires isolation of the FT3 gene from T ni, which is needed to produce a fransgenic insect that constitutively expresses a fragment of this gene as DS RNA. A partial sequence of a Trichoplusia ni core αl,3 fucosylfransferase has been cloned and sequenced. Amino acid sequences from the demonstrated core αl,3 fucosylfransferase from Drosophila melanogaster (Fabini et al. (2001) J Biol Chem 276, 28058) and putative core l,3 fucosyltransferases of Anopheles gambi e and Apis mellifera were aligned with each other by ClustalW. Regions of high sequence conservation among the three sequences were identified and used to design degenerate oligonucleotid.es for PCR. Degenerate PCR with one pair of primers yi elded a product of the predicted size. When this 218 bp PCR product was cloned and sequenced, it found to encode an amino acid sequence with a high level of sequence identity to the other insect core αl,3 fucosylfransferase sequences. When a BLAST-p homology search of non-redundant amino acid sequences was carried out using the T. ni PCR product amino acid sequence as a query, the highest match was with the Apis mellifera core αl,3 fucosylfransferase. The sequence of the T ni fragment is:

GTGGCGTGGTTTGTTTGGAACTGCCACGCCCGCAACCGCCGCCTGCAGTACGCGCGG CAGCTCAGCAGGCACATCCAGGTGGACATCTACGGTGCGTGCGGCTCGCACCACTG CCCCCGCACTGACCCCAACTGCCTGGAGATGCTCGACAGGGACTACAAGTTCTACCT CGCATTTGAAAATTCTAACTGTCGTGATTACATCACAGAGAAGTTCTT (SEQ ID NO:l) siRNAs are designed, using conventional procedures, that are specific for the entire sequence of SEQ ID NO: 1, or for fragments thereof. The siRNAs are first tested for efficacy in cell culture, and are then introduced into insects of the invention. Other conventional methods for suppressing FT3 expression are also employed. These methods include, e.g., the use of antisense nucleic acid, or generating "knockouts" of the gene by, e.g., homologous recombination. The elimination of FT3 activity is useful, not only in the context of insects that produce mammalianized glycoproteins, but also for insects that are not modified to produce mammalianized glycoproteins. For example, insect-like glycoproteins that have been treated to remove alpha 1,3-linked fucose residues, and thus lack that immunogenic carbohydrate epitope, can be useful as vaccines; the major epitopes in such a vaccine are from the polypeptide of interest, itself, rather than the "non-mammalian" carbohydrate residue. Accordingly, a form of "non-insectivized" polypeptide is one in which alpha 1,3-linked core fucose residues are absent from the linkage sugar of an N-linked glycan. Such a "non- insectivized" heterologous polypeptide can be generated in an insect (e.g., a transgenic insect), wherein the insect is selected or modified so as not to express FT3 in its cells, using any of the methods described above. Optionally, such an insect may also express in its cells suitable recombinant glycosylation enzymes, as is discussed elsewhere herein.

