T 0989/16 (Post translationally modified glycoproteins/LIFE TECHNOLOGIES CORPORATION) 04-12-2020

I. European patent No. 1 991 273, based on European patent application No. 07825657.5, was opposed on the grounds of Articles 100(a), (b) and (c) EPC. An opposition division considered that the main request before it (patent as granted) contravened both Article 83 EPC and Article 56 EPC and took the view that the auxiliary request and the description adapted thereto complied with the requirements of the EPC.

II. The patent proprietor (appellant I), and opponents 1 and 2 (appellants II and III respectively) lodged an appeal against the decision of the opposition division. Appellant II submitted new documents E28 to E30 in support of its case.

III. The patent proprietor (appellant I) replied to appellant's II and III statements of grounds of appeal.

The parties were summoned to oral proceedings.

In response to the board's communication pursuant to Article 17(1) RPBA 2020, appellant I submitted new auxiliary requests 1, 1a, 2 and 2a.

Appellants II and III announced that they would not attend the oral proceedings.

IV. Oral proceedings took place on 4 December 2020, in the absence of appellants II and III.

V. Claims 1, 9 and 10 of the main request (claims as granted) read as follows:

"1. A method for enzymatically labeling a glycoprotein with an azide moiety, wherein the method comprises:

a) contacting the glycoprotein with UDP-GalNAz in the presence of a beta-1,4-galactosyltransferase Y289L mutant enzyme to form an azido modified glycoprotein.

9. An azido modified glycoprotein obtainable by the method of any of claims 1-8.

10. A kit comprising: UDP-GalNAz; a beta-1,4-galactosyltransferase Y289L mutant enzyme; an azide reactive reporter molecule, carrier molecule or solid support."

VI. Claims 1 and 9 of auxiliary request 1 read:

"1. A method for enzymatically labeling a glycoprotein comprising a terminal GlcNAc group with an azide moiety, wherein the method comprises:

a) contacting the glycoprotein with UDP-GalNAz in the presence of a beta-1,4-galactosyltransferase Y289L mutant enzyme to form an azido modified glycoprotein.

9. An azido modified glycoprotein obtainable by the method of Claim 1."

VII. The following documents are cited in this decision:

D1 H.C. Hang and C.R. Bertozzi "The chemistry and

biology of mucin-type O-linked glycosylation."

Bioorganic Medicinal Chemistry 2005 Sep 1;

13(17), pages 5021-34;

D2 H.C. Hang et al. "A metabolic labeling approach

toward proteomic analysis of mucin-type O-linked

glycosylation", PNAS, Dec 2003, 100 (25), pages

14846-14851;

D3 R. Chatterjee et al., "Mutation of the ptsG gene

results in increased production of succinate in

fermentation of glucose by Escherichia coli"

Appl. Environ. Microbiol. 2001; 67(1), p.148-154;

D4 US2005/0130235 (published 16 June 2005);

D16 A. Varki et al., Essentials of Glycobiology,

(1999), Cold Spring Harbor, Chapter 38,

pages 581-598;

E1 Experimental Evidence submitted by opponent 2/

Appellant III;

E4 E. Saxon et al. "Investigating cellular

metabolism of synthetic azidosugars with the

Staudinger ligation." J Am Chem Soc. 2002 Dec

18; 124(50), pages 14893-902.

VIII. The submissions made by appellant I, insofar as relevant to the present decision, may be summarized as follows:

Main request

Sufficiency of disclosure (Article 100 (b) EPC)

