T 0938/08 (Subtilisin/NOVOZYMES) 15-04-2010

Summary of Facts and Submissions

I. The patentee (appellant) lodged an appeal against the decision of the opposition division dated 14 March 2008, whereby European patent 0 945 502 was revoked. The patent had been granted on European patent application No. 99 102 452.2 entitled "A mutated subtilisin protease" which was filed as a divisional application to the parent application 90 910 604.9 published as the international publication WO 91/00345.

II. The patent had been opposed by one opponent. The grounds for opposition relied on were lack of inventive step (Article 100(a) EPC), insufficiency of disclosure (Article 100(b) EPC) and presence of added matter (Article 100(c) EPC).

III. In the decision under appeal, the opposition division found that the main request (claims as granted) and the sole auxiliary request then on file (filed as auxiliary request 2 on 12 December 2007) lacked an inventive step (Article 56 EPC).

IV. The statement of grounds of appeal was filed on 22 July 2008. It was accompanied by two auxiliary requests (1 and 2), of which auxiliary request 2 corresponded to the auxiliary request refused by the opposition division. 16 documents were attached thereto.

V. The opponent (respondent) replied thereto on 16 December 2008. It argued that the three requests on file, i.e. the main request (claims as granted) and the two auxiliary requests filed with letter of 22 July 2008, did not comply with the requirements of Articles 56 and 83 EPC. Three further new documents were attached thereto.

VI. On 21 September 2009, the Board issued a communication pursuant to Article 15(1) of the Rules of Procedure of the Boards of Appeal, in which a provisional and non-binding opinion on issues to be discussed at the oral proceedings was expressed.

VII. In reply to the board's communication, the appellant with a letter dated 15 March 2010 filed further submissions together with five additional sets of claims as auxiliary requests 1 and 4 to 7, and renumbered the previous auxiliary requests 1 and 2 of 22 July 2008 as its auxiliary requests 2 and 3, respectively. 14 new documents were enclosed with the letter.

VIII. Then, on 23 March 2010, the appellant filed a further document to confirm the exact date of publication of one of the documents filed on 15 March 2010.

IX. Oral proceedings took place on 15 April 2010, at which the appellant withdrew its auxiliary requests 1 to 3 and 5 to 7, and adopted as its auxiliary request 1 the set of claims according to its previous auxiliary request 4.

X. Claim 1 of the main request (claims as granted) read:

"1. A method of preparing a mutated subtilisin protease for use in a detergent composition, which comprises changing the net electrostatic charge of the protease in comparison to the parent protease at the same pH, such that, in said protease there are, relative to said parent protease, fewer or more positively-charged amino acid residue(s) and/or more or fewer negatively-charged amino acid residue(s), wherein the parent protease is selected from subtilisin BPN', subtilisin amylosacchariticus, subtilisin 168, subtilisin mesentericopeptidase, subtilisin Carlsberg, subtilisin DY, subtilisin 309, subtilisin 147, thermitase, aqualysin, Bacillus PB92 protease, proteinase K, Protease TW7, and Protease TW3 characterised in that one mutation is effected at a position corresponding to position 252 in subtilisin BPN' by substitution, whereby said mutated subtilisin protease has an isoelectric point (pIo) lower or higher than that of said parent protease and the pIo of the mutated subtilisin protease is closer to the pH of the wash liquor formed by the detergent composition than is the pIo of the parent protease, such that the mutated subtilisin protease exhibits improved wash performance relative to the parent protease in said wash liquor."

(underlining added by the board)

XI. The set of claims according to the auxiliary request 1 consisted of 6 claims which differed from the claims as granted only in that the expression "or higher" in claim 1 (see the underlined expression in claim 1 as granted in section X supra) had been deleted.

Claims 2 to 6 were dependent on claim 1 and directed to particular embodiments thereof.

XII. The following documents are referred to in the present decision:

(D1) A. J. Russell and A. R. Fersht, Nature, Vol. 328, 6 August 1987, pages 496 to 500

(D13) Report submitted by the appellant during the examination proceedings together with its letter of 10 October 2003

(D43) Alessandra Bossi et al., Electrophoresis, Vol. 15, 1994, pages 1535 to 1540

(D44) B. Bjellqvist et al., Electrophoresis, Vol. 15, 1994, pages 529 to 539

(D46) A.W. Kenchington and A.G. Ward, Biochem. J., Vol. 58, No. 2, 1954, pages 202 to 207

XIII. The submissions made by the appellant (patentee), insofar as they are relevant to the present decision, may be summarised as follows:

Main request

Inventive step (Article 56 EPC)

The data provided in the application as filed showed that, in almost all cases, mutations to parent subtilisin enzymes that shifted the pIo resulted in enhanced wash performance in a detergent solution where the shifted pIo was closer to the pH of the detergent solution, and the converse where the shifted pIo was further away from the pH of the detergent solution. The patentee had found that the relationship between pIo and the operational pH was an important contributing factor to optimising wash performance.