Example XIV - Methods for introducing polypeptides of interest into a transgenic animal that expresses mammalianizing glycoproteins Typically, inoculation of larvae has been done by injection with budded virus or feeding of occluded virus. Preferably, a different route is used in methods of the invention, because automated injection of larvae is not feasible and oral infection with occluded virus is detrimental for product protein yield (competition of polyhedrin synthesis) and complicates sanitation. A preferable form of inoculation is oral inoculation, using a pre-occluded virus (POV) form. This is virus localized in the nucleus and destined to be occluded in a paracrystalline matrix of the protein polyhedrin, except that the polyhedrin gene is deleted from the viral genome. Conventional methods may be used. Guidance regarding oral inoculation with POV inoculum is provided in USP 6,090,379 and 5,593,669. An exemplary embodiment is described below: Preparation of POV inoculum Early 5^th instar T. ni larvae are injected with budded virus and incubated. Larvae are monitored for symptoms of infection and mortality. Moribund larvae are collected and frozen. The frozen larvae are then lyophilized. When lyophilized cadavers are removed from the freeze dryer the % solids of the lyophilized cadavers is confirmed to be between 21% and 23%. The dry cadavers are then milled into bulking material to form a wettable powder which serves as the POV inoculum. The WP is stored at -80C and serves as POV inoculum stock. Inoculation of larvae with POV A suspension of the POV stock is prepared in water containing 2.5% sucrose. This suspension is screened through a 48 mesh sieve to remove debris that would plug hypodermic needles on the inoculator, and is then ready for use. The virus inoculator consist of four parts: a) a pump, connected to b) a manifold with hypodermic needles in a pattern fitting that of the wells in the trays with insects c) a platform that moves the manifold with the needles up and down d) guardrails that allow the trays with insects to be placed directly under the needles The operation of the machine depends on a foot-pedal switch-activated and compressed air-powered depression of the platform. This action forces the hypodermic needles through the topfilm of the trays while at the same time a defined amount of inoculum is sprayed in the chambers. The platform then pulls the needles out of the wells, and a new frays can be placed under the platform. The sequence and coordination of events is controlled by microswitches. The effective dose applied by the inoculator machine to each well is equivalent to 33ug lyophilized cadaver/well. This may be adjusted based on potency of POV inoculum Example XV - Growth conditions (Insect mass rearing: Process variables) The mass rearing of Trichoplusia ni for protein manufacturing falls into two functionally different processes. The first is maintenance of a breeding colony, based on the insect's life cycle of approximately 4 weeks. The second process pertains to the diversion of large numbers of larvae from the breeding colony to serve as production larvae. Maintenance of the breeding colony. Breeding methods for some Lepidoptera are well established, and in this category fall noctuid moths such as the cabbage looper (Trichoplusia ni). The life cycle consists of 4 stages: egg stage, larval stage, pupal stage, and adult stage. The inventors have determined optimal conditions under which the insects need to be kept, and have established protocols for handling of the insects during each of these stages. The following lists the tolerances in conditions and indicates some alternative handling procedures for T. ni. Eggs. The egg stage is short (about 3 days). Eggs are typically laid on a solid substrate such as paper towels or muslin cloth. T. ni deposits its eggs separately on the substrate to which the eggs stick. Eggs are removed fro the substrate and collected using a dilute bleach solution. After rinsing the eggs they are incubated in a moist bulking agent until one day before egg hatch. Then the eggs are "packaged" by a form-fϊll-and-seal machine in a continuous, automated process. This process starts with indentations (wells) being thermoformed in a sheet of PVC film (the web), and flash-sterilized, liquid, semi-synthetic insect diet is distributed into the wells via a manifold. The web then moves through a cooling tunnel where the diet solidifies. Next the eggs in the bulking agent are deposited onto the diet which has solidified. Finally, at the end of the line, perforated film is thermosealed over the wells. For a period of approximately one day the eggs remain on the diet under standard incubation conditions until the larvae hatch. Process variables to be optimized include: Substrate for oviposition; egg removal procedure (% bleach, immersion time); bulking agent; type of diet; type of top film and perforations (gas exchange); and incubation conditions (temperature, relative humidity, light regimen) Larvae. Larvae hatch as neonates and after eating the remains of the egg shell, they start feeding on the synthetic diet. The larvae when incubated under standard conditions grow over a period of 12 days through 5 instars and pupate. Process variables to be optimized include: incubation conditions, such as temperature, relative humidity, and light regimen. Pupae. Pupae embedded in a cocoon stay in average for 3 days in the wells under the same conditions as for larval growth. Then the pupae are released from their cocoons and placed into the adult emergence cages. Process variables to be optimized include: cocoon removal procedure (manually, % bleach, immersion time); and incubation conditions (temperature, relative humidity, light regimen). Adults. After 1-2 days both female and male adults emerge and they are allowed to mate and lay eggs. Eggs are collected daily. Process variables to be optimized include: type of adult emergence cage (carton, wire cage); number of adults per cage; incubation conditions (temperature, relative humidity, light regimen). Production larvae. Massive numbers of larvae are diverted from the colony maintenance cycle and are used as the hosts for protein production. The sheer numbers involved make automation a necessity. Essentially 99.9% of the insects packaged in the form-fill-and-seal- machine are inoculated with recombinant baculovirus as late instars. While the inoculated larvae keep eating and growing for several days more, their development is halted by the viral infection and they do not pupate. These larvae are harvested at the appropriate time, frozen and are then ready to enter the process of protein purification. Process variables to be optimized include: incubation conditions (temperature, relative humidity, light regimen); inoculum dose; inoculum timing; harvest timing.