The skilled person would have interpreted claim 1 when reading the patent application, based on its technical knowledge with respect to the enzyme specificity, as excluding methods relating to glycoproteins that could not be labelled by an enzyme-based reaction with UDP-GalNAz. These glycoproteins, including antibodies and alpha-crystallin, lacked an enzyme accessible terminal GlcNAc, especially an O-GlcNAc and/or at least one terminal N-acetylglucosamine residue (see paragraphs [0108], [0112], [0280], Figures 6A-B, Figure 11B, Figure 17, Examples 18, 19 and 26). Document D16 - a general reference on glycobiology - taught that treatment of glycoproteins with endoglycosidases, e.g. endoglycosidase H, resulted in a mobility change on one or more of the bands on the SDS PAGE gel, indicating the presence of N-glycans (see page 584, Figure 38.1). Endo-H enzyme was known to specifically cleave the chitobiose core of N-linked glycans, exposing an N-linked terminal GlcNAc group. The "Endo-H (endo-beta-N-acetylglucosaminidase H) enzyme could be used to generate a truncated chain which terminates with one N-acetylglucosamine residue (Figure 11B)", which inferred that the "terminal GlcNAc group" was either O- or N-linked to the remainder of the glycoprotein (see paragraph [0112]). Example 26 established that the Endo Hf deglycosylated goat antibody had a terminal N-acetylglucosamine residue, which was then labelled with a beta-1,4-galactosyltransferase Y289L mutant and UDP-GalNAz (see paragraph [0280], Figure 17 A, lane 3, and paragraph [0112]).

The term "azide moiety" specified that a glycoprotein was contacted with a UDP-GalNAz in the presence of the claimed mutant enzyme to form an azido modified glycoprotein. The azide moiety was accordingly clearly defined.

Moreover, independent claim 9 had to be interpreted as a product formed in accordance with the method of claim 1 based on the same original starting glycoproteins, whereas independent claim 10 was defined by its components: UDP-GalNAz and beta-1,4-galactosy1transferase Y289L mutant enzyme, which, for consistency reasons,

had to be interpreted as in claim 1.

Auxiliary request 1

Amendments (Article 123(2) EPC)

The features of amended claim 1 could be found in paragraph [0009] of the patent application, except that the enzyme is "a modified Gal T enzyme" rather than the claimed "beta-1,4-galactosyltransferase Y289L mutant enzyme". The missing feature was disclosed in paragraphs [0111] and [0115] of the patent application, where it was stated that "... antibodies can be enzymatically labeled with UDP-GalNAz using GalT via reaction of the terminal O-GlcNAc molecules present on antibody carbohydrates" and "... if no O-GlcNAc sugars are present on an antibody, then the use of the Endo-H (endo-beta-N-acetylglucosaminidase H) enzyme may be used to generate a truncated chain, which terminates with one N-acetylglucosamine residue (Figure 11B)". The enzymatic labeling of goat IgG antibody with UDP-GalNAz in the presence of a GalT1 Y289L mutant, which "can catalyze the transfer of galactose from uridine diphosphate-GalNAz (UDP-GalNAz) to terminal GlcNAc groups" was expressly disclosed in paragraph [0111] and Example 26 of the patent application (see Figure 17).

The methods could be used to obtain an azido modified glycoprotein (see paragraph [0009], lines 1 to 4), while paragraph [00231] referred to a kit containing "a GalT enzyme", a term which included specific GalT enzymes, such as beta-1,4-galactosyltransferase (GalT) Y289L (see paragraph [0111] of the patent application).

Hence claims 1, 9 and 10 were based on the disclosure of the patent application.

Clarity (Article 84 EPC)

The expression "terminal GlcNAc group" designated a "terminal" location of a GlcNAc group covalently attached at a terminal glycan position of another glycan residue or of an amino acid residue of the glycoprotein, which could be O-linked or N-linked, but lacking any further residue attached to it. This expression reliably establishes what falls under the scope of claims 1 and 9. This interpretation was confirmed by paragraph [0112] of the patent.