Based on the data in Table III alone, 21 different mutants having mutations at one or more of eleven different positions, that caused the pIo of the protein to be shifted compared to the wild-type protein, had all been demonstrated to follow the trend underlying the present invention by showing an associated shift in performance.

It was difficult to draw any clear conclusions from the data in Table XI, due to the small sample size. Nevertheless, the results in Table XI were almost consistent with the general principle underlying the invention, particularly when taking the pH drift towards acidity into account for the lower pH values (where the initial pH was close to the pIo of the tested enzymes and thus the pH drift became influential), whereas results at the higher pH values of 11.5 and 12.0 (more distant from the tested pIo values, and thus less influenced by the effect of the pH drift towards acidity) showed that there was a clear trend of increasing wash performance as pIo approached the operational pH.

In view of the fact that mutations could alter a protein's wash performance by both pIo-dependent and pIo-independent mechanisms, it would not have been at all surprising if some mutations did not follow exactly the general trend of pIo to pH wash relationship compared to other mutants. A small amount of non-conforming data did not detract for the vast majority of data that clearly supported the trend.

The quality of the evidence in the application (which provided multiple proven solutions to the problem) was of a suitably high quality to make "at least plausible" that the effect demonstrated for other positions could also be achieved at the claimed position 252. The application as filed provided multiple examples of making mutations at other positions (including adjacent position 251) in order to solve the problem, and there were no data to suggest that the other positions that were listed, but not tested in the application as filed would be unsuitable alternatives to those positions proven to provide the effect. Accordingly, it was proper to admit the post-filed evidence D13 that merely confirmed the successful effect of a previously untested, but structurally-related solution that was explicitly disclosed in the application as filed.

Starting from document D1, the technical problem was seen as the provision of a method of generating subtilisin mutants with improved wash performance at a given pH. The solution provided by claim 1 was to modify the subtilisin by mutation at position 252, so that the pIo of the mutant was closer to the pH of the wash liquor than the pIo of the parent enzyme. The question to be answered, therefore, was whether there was anything in the art that would have motivated the skilled person to modify a subtilisin enzyme in order to move the pIo of the enzyme closer to the wash pH, in particular by making a mutation at position 252, in the expectation of achieving some benefit.

There was nothing in the prior art that would have motivated the skilled person to modify the teaching of document D1 in such a way as to arrive at a method having all of the features of claim 1.

The experimental report D13 fully supported a technical effect associated with the claimed subject-matter when modifying position 252. Mutants N252D and N252E did, indeed, perform better than the wild-type parent subtilisin because their pIo was closer to the pH of the wash liquor, whereas N252K and N252R did, indeed, perform worse than the wild-type parent subtilisin because their pIo was further away from the pH of the wash liquor.

The rationale for mutating the 66 amino acid residues listed in original claim 1, which included position 252, in the context of the present invention was that they were located on or close to the surface of the resulting enzyme and, consequently, modifications to the charge of these amino acids could have an impact on the pIo of the protein. Those amino acids did not represent an arbitrary selection from the totality of amino acids in subtilisin. The skilled person would have immediately and unambiguously understood that position 252 was a position that could be modified in the practice of the invention.

It followed that it was permissible to take into account the data in document D13 as evidence when assessing inventive step, since it merely confirmed that the technical effect taught in the application as filed to be associated with modifying position 252 was, indeed, correct.

There was nothing in decision T 1329/04 of 28 June 2005, as relied on by the opponent, to support its position that document D13 should not be taken into account. There was indeed nothing in the application (or elsewhere in the art) that would have led the skilled person to the conclusion that it was implausible that the observed effects might not be obtained by making modifications at residue 252.

The data in document D13 confirmed that the claimed invention did, indeed, solve the problem that was taught to be solved by the application as filed, and so supported a finding in favour of inventive step.

Auxiliary request 1

Allowability of amendments (Article 123(2) EPC)

Support for claim 1, in particular with regard to the feature 'the pIo of the mutated subtilisin protease is closer to the pH of the wash liquor formed by the detergent composition than is the pIo of the parent protease', could be found in various passages of the application as filed, namely on page 74, lines 15 to 19; page 75, lines 36 to 38; page 77, lines 17 to 19 and, more particularly, on page 22, lines 20 to 28.

The term "low" as used on page 22, line 24 to qualify the wash liquor pH was a relative term for which no particular definition had been given in the application as filed. As such, it could be ignored.

Sufficiency of disclosure (Article 83 EPC)

It was only required that the invention was disclosed in a manner that was sufficiently clear and complete for it to be 'performable' by the skilled person. Thus, for the assessment of sufficiency, the issue was whether the application provided adequate guidance to allow the skilled person to perform the claimed invention without undue burden.