Example XVI - Transformed Bomhyx mori with mammalian glycosylation capabilities for production of mammalian proteins in the silk gland The p25 promoter of the silkworm, Bombyx mori, is used to obtain organ-specific expression of genes in the posterior silk gland. The silk fibroin light or heavy chain gene promoter is used to obtain organ specific expression of genes in the median silk gland. The piggyBac transposon vector technology, as described elsewhere herein, is used. Silkworms can be maintained in the absence of silk production. Using conventional procedures, piggyBac-based transformation vectors are constructed which can introduce mammalian glycosylation enzymes for restricted expression (or restricted and controlled expression) in the silk gland. Bivalent promoter cassettes are constructed that allow for the expression of two mammalian glycosylation enzymes simultaneously from one transformation vector and a selectable fluorescent marker gene. Using conventional microinjection protocols, the vector and a helper plasmid that provides the fransposase protein are introduced into embryos of Bombyx mori and transformed insects are selected. Conventional tests (e.g., PCR. Protocols) are used to test for expression of the mammalian glycosylation enzymes in the silk gland of transformed insects. A second vector is then applied which contains additional mammalian glycosylation enzymes, and successful transformants are selected for as above. These steps are repeated until all the desired mammalian glycosylation enzymes are established in the genome of Bombyx mori. Transformed strains expressing individual combinations of glycosylation enzymes are mated to establish a single strain expressing all the desired mammalian glycosylation enzymes. Alternatively, a single strain is transformed to establish a multiply transformed strain expressing all the desired mammalian glycosylation enzymes. In one embodiment, fucosylation is inhibited by the expression of RNAi to knock out expression of the Bombyx mori endogenous fucosylation enzymes. Optionally, mammalian fucosylation enzymes are inserted and expressed above. A skilled worker will recognize than any of the methods described herein with regard to T ni or other insects can be adapted for use in B. mori, following conventional procedures. Example XVII - TRANSPILLAR Larvae Commercialization Chesapeake PERL has developed an automated process to generate large numbers of T. ni larvae in thermoformed habitats. These larvae are inoculated at the appropriate stage and harvested in a labor-extensive, semi-automated step. Finally, after processing the larvae, the protein product is recovered and purified to the required purity. This process is currently operational and enables at capacity the rearing and processing of circa 1 million larvae per week. While yields vary significantly for different types of proteins, 200 μg/larvae is a reasonable average yield estimate, in our experience. This indicates a production capacity of ca 200 grams of recombinant protein per week. This methods disclosed herein help solve the crisis in biopharmaceutical manufacturing by making the development cycle for new biotechnology-based therapeutics more predictable and less difficult. A suite of technologies is developed based on inventive transgenic-modified caterpillars — TRANSPILLARs. Combining TRANSPILLAR larvae with fully developed protein manufacturing process enables efficient, high- volume, cost-competitive development of a broad range of biopharmaceuticals. It eliminates scale-up issues and allow the entire development cycle, from discovery to manufacturing, carried out using one expression process. This helps the biopharmaceutical industry fulfill its promise of improved health and eradication of disease by removing years from drug development, reducing costs by millions, and in some cases ensuring the marketing of new therapeutics from emerging innovators that would have otherwise failed. The process uses whole cabbage looper caterpillars in an assembly line type procedure, which transforms the caterpillars into near-perfect, self-regulating "mini-bioreactors," with self- optimized cell growth and protein expression. The transformation occurs via infection with a baculovirus. The baculovirus vector delivers the gene encoding the protein of interest to susceptible host cells while providing the control elements needed to express at extraordinarily high levels. The infected cell provides the complex enzymatic machinery for expression and post-franslational processing. Each insect serves as a discrete and predictable unit of production: it sustains exquisite homeostasis; it has a rudimentary immune system, which maintains internal sterility; it respires, which maintains optimal dissolved oxygen for cell growth; and it eats and excretes, which maintains optimal pH and nutrient concenfration. And by being more densely packed than any possible concenfration in vitro, the insect system optimizes space. Further, using a whole, self- contained organism greatly reduces operator intervention, sterile handling, process controls, and ultimately possible process variables and deviations. The overall process is enabled by the patent-protected use of the orally infectious pre-occluded virus morphotype (POV) used to infect cells via the diet (rather than physical injection). Mass production of protein in insects is similar to bioreactor-based cell culture, but there are important differences. Both require a vector, growth phase, infection/induction, expression and harvest, and clarification and product separation. However, unlike insect-based production, cell culture processes require sterile seed trains, multiplicity of infection, cell counts, more stringent process controls, and more capital and labor. Moreover, the inventive manufacturing method requires fewer steps, and most importantly, it vastly improves scale-up, because no process engineering is required. Instead, with each larva treated as a unit, you scale up simply by growing more TRANSPILLAR larvae. In other words, the system completely removes the exorbitant process development and scale-up cycle. Reactor scale up issues, such as oxygen mass transfer, shear, and gas mixing, are obviated, as the system has demonstrated scalability from microgram research quantities to multiple kilograms — commercial scale quantities — easily and linearly, within one to two weeks. Three commercially valuable components of the invention are TRANSPILLAR larvae (Transgenic Insects): A tool for biopharmaceutical manufacturing. A stable line of transgenic-modified caterpillars to be used as a platform that expresses recombinant proteins with human glycosylation without the immunogenicity associated with insect-mediated expression. PERL SOLUTIONS (Process Out-Licensing): A complete commercial process licensing package, optimized for efficient manufacture of proteins using TRANSPILLAR larvae, and constant regardless of the protein. C-PERL CONTRACT MANUFACTURING: Complete contract manufacturing expanding over time: from research grade, to final semi-purified bulk for final purification, to final purified bulk API for fill and finish. TRANSPILLAR larvae decrease development, scale-up, and rework costs. Because failures account for 75% of the $880 million to develop a new drug, TRANSPILLAR larvae should therefore drastically reduce costs, thus often enabling market entry before funding is exhausted. The benefits to direct customers during each phase of drug development: 1) Discovery: Biotechnology companies, universities, federal laboratories, and research institutions discover proteins that scientists seek to produce in the effort to find those with applications to treat diseases. They need to produce milligram quantities of large numbers of widely varying proteins for testing. As experts in protein expression vectors, they can readily transition to the baculovirus vector required to utilize the TRANSPILLAR larvae. They can use the TRANSPILLAR larvae kit on the bench top without additional capital investment or specialized equipment. The TRANSPILLAR larva is easier to use, does not need sterile conditions, and can produce the needed glycosylation. The customer benefits from ease of use, low capital investment, high yields, and a clear path to development. 2) Development (defining and producing proteins that may become drugs): Development requires larger amounts of proteins, mg to gram quantities, which can be produced in the TRANSPILLAR larva, without a the pilot plant, at about half the size and cost. The cycle time is weeks rather than months. 3) Preclinical (early FDA-mandated safety and properties testing): Depending on the type of protein, anywhere from ten to hundreds of grams are needed. These could be produced by contract manufacturing or under license, in-house, with 10,000 to 100,000 insects in one bench top incubator. 4) Clinical trials (three phases): Companies produce enough material for all three trials, often 100s of grams, and with TRANSPILLAR larvae can use essentially the same process. For example: Company X identifies a promising recombinant protein drug candidate.