Sufficiency of disclosure (Article 83 EPC)

The patent disclosed alpha-crystallin labelled at its terminal O-GlcNAc with an azide from UDP-GalNAz, using a modified b-GalT1 enzyme in accordance with the method of claim 1 (see example 18), and the generation of a terminal N-linked GlcNAc with Endo Hf, followed by azide labeling of the glycoprotein by contacting the glycoprotein with UDP-GalNAz in the presence of a GalT1 Y289L mutant (see Example 26). These examples used the Y289L mutant b-GalT1, since it was the only tangible modified enzyme of the patent (see paragraphs [0108] and [0280]), and two glycoproteins with different types of terminal GlcNAc (see paragraphs [0108], [0111] and [0112]).

Inventive step Article 56 EPC

Document D4 was the closest prior art. It concerned "methods for rapid and sensitive detection of post-translationally modified proteins ... that were at the limits of detection using traditional models" (D4, abstract). The aim of document D4 was achieved by using a glycosyl donor comprising an unnatural ketone functionality. Even if GalT having a Y289L mutation could tolerate substrates with minor substitutions at the C-2 position, specifically 2-deoxy, 2-amino, and 2-N-acetyl substituents (D4, paragraph [0131]), these substituents differed significantly from the C-2 substituent both in size and by means of its polar azide group. Due to this difference, it was unknown whether GalT or a mutant thereof could use the UDP-GalNAz substrate with a reasonable expectation of success. Document D15 confirmed that the polarity of the substrate played an important role in the progress of the enzymatic reaction. Thus, the polar azide group of the UDP-GalNAz interfered with the network of polar interactions in the substrate binding pocket of GalT and thereby destroyed its ability to transfer the GalNAz group of UDP-GalNAz to an acceptor glycoprotein.

The subject-matter of claim 1 was for this reasons inventive over the content of document D4 alone.

Document E4 related to the metabolic incorporation of sugar analogues, including Ac4ManNAz and Ac4GlcNAz (see page 14897, Figures 3 and 5)

Thus, documents D4 and E4 failed to disclose the use of UDP-GalNAz, much less the use of UDP-GalNAz in conjunction with Y289L GalT to form an azido labelled glycoprotein.

Document D1 discussed mucin-type O-linked glycosylation initiated by the polypeptide N-acetyl-alpha-galactosaminyltransferases (ppGalNAcTs) (abstract, third sentence). ppGalNAcTs are not beta-1,4-galactosyltransferases. Documents D2 and D3 described the transfer of GalNAz with a ppGalNAcT using a UDP-GalNAz (see document D2, Figure 1; document D3, page 6, column 1, second paragraph, line 5).

Based on the enzyme specificity used in documents D1 to D3, the skilled person would not have considered combining any of them with document D4. But even if the skilled person considered such a combination possible, it was still unknown whether UDP-GalNAz represented a valid substrate for the mutant Y289L enzyme of document D4 or not. In any event, none of the R groups of the modified UDP-Gal compounds were attached to the glycosyl group via a nitrogen atom. The UDP-GalNAz of claim 1 was not covered by the Genus A derivatives of document D4.

Thus, applying the problem-solution approach, the skilled person would not have considered any of these citations in combination with D4.

This reasoning applied mutatis mutandis to product-by-process claim 9 and kit claim 10. All of the claims are therefore inventive.

IX. The submissions made by appellant II, insofar as relevant to the present decision, may be summarized as follows:

Main request

Article 100(b) EPC

Claim 1 extended to the labelling of glycoproteins without a specific terminal glycosyl residue, such as O-GlcNAc or N-GlcNAc. However, only the O-GlcNAc terminal residue and the GlcNAc terminal residue resulting from an Endo H or Endo Hf enzymatic glycosyl chain truncation on glycoprotein were a substrate for the claimed method (see patent paragraphs [0108], [0112], [0280]). There was no support in the patent for other glycoproteins, known or yet to be identified, which could be used in the method of claim 1. Since the term "glycoprotein" had a recognized meaning in the prior art, it was inappropriate to apply a narrower interpretation than the definition given to this term in the prior art (see paragraphs [0085], [0108], Figures 6A-B, Figure 11b, examples 18, 19 and 26). Thus, the subject matter of claim 1 and its dependent claims was insufficiently disclosed.