The drug has the desired pharmacological characteristic, and Company X is ready to produce milligram quantities for lead optimization and preclinical studies. The preferred method of expression: C-PERL Solutions from Chesapeake PERL (Protein Expression and Recovery Labs). C-PERL's fransgenic insects (TRANSPILLAR larvae) produce the same quality product as cell culture, without immunogenicity, and add full mammalian-type glycosylation for full biologic activity and stability in serum. Company X is now ready to develop the drug. During discovery and development, the R&D staff purchases kits from C-PERL, uses a few dozen TRANSPILLAR larvae on the bench top, and gets the same quality product as the pilot plant. Next, the plant purchases TRANSPILLAR larvae and licenses the completely developed C-PERL Solutions process to manufacture early clinical material—without process development. Further, because the process scales linearly, Company X knows the needed commercial manufacturing capacity early in development. They decide to contract manufacture with C-PERL, and save 5 years in the development phase and millions cutting out technology transfer and scale-up. Their completed Phase I trial data and clear path to commercial manufacturing help oversubscribe a Series B round of financing. C-PERL Solutions manufactures enough drug for Phase II and III trials, and immediately after Phase I Company X begins treating patients under the Investigational New Drug Treatment policy. As they break ground on a new research facility for new lead compounds, the Agency approves the Biologies License Application.

From the foregoing description, one skilled in the art can easily ascertain the essential characteristics of this invention, and without departing from the spirit and scope thereof, can make changes and modifications of the invention to adapt it to various usage and conditions. Without further elaboration, it is believed that one skilled in the art can, using the preceding description, utilize the present invention to its fullest extent. The preceding preferred specific embodiments are, therefore, to be construed as merely illustrative, and not limitative of the remainder of the disclosure in any way whatsoever. The entire disclosure of all applications, patents and publications (including priority document, U.S. provisional application 60/514,741), cited above and below and in the figures are hereby incorporated by reference.

Claims

WE CLAIM:

1. A transgenic insect, or progeny thereof, whose somatic and germ cells contain recombinant nucleic acid encoding A. two or more of the glycosylation enzymes: a) beta-l,2-N-acetylglucosaminyltransferase I, b) beta-l,2-N-acetylglucosaminyltransferase II, c) a βl, 4-galactosyltransferase, and/or d) a sialyltransferase, or B. one or more of the glycosylation enzymes: a) beta-l,2-N-acetylglucosaminyltransferase I, b) beta-l,2-N-acetylglucosaminyltransferase II, and/or d) a sialyltransferase, wherein each recombinant nucleic acid encoding a glycosylation enzyme is integrated in the insect genome, and is present in one or more copies, wherein each recombinant nucleic acid encoding a glycosylation enzyme is operably linked to an expression control sequence, and wherein expression of said glycosylation enzymes allows for production of a partially or completely mammalianized glycosylated protein in the insect.

2. The transgenic insect of claim 1, wherein enzyme c) is a β4-galactosyltransferase; and/or enzyme d) is an alpha 2,6-sialyltransferase and/or an alpha 2,3-sialyltransferase.

3. The fransgenic insect of claim 1, wherein the glycosylation genes are expressed in catalytic amounts.

4. The fransgenic insect of claim 1, whose somatic and germ cells comprise genomically integrated recombinant nucleic acid encoding enzyme a); enzyme a) and enzyme b); enzyme a), enzyme b) and enzyme c); or enzyme a), enzyme b), enzyme c) and enzyme d).

5. The transgenic insect of claim 1, whose somatic and germ cells contain at least one genomically integrated nucleic acid encoding enzyme a), enzyme b), enzyme c), and enzyme d).

6. The transgenic insect of claim 1, whose somatic and germ cells further comprise recombinant nucleic acid encoding one or more of the following glycosylation enzymes: e) a sialic acid synthase and/or f) CMP-sialic acid synthetase, wherein each recombinant nucleic acid encoding a glycosylation enzyme is integrated in the insect genome, and is present in one or more copies, and wherein each recombinant nucleic acid encoding a glycosylation enzyme is operably linked to an expression control sequence.

7. The transgenic insect of claim 6, wherein the somatic and germ cells comprise recombinant nucleic acid encoding enzyme e) and enzyme f).

8. The transgenic insect of claim 1 or claim 6, whose somatic and germ cells further comprise recombinant nucleic acid encoding one or more of the following auxiliary glycosylation proteins: g) UDP-N-acetylglucosamine 2 epimerase/N-acetylmannosamine kinase; h) beta-l,4-N-acetylglucosaminyltransferase III; i) beta-l,4-N-acetylglucosaminyltransferase IV; j) beta-l,6-N-acetylglucosaminyltransferase V; k) beta-l,4-N-acetylglucosaminyltransferase VI; 1) a beta 1,4-N-acetylgalactosaminyltransferase; m) CMP-sialic acid transporter; n) UDP-galactose transporter, wherein each recombinant nucleic acid encoding an auxiliary glycosylation protein is genomically integrated in the insect genome, and is present in one or more copies, and wherein each recombinant nucleic acid is operably linked to an expression control sequence.

9. The transgenic insect of claim 1 or claim 6, which is a lepidoptera, coleoptera, hymenoptera or diptera.

10. The transgenic insect of claim 1 or claim 6, which is a Lepidoptera.

11. The transgenic insect of claim 10, which is T ni.

12. The transgenic insect of claim 1 or claim 6, which is an egg cell, a larva, a pupa, or an adult insect.

13. The transgenic insect of claim 1 or claim 6, wherein at least one of the mammalianizing glycosylation protein genes is under the confrol of a constitutive promoter.

14. The transgenic insect of claim 13, wherein the constitutive promoter is a polh, plO or Iel baculovirus promoter.

15. The transgenic insect of claim 1 or claim 6, wherein at least one of the mammalianizing glycosylation protein genes is under the control of an inducible expression confrol element.

16. The transgenic insect of claim 15, wherein the inducible expression confrol element comprises a baculovirus-specific late or very late promoter.

17. The transgenic insect of claim 16, wherein the humanizing glycosylation genes are not expressed until the transgenic insect is infected with a baculovirus expressing a heterologous gene of interest.

18. The transgenic insect of claim 15, wherein the inducible expression confrol element comprises an hsp70 promoter.

19. ' The transgenic insect of claim 15, wherein the inducible expression control element comprises a constitutive promoter that is regulated by Tet.

20. The transgenic insect of claim 19, wherein the inducible expression confrol element comprises a Tet-CMV-IE promoter or a Tet-baculovirus Iel promoter.

21. The fransgenic insect of claim 1 or claim 6, which is heterozygous for the sequences encoding the glycosylation enzyme(s).

22. The fransgenic insect of claim 1 or claim 6, which is homozygous for the sequences encoding the glycosylation enzyme(s).