The method of claim 2 was directed to a method of forming a glycoprotein conjugate. It defined the method of claim 1 as its first step, resulting in a glycoprotein with an azide functional group, which was then involved in a second conjugation step resulting in a product having consumed the azide functional group. Thus, the subject-matter of claim 9, directed at an azido modified glycoprotein inter alia obtainable specifically by a method of claim 2 was not enabled.

Thus, the set of claims failed to meet the requirements of Article 83 EPC.

Auxiliary request 1

Amendments (Article 123(2) EPC)

Claim 1 was amended to a method of enzymatically labeling a "glycoprotein comprising a terminal GlcNAc group". The amendment contravened the requirements of Article 123(2) EPC as there was no basis in the patent application for using glycoproteins with a terminal GlcNAc other than O-GlcNAc groups. The sole explicit disclosure of the term "terminal GlcNAc", though embedded in an otherwise erroneous technical statement, was found in the patent application in paragraph [0111]. First, even if a beta-1,4-galactosyltransferase (GalT) enzyme could be used for this purpose, there was no direct and unambiguous disclosure to use a mutant enzyme, let alone a GalT Y289L mutant enzyme. Second, even if the use of a Y289L mutant of GalT on substrates (plural) was mentioned, it was nowhere specifically combined with a UDP-GalNAz substrate. Third, Figure 6 showed a geometric placeholder to represent the entire molecule except for the azide group of UDP-GalNAz. Fourth, Figure 11A depicted an erroneous side chain: -(C=O)-CH2-N3 instead of: -N-(C=O)-CH2-N3 for UDP-GalNAz, but there was no reason to consider Figure 11 rather than the corresponding written disclosure to be erroneous. Thus, both Figures could not serve as an adequate basis for amended claim 1 either.

Finally, paragraph [0115] of the patent application referred to antibodies, or otherwise to a "reaction of the terminal OGlcNAc molecules present on antibody carbohydrates". "If no OGlcNAc sugars are present on an antibody, the use of the Endo-H (...) enzyme will be used to generate a truncated chain which terminates with one O-linked N-acetylglucosamine residue (Figure 11B)" (see Example 26). Thus, paragraph [0115] only provided a basis for glycoproteins comprising O-GlcNAc groups but not for glycoproteins with other GlcNAc groups (see also paragraphs [0114], [0003] and [0004]). The same conclusion was applicable to independent claim 9.

Clarity (Article 84 EPC)

The patent provided no specific definition for the term "terminal GlcNAc". In consequence, the GlcNAc group was either located somewhere in a terminal part of a complex glycan motif on a glycoprotein or it represented the terminus itself, which corresponded to its ordinary meaning. In some examples of the patent it defined the terminus of a truncated chain, illustrated as terminal O-GlcNAc, but this was not its sole logical interpretation (see paragraph [0112] and Figure 11, Figures 6A-B in paragraph [0108]). Thus, there was no reason to limit the scope of claim 1 to glycoproteins comprising specific positions on glycan allowing a transfer of the substituted monosaccharide UDP-GalNAz onto that glycoprotein by galactosyltransferases only, let alone by a Y289L mutated form. The term "terminal GlcNAc" introduced into claim 1 contravened Article 84 EPC.

The same conclusion applied to the dependent claims and to claim 9.

Article 83 EPC

Claim 1 referred to a terminal GlcNAc, encompassing O-GlcNAc and other terminal GlcNAc, such as N-GlcNAc, as well as simple or complex branched glyco-chains, wherein GlcNAc might be only one of several end groups. The patent provided no clear and unambiguous teaching regarding the nature of terminal GlcNAc groups that terminal N-GlcNAc and branched terminal GlcNAc were also suitable substrates for the claimed method. The labelling of a goat antibody, known to be O- and N-glycosylated, which was deglycosylated using Endo Hf enzyme was shown to occur on terminal O-GlcNAc residues (see Figure11, example 26). This example could not prove that the labelling of a terminal N-GlcNAc comprising glycoprotein also took place. Thus, the subject matter of claim 1 was not enabled across the entire scope of the claim 1. The same conclusion applied mutatis mutandis to claims referring to it and dependent thereon.