23. An isolated cell, or progeny thereof, of a transgenic insect of claim 1 or claim 6.

24. A fransgenic insect of claim 1 or claim 6, whose somatic and germ cells further comprise genomically integrated recombinant nucleic acid encoding a heterologous polypeptide(s) of interest, which is operably linked to an expression confrol sequence.

25. A method for producing, in an insect larva, a partially or completely mammalianized glycosylated form of a polypeptide of interest that is endogenous to the insect, comprising cultivating a fransgenic insect of claim 1 or claim 6, which is a larva, under conditions effective to produce a mammalianized glycosylated form of said polypeptide of interest.

26. A method for producing, in an insect larva, a partially or completely mammalianized glycosylated recombinant polypeptide of interest, comprising introducing into a fransgenic insect of claim 1 or claim 6, which is a larva, a vector comprising nucleic acid encoding said recombinant polypeptide, operably linked to an expression control sequence.

27. The method of claim 26, wherein the recombinant polypeptide is endogenous to the insect.

28. The method of claim 26, wherein the recombinant polypeptide is heterologous to the insect.

29. The method of claim 26, wherein the vector is a baculovirus vector.

30. The method of claim 26, wherein the vector is a transposon-based vector.

31. The method of claim 26, wherein the vector is a piggyBac vector.

32. The method of claim 26, wherein the molar ratio of the polypeptide of interest to the glycosylating enzyme(s) is greaterthan about 100:1.

33. The method of claim 26, wherein the vector further comprises a detectable marker protein, operably linked to an expression control sequence.

34. The method of claim 26, further comprising culturing the infected insect under conditions effective for expressing the heterologous protein and glycosylating it in a mammalianized fashion, and harvesting the mammalianized glycosylated heterologous polypeptide.

35. The method of claim 26, wherein the polypeptide of interest is an antibody, cytokine, blood clotting factor, anticoagulant, viral antigen, enzyme, receptor, vaccine, hormone, or viral insecticide.

36. The method of claim 26, wherein the glycosylation enzymes are expressed at a low level before the vector encoding the polypeptide of interest is introduced into the insect.

37. The method of claim 26, wherein the glycosylation enzymes are not expressed until the vector encoding the polypeptide of interest is introduced into the insect.

38. The method of claim 37, wherein the nucleic acids encoding the glycosylation enzyme(s) are under the confrol of late or very late baculovirus promoters, and the polypeptide of interest is in a baculovirus vector, such that the infection of the insect by the baculovirus vector induces expression of the glycosylation enzyme(s).

39. A fransgenic insect of claim 1 or claim 6 which is infected with a vector comprising nucleic acid encoding a heterologous polypeptide of interest, operably linked to an expression confrol sequence.

40. The transgenic insect of claim 39, wherein the vector is a baculovirus vector.

41. The fransgenic insect of claim 39, wherein the vector is a transposon-based vector.

42. The method of claim 39, wherein the vector is a piggyBac vector.

43. A method for producing, in an insect larva, a partially or completely mammalianized glycosylated polypeptide of interest that is heterologous to the insect, comprising cultivating a fransgenic insect of claim 24, which is a larva, under conditions effective to produce a mammalianized glycosylated form of said polypeptide of interest.

44. A method for producing, in an insect larva, a partially or completely mammalianized glycosylated polypeptide of interest that is heterologous to the insect, comprising introducing into an insect larva a construct comprising nucleic acid encoding A. two or more of the glycosylation enzymes: a) beta- 1,2-N-acetylglucosaminylfransf erase I, b) beta- 1,2-N-acetylglucosaminylfransf erase II, c) a βl, 4-galactosyltransferase, or d) a sialyltransferase, or B. one or more of the glycosylation enzymes: a) beta-l,2-N-acetylglucosaminyltransferase I, b) beta-l,2-N-acetylglucosaminyltransferase II, or d) a sialyltransferase, wherein each nucleic acid sequence encoding a glycosylation enzyme is operably linked to an expression confrol sequence, and a construct comprising a nucleic acid encoding the polypeptide of interest, operably linked to an expression confrol sequence, under conditions effective to produce a mammalianized glycosylated from of said polypeptide of interest.

45. The method of claim 44, wherein enzyme c) is a β4-galactosyltransferase; and/or enzyme d) is an alpha 2,6-sialyltransferase and/or an alpha 2,3-sialyltransferase.