Inventive step Article 56 EPC

Document D4 was the closest prior art. It described a method for the labeling of post-translationally modified proteins (e.g. paragraphs [0004], [0008]) with a chemical handle (see paragraph [0042]), wherein the method comprises contacting the post-translationally modified protein with a UDP-sugar derivative (see paragraph [0038]) in the presence of mutant Y289L beta-l,4-galactosyltransferase (see paragraph [0097]) to form a modified glycoprotein (e.g. Figure 9, paragraphs [0183]-[0187]). The post-translationally modified protein was a glycoprotein (see paragraph [0093]), "having a pendant moiety", such as a glycosyl group or GlcNAc (paragraphs [0010]-[0013], [0027]-[0029]; claim 19). O-GlcNAc-glycosylated proteins were preferred embodiments and were used with Y289L mutant GalT and GalNAc sugar substrates, such as labelling agent 1, falling under the formula of genus A or A' (e.g. paragraphs [0094], [0125] to [0131], Figure 1B, [0038]). The R substituent was selected from the group consisting of straight chain or branched C1-C12 carbon chain bearing a carbonyl group, azide group, straight chain or branched C1-C2 carbon chain bearing an azide group, straight chain or branched C1-C12 carbon chain bearing an alkyne, and straight chain or branched C1-C12 carbon chain bearing an alkene (see paragraph [0039]). The chemical handle attached to the UDP-sugar derivative was also described as an azide moiety (see paragraphs [0020], [0105]-[0107]) defined in paragraph [0104]. It explicitly disclosed the use of UDP-galactose derivatives, wherein the substituent R carrying the chemical handle appended at the C-2 position of the sugar moiety was an azide group or a linear or branched alkyl substituent bearing an azide group (see paragraph [0136]) inter alia in conjunction with a mutant GalT (see paragraph [0137]). It was immaterial that "the main focus appears to lie on ketone substituents".

The sole difference between the content of document D4 and the patent was the use of UDP-GalNAz. The sugar derivative of document D4 carried a chemical handle appended at the C-2 position of the galactose moiety whose purpose was to allow further modifications e.g. via Staudinger reaction (see paragraph [0110], second sentence). The use of GalNAz compared to the substrates disclosed in document D4 was not disclosed to have any advantage.

Thus, starting from document D4, the problem to be solved was to provide an alternative azido sugar substrate.

Document D4 disclosed that the engineered enzyme and the corresponding substrate "capitalize on the substrate tolerance of GalT, which allows for chemoselective installation of a non-natural functionality, such as a ketone chemical handle, to O-GlcNAc pendant moiety on modified proteins (Figure 1B)" (see paragraph [0130]) extending to "... unnatural substrates containing minor substitutions at the C-2 position, including 2-deoxy, 2-amino and 2N-acetyl substituents". The Y289L mutation enlarged the binding pocket of GalT and enhanced the catalytic activity toward GalNAc substrates without compromising specificity. Thus the skilled person was motivated to consider alternative substitutions at the C-2 position and was aware of the favourable properties of UDP-GalNAc sugar substrates. The chemical handle could be an azido group capable of reacting in a Staudinger reaction, which was referring explicitly to document E4 (see paragraph [0107]) and thus incorporating its teaching (see decision T 153/85). The use of a substituent with a C1-C12 carbon chain bearing an azide group was not expected to encounter steric hindrance (see document D4, paragraph [0136]).

Document E4 described azido modified N-acetyl analogues ManNAz and GlcNAz with an azido handle capable of reacting in a Staudinger reaction (see abstract).