46. An insect comprising, in at least some of its cells, A. two or more of the glycosylation enzymes: a) beta-l,2-N-acetylglucosaminyltransferase I, b) beta- 1 ,2-N-acetylglucosaminyltransferase II, c) a βl, 4-galactosyltransferase, or d) a sialyltransferase, or B. one or more of the glycosylation enzymes: a) beta-l,2-N-acetylglucosaminylfransferase I, b) beta-l,2-N-acetylglucosaminyltransferase II, or d) a sialyltransferase, and a heterologous polypeptide of interest, wherein the glycosylation enzymes are effective to glycosylate the heterologous polypeptide of interest in a mammalian-like glycosylation pattern.

47. The insect of claim 46, wherein enzyme c) is a β4-galactosyltransferase; and/or enzyme d) is an alpha 2,6-sialyltransferase and/or an alpha 2,3-sialyltransferase.

48. The insect of claim 46 or 47, at least some of whose cells further comprise effective amounts of e) a sialic acid synthase and/or f) CMP-sialic acid synthetase.

49. The insect of claim 46, 47, or 48, at least some of whose cells further comprise effective amounts of g) UDP-N-acetylglucosamine 2 epimerase/N-acetylmannosamine kinase; h) beta-l,4-N-acetylglucosaminylfransferase III; i) beta-l,4-N-acetylglucosaminylfransferase rV; j) beta-l,6-N-acetylglucosaminylfransferase V; k) beta-l,4-N-acetylglucosaminylfransferase VI; 1) a beta 1,4-N-acetylgalactosaminylfransferase; m) CMP-sialic acid transporter; and/or n) UDP-galactose transporter.

50. An insect comprising in at least some of its cells recombinant nucleic acid encoding a protein of interest operably linked to an expression confrol sequence, and recombinant nucleic acid encoding A. two or more of the glycosylation enzymes: a) beta-l,2-N-acetylglucosaminylfransferase I, b) beta-l,2-N-acetylglucosaminyltransferase II, c) a beta-1, 4-galactosyltransferase, or d) a sialyltransferase, or B. one or more of the glycosylation enzymes: a) beta-l,2-N-acetylglucosaminylfransferase I, b) beta-l,2-N-acetylglucosaminyltransferase II, or d) a sialyltransferase, wherein each recombinant nucleic acid encoding a glycosylation enzyme is operably linked to an expression control sequence, and wherein the insect produces partially or completely mammalianized glycosylated protein of interest.

51. A method for producing, in an insect larva, a partially or completely mammalianized glycosylated polypeptide of interest that is endogenous or heterologous to an insect as described herein, or an insect as described herein, wherein the insect is not Bombyx mori.

52. A method for producing, in an insect larva, a partially or completely mammalianized glycosylated polypeptide of interest that is heterologous to the insect, comprising introducing a vector comprising nucleic acid encoding said heterologous polypeptide, operably linked to an expression confrol sequence, into a transgenic insect larva, or progeny thereof, whose somatic and germ cells contain recombinant nucleic acid encoding one or more of the glycosylation enzymes: a) beta- 1 ,2-N-acetylglucosaminylfransferase I, b) beta-l,2-N-acetylglucosaminylfransferase II, c) a βl, 4-galactosyltransferase, or d) a sialyltransferase, wherein each recombinant nucleic acid encoding a glycosylation enzyme is integrated in the insect genome, and is present in one or more copies, wherein each recombinant nucleic acid encoding a glycosylation enzyme is operably linked to an expression control sequence, wherein expression of said glycosylation enzymes allows for production of a partially or completely mammalianized glycosylated protein in the insect, and wherein, if the insect is B. mori, and the insect contains genomically integrated nucleic acid encoding enzyme c), then the insect also contains genomically integrated nucleic acid encoding at least one of enzymes a), ^"b) or d).

53. A method for producing, in an insect larva, a partially or completely mammalianized glycosylated polypeptide of interest that is heterologous to the insect, comprising introducing a vector comprising nucleic acid encoding said heterologous polypeptide, operably linked to an expression control sequence, into a transgenic insect larva, or progeny thereof, whose somatic and germ cells contain recombinant nucleic acid encoding one or more of the glycosylation enzymes: a) beta-l,2-N-acetylglucosaminyltransferase I, b) beta-l,2-N-acetylglucosaminylfransferase II, c) a βl, 4-galactosyltransferase, or d) a sialyltransferase, wherein each recombinant nucleic acid encoding a glycosylation enzyme is integrated in the insect genome, and is present in one or more copies, wherein each recombinant nucleic acid encoding a glycosylation enzyme is operably linked to an expression confrol sequence, wherein expression of said glycosylation enzymes allows for production of a partially or completely mammalianized glycosylated protein in the insect, and wherein if the insect is B. mori, the glycosylated polypeptide is not expressed specifically in the silk glands.