Documents D4 and E4 prompted the skilled person to employ GalNAz instead of Gal-X-NAz and did not prevent him from using UPD-GalNAz in combination with the mutant Y289L GalT, since its binding pocket was enlarged, which enhanced its catalytic activity toward GalNAc substrates (see paragraphs [0130], [0131]; document D10).

Since GalNAz can be produced from GalNAc with an azide moiety attached at the N-acetyl group using a method known in art and the advantages of azide groups were known, the choice of an azide chemical handle cannot justify an inventive step (see documents D1 to D3, especially D3, col.1, second paragraph).

Since the complexity of the technical problem to be solved was low and the labeling procedure completed in less than 24 hours, there was no technical prejudice which would have jeopardized the skilled person's expectation of success of a 'try and see' approach.

The subject matter of claim 1 was accordingly not inventive in the light of document D4 in combination with documents D1 to D3 or E4, explicitly referred to in document D4 in paragraph [0107]. For the same reasons dependent claims 2-8 lacked inventive step as well.

The submissions made by appellant III, insofar as relevant to the present decision, may be summarized as follows:

Main request

Sufficiency of disclosure (Article 100(b) EPC)

The "modified b-GalT1 enzyme" used in example 18 could not only be reduced to the "beta-1,4-galactosyltransferase Y289L mutant enzyme". The term encompassed a large number of enzymes characterized by their modifications. There was no reason to assume that the Y289L mutant, mentioned also elsewhere in the description and in the context of another embodiment, was the enzyme used in Example 18. Example 26 related to an experiment where deglycosylated goat IgG antibodies, whose glycosyl chains were first trimmed by Endo Hf to terminal sugar moieties, were then enzymatically labelled with a GalNAz moiety by a beta-1,4 galactosyltransferase Y289L mutant enzyme to form an azido modified glycoprotein. Since the method of claim 1 defined neither the use of a deglycosylated glycoprotein nor the trimming step of a glycoprotein, a skilled person, applying the method of claim 1 to an antibody, would not know how to label the glycoprotein with a GalNAz moiety, i.e whether a prior deglycosylation or the introduction of a terminal GlcNAc was necessary or not. Example 26 was not sufficient to enable the skilled person to carry out the method of claim 1 over its entire scope without undue burden.

Examples 18 to 22 referred to a glycoprotein labelled on a terminal O-GlcNAc group, but provided no support on how the remaining embodiments falling also under claim 1, e.g. glycoproteins having a high-mannose comprising two innermost GlcNAc moieties (see document E1, RNAse B) or a terminal sugar moiety other than a terminal O-GlcNAc group, could be amenable to the claimed method without undue burden. The legend of Figure 11 related to an antibody and required the presence of terminal O-GlcNAc molecules on the antibody for its subsequent enzymatic labelling with UDP-GalNAz (see patent paragraph [0026]). Absent a disclosure how a non-terminal GlcNAc group was labelled and of the experimental conditions for achieving this labelling, the skilled person was not in a position to reproduce the claimed method over the entire breadth of the claim.

The Examples of the patent related to reactions where O-GlcNAc, but not "GlcNAc" or "N-GlcNAc" groups, were modified (see Figure 17; Figure 11 A or B). Except for glycoproteins comprising a O-GlcNAc, the skilled person had to identify first which sugar moieties allowed azido groups to be transferred and which experimental conditions were needed for glycoproteins to be modified. Document El supported this view. Hence, the skilled person could not carry out the invention of claim 1 over its entire scope pursuant to Article 83 EPC.

Claim 2 referred to "the method of claim 1 that is a method of forming glycoprotein conjugate", wherein the azido modified glycoprotein was contacted with a reporter molecule, carrier molecule or solid support comprising an azide reactive moiety to form a conjugate.

The reaction between an azido modified glycoprotein and an azide reactive moiety resulted in consumption of the azido group.

Claim 9 referred to an azido modified glycoprotein obtainable inter alia by the method of claim 2. As the method of claim 2 consumed the azido reactive group on the glycoprotein, claim 9 could not be implemented without carrying out an additional azido modification step.

Thus, the subject matter of claim 1 was not enabled across its entire scope, contrary to the requirements of Article 83 EPC. The same conclusion applied to claims 2-10.

Auxiliary request 1

Amendments (Article 123(2) EPC)

Paragraph [0004] of the background art of the patent application could not serve as a basis for the claimed invention.

The patent application stated erroneously that "beta-1,4-galactosyltransferase (GalT) is an enzyme that can catalyze the transfer of galactose [sic!] from uridine diphosphate-GalNAz (UDP-GalNAz) to terminal GlcNAc groups" (see paragraph [0111]). This is technically impossible. Thus, even if a beta-1,4-galactosyltransferase (GalT) enzyme could be used for this purpose, there was no direct and unambiguous disclosure to use a mutant enzyme, let alone a GalT Y289L mutant enzyme. Even if the use of a Y289L mutant of GalT on substrates (plural) was mentioned, it was not specifically combined with a UDP-GalNAz substrate. Neither example 18 and Figure 6 nor example 26 disclosed a method combining the specific features of claim 1.

In the method for enzymatically labeling a glycoprotein "the appropriate enzyme" is in one aspect a "Gal T enzyme" and in another aspect "a modified enzyme" (see paragraph [0009]). There was no direct and unambiguous basis for a modified enzyme being a mutant enzyme, let alone a Y289L mutant GalT enzyme. There was no basis in claim 1 of the patent application for a generic method for enzymatically labeling a glycoprotein with an azide moiety comprising step a) which is lacking its step b).

The subject-matter of claim 9 did not have a basis in the patent application, as no generic disclosure could be identified in the patent application for an "azido-modified glycoprotein" obtainable by the method according to claim 1. This was exacerbated by the fact that the method of claim 1 used the term "comprising", allowing the product of step a) to be modified by any additional process step, resulting in any potential azido modified glycoprotein.

The kit of claim 10 comprised added matter as it comprised a beta-1,4-galactosyltransferase Y289L mutant enzyme, whereas the kit of claim 40 of the patent application comprised "a GalT enzyme" (see claim 40; paragraph [00231]).

Hence the subject-matter of claims 1, 9 and 10 violated the provisions of Article 123(2) EPC.

Clarity (Article 84 EPC)

Decision G 3/14 stipulated that amended claims had to comply with the requirements of Article 84 EPC, if the introduced amendments resulted in a lack of clarity of the claims.

Claim 1 was amended to "[a glycoprotein] comprising a terminal GlcNAc group" only. However, the term "terminal" was highly ambiguous. It might either refer to the N-terminus or the C-terminus of the polypeptide chain of the glycoprotein or relate to the specific sugar moiety at the terminal position in a (branched or straight) chain of any polymer associated with the glycoprotein, e.g. a carbohydrate chain. It was equally unclear, which "terminus" of the glycoprotein the "terminal GlcNAc group" had to occupy, i.e. on a glycoside or a protein moiety and whether one single GlcNAc monomer attached to a side chain of the glycoprotein qualified as "terminal GlcNAc group" or not. It was furthermore unclear whether the "terminal GlcNAc group" of a glycoprotein had to be present at the outset of the method or whether it could be introduced by process steps not explicitly mentioned in claim 1. A skilled person would not interpret a terminal single monomeric GlcNAc group, having a relatively small size, buried in, or exposed on a larger glycoprotein at the terminus of a chain of units or on an amino acid residue on a glycoprotein as a "terminal GlcNAc group". Since the term "terminal" was assigned multiple definitions, it was unclear whether a single monomeric unit on a protein qualified as "terminal" (see Example 26 of the patent). Thus, claim 1 and all the claims dependent thereon contravened the requirements of Article 84 EPC.