54. A fransgenic insect, or progeny thereof, whose somatic and germ cells contain recombinant nucleic acid encoding A. two or more of the glycosylation enzymes: a') a beta-l,2-N-acetylglucosaminylfransferase, c) a βl, 4-galactosyltransferase, or d) a sialyltransferase, or B. one or more of the glycosylation enzymes: a') a beta-l,2-N-acetylglucosaminylfransferase, or d) a sialyltransferase, wherein each recombinant nucleic acid encoding a glycosylation enzyme is integrated in the insect genome, and is present in one or more copies, wherein each recombinant nucleic acid encoding a glycosylation enzyme is operably linked to an expression confrol sequence, and wherein expression of said glycosylation enzymes allows for production of a partially or completely mammalianized glycosylated protein in the insect.

55. The method of claim 1, further wherein the expression of endogenous 1,3-fucosyltransferase expression or activity is inhibited.

56. A method comprising: producing a TRANSPILLAR larva expressing a glycosylated protein of interest, and scaling up production of the larva to arrive at sufficient numbers of larva to produce enough of the glycosylated protein for pre-clinical studies, clinical trials, and for commercialization.

57. A method comprising: receiving a request for production of a glycosylated protein, producing a TRANSPILLAR LARVA expressing the glycosylated protein, generating revenue from the TRANSPILLAR larva either by rearing TRANSPILLAR larvae and isolating and selling the glycosylated protein, or by selling TRANSPILLAR eggs or larvae.

58. A library of different types of TRANSPILLAR larvae expressing a variety of different glycosylated proteins.

59. A library of different types of TRANSPILLAR larvae glycosylating proteins in a variety of patterns.

60. A fransgenic insect, or progeny thereof, whose somatic and germ cells contain recombinant nucleic acid encoding: A. two or more of the glycosylation enzymes: a beta- 1 ,2-N-acetylglucosaminyltransferase; a beta- 1 ,4-galactosyltransferase; a sialyltransferase; or B. one or more of the glycosylation enzymes: a beta- 1 ,2-N-acetylglucosaminylfransferase; a sialyltransferase, wherein each recombinant nucleic acid encoding a glycosylation enzyme is integrated in the insect genome, and is present in one or more copies, wherein each recombinant nucleic acid encoding a glycosylation enzyme is operably linked to an expression control sequence, and wherein expression of said glycosylation enzyme(s) (e.g., in a catalytic amount) allows for production of a partially or completely mammalianized glycosylated protein in the insect.

61. The fransgenic insect, or progeny thereof, of claim 60, whose somatic and germ cells contain recombinant nucleic acid encoding A. two or more of the glycosylation enzymes: beta-l,2-N-acetylglucosaminyltransferase II, a βl, 4-galactosyltransferase, an alpha 2,6-sialyltransferase, an alpha 2,3-sialyltransferase, or B. one or more of the glycosylation enzymes: beta- 1 ,2-N-acetylglucosaminyltransferase II, an alpha 2,6-sialyltransferase, an alpha 2,3-sialyltransferase.

62. The transgenic insect, or progeny thereof, of claim 60, wherein if nucleic acid encoding a βl, 4-galactosyltransferase is present, nucleic acid encoding at least one of the enzymes: a beta- 1,2- N-acetylglucosaminylfransferase or a sialyltransferase is also present.

63. The fransgenic insect of claim 61, or progeny thereof, wherein the somatic and germ cells contain recombinant nucleic acid encoding beta-l,2-N-acetylglucosaminylfransferase II.

64. The transgenic insect of claim 61, or progeny thereof, wherein the somatic and germ cells contain recombinant nucleic acid encoding beta-l,2-N-acetylglucosaminyltransferase II and a βl, 4-galactosyltransferase.

65. The transgenic insect of claim 61, or progeny thereof, wherein the somatic and germ cells contain recombinant nucleic acid encoding beta-l,2-N-acetylglucosaminylfransferase II, a βl, 4- galactosyltransferase, and an alpha 2,6-sialyltransferase.

66. The fransgenic insect of claim 61, or progeny thereof, wherein the somatic and germ cells contain recombinant nucleic acid encoding beta- 1 ,2-N-acetylglucosaminylfransferase II, a β 1 , 4- galactosylfransferase, an alpha 2,6-sialyltransferase, and an alpha 2,3-sialyltransferase and, optionally, beta-l,2-N-acetylglucosaminyltransferase I.

67. The fransgenic insect of claim 66, or progeny thereof, wherein the somatic and germ cells further contain recombinant nucleic acid encoding a sialic acid synthase and CMP-sialic acid synthetase.