Sufficiency of disclosure (Article 83 EPC)

The objections raised against the main request under Article 83 EPC were applicable mutatis mutandis to this request.

Inventive step Article 56 EPC

Document D4 was the closest prior art. It related to methods for labeling and detecting glycosylated proteins comprising a GlcNAc group (see paragraph [0008]). The Y289L mutant enzyme was used to transfer galactose or its analogue to terminal GlcNAc groups (see paragraph [0097]) and taught that "the single Y289L mutation enlarged the binding pocket of GalT and enhanced the catalytic activity toward GalNAc substrates without compromising specificity" (see paragraph [0131]). It provided specific embodiments of substrates for Y289L mutant enzyme, such as UDP analogues according to "Genus A" or "Genus A'" (see paragraphs [0135] to [0138]), wherein the C2 substituent was - amongst others - a C2 carbon chain bearing an azide group. Thus, document D4 disclosed generically substrates including UDP-GalNAz and an UDP-Gal analogue with a ketone chemical handle (see formula 1, paragraph [0133]).

The only difference between the specific embodiment disclosed in document D4 and the subject-matter of claim 1 was the "azido modified N-acetyl group" at position C2 instead of the "ketone chemical handle". The choice of an azide handle (instead of a ketone handle), both representing well known chemical handles in sugar chemistry, was merely an arbitrary alternative, as there is no unexpected effect connected to that structural substitution.

The objective technical problem was therefore the provision of an alternative UDP-Gal analogue.

It was common technical knowledge that in addition to its galactosyltransferase activity, the mutant enzyme exhibited GalNAc transferase activity. Changes in enzymatic activity and substrate specificity of the Y289L mutant enzyme accepting other Gal C2 analogues was known (see document D4, paragraph [0131]): "Importantly, the single Y289L mutation enlarges the binding pocket of GalT and enhances the catalytic activity toward GalNAc substrates without compromising specificity."

The plural of the expression "GalNAc substrates" in the above sentence, provided a further hint for the skilled person that GalNAc itself as well as other structurally similar compounds could be used, all of which acting as substrates for that mutant enzyme.

There was a clear pointer in document D4 to use UDP-Gal analogues to replace UDP-Gal ketone and for the use of GalNAc transferases. The GalNAc transferases and the analogue UDP-GalNAz as substrate for GalNAc transferases were reported in documents D1, D2 and D3 (see D3, page 6, right column, first paragraph; Figure 2; D1, page 5026, right column, last sentence; D2, page 14850, left column, penultimate paragraph). The skilled person would therefore have selected the UDP-GalNAz as inevitable "chemical handle" candidate analogue substrate with a reasonable expectation of success (cf. decision T 249/88) and was aware of all the advantages of azide over ketone as a chemical handle.

The subject-matter of claim 1 did not involve an inventive step, as (i) its underlying object was not solved over the entire scope, (ii) the subject-matter claimed did not represent any advantageous effects and (iii) its subject-matter was most obviously suggested by the prior art.

In view of the above arguments, the azido modified glycoprotein obtainable by the method of claim 1 lacked an inventive step as well.

Finally, both the Y289L mutant enzyme and the use of UDP-GalNAz were known in the art, rendering their use in a kit together with an azide reactive reporter molecule, carrier molecule or solid support obvious. Thus, claims 1, 9 and 10 lack an inventive step.

X. Appellant I (patent proprietor) requested, as its main request, that the decision under appeal be set aside and the patent be maintained as granted, or alternatively, that appellant II's and III's appeals be dismissed, or that the patent be maintained on the basis of any of auxiliary requests 1a, 2 and 2a, filed under cover of letters dated 4 March 2020 and 30 September 2016 respectively.

XI. Appellant II requested that the decision under appeal be set aside and the patent be revoked.

XII. Appellant III requested that the decision under appeal be set aside and the patent be revoked as well as that documents E23 to E27 be admitted into the appeal proceedings.