In this respect, the skilled person was readily able to determine whether the isoelectric point of the mutant was lower or higher compared to the parent protein, using entirely ordinary methodologies that were well known at the filing date. Paragraph [0138] of the patent made it clear that calculated pIo values were, at the very least, useful to indicate in what direction the pIo for a given mutant enzyme will move in comparison to the pIo of the parent enzyme. A calculation of the exact modified pIo value was not necessarily required to perform the invention. It could be enough, for the purposes of practicing the present invention, to know that the pIo had moved in a particular direction, i.e. either towards or away from the operational pH value of interest, and this could be readily achieved using calculated pIo values. In any case, methods of accurately calculating the pIo of proteins were known in the art at the filing date (see document D44) and if an accurate pIo value was required, the skilled person would have been able to determine such value by well known experimental protocols. As discussed at paragraph [0140] of the patent, pIo values could also be measured (i.e. "observed") if an accurate value was deemed necessary.

All steps of the claimed method were readily performable, without undue burden, by the skilled person, and so the claims fully complied with Article 83 EPC.

Inventive step (Article 56 EPC)

The test data for mutants that had a lower pIo than their parents clearly demonstrated that they reliably possessed improved wash activity compared to their parents in wash liquors that had a pH value lower than the pIo of the parent protein. It was clear that the data in the application as filed made it at least "plausible" that the same effect could be achieved by making similar modifications at position 252. Accordingly, a proper assessment of inventive step of the claimed invention should take into account the post-filing data provided by document D13. That document confirmed that those modifications at position 252, which reduced the pIo of the protein and thereby moved the pIo closer to the pH of the wash liquor than was the pIo of the parent protein, did lead to improved performance. In the application as filed, such a technical effect was ascribed to this modification.

XIV. The submissions made by the respondent (opponent), insofar as they are relevant to the present decision, may be summarised as follows:

Main request

Inventive step (Article 56 EPC)

The application proposed that the wash performance of a subtilisin protease at a particular pH could be improved by altering its isolelectric point to lie closer to that pH. This hypothesis required that a shift in the pIo would result in enhanced wash performance in a detergent solution where the shifted pIo was closer to the pH of the detergent solution and the converse where the shifted pIo was further away from the pH of the detergent solution. The hypothesis predicted that the peak of activity occurred where pH was equal to pIo and that activity should fall away on either side of this value. If the patent's hypothesis were correct, then it should always be possible to improve wash performance by moving the pIo of an enzyme towards the pH of the wash liquor. It followed that the optimum performance should be found when the pIo and the pH were equal. This, however, was not the case, as could be clearly illustrated by the data in Table VI (see page 37 of the patent specification). Thus, the main request did not involve an inventive step.

The post-filed data provided in document D13 certainly did show that technical benefits could be achieved, under some circumstances, by substitution at position 252. The question to be addressed was whether the application provided the kind of disclosure in respect of position 252 which was suitable for support by those post-filed data. Because the link between pIo, pH and wash performance was flawed, the application did not teach (in the sense of providing a real substantial technical teaching) that mutation at position 252 could improve wash performance. It simply proposed 252 as a candidate position for mutation, which could, or could not lead to an improvement of the wash performance. Thus, the improvements found by mutation at position 252 were a "new effect" not disclosed in the application as filed because they were position- and substituent-specific and had nothing to do with pIo. While the application as filed merely put forward a problem to be solved, it was only the data in document D13 which provided the basis to establish that the claimed subject-matter solved indeed the problem it purported to solve. Thus, in line with decision T 1329/04 (see above), document D13 could not be used to support an inventive step.

Auxiliary request 1

Allowability of amendments (Article 123(2) EPC)

The passage on page 22 lines 20 to 28 of the application as filed and the equivalent passage in the parent application as filed (see WO 91/00345, page 22, lines 18 to 26) referred to a wash liquor of low pH and therefore provided support only for a specific embodiment of the method according to claim 1.

Sufficiency of disclosure (Article 83 EPC)

The method of calculating the pIo referred to in the patent-in-suit relied primarily on the assignment of a pK value to each potentially charged amino acid residue, while the pK of a given amino acid residue would depend on its neighbours. As regards the Tyr residues, only three pK values were given, whereas the protein contained seven such residues. As the patent did not specify which Tyr residue corresponded to which pH value, the skilled person would not have known how the pIo would be affected by mutation of any particular Tyr residue. Therefore, no valid method of calculating the pIo was provided by the patent. Nor was the skilled person looking for an alternative method taught how to experimentally measure the pIo. In this respect, document D46 was of no relevance, as the titration technique it described was for gelatine, a non-enzymatic protein, which moreover was tested in a denatured form. Post-published document D43, which described attempts to measure the pIo of subtilisins by isoelectric focusing, illustrated the lack of reliability of such a determination, as the measured value (11.15) was more than a whole pH unit away from the predicted calculated value as given in the patent (10.06 or 10.02), and substantially different from the "observed value" of 9.7 quoted in the patent.

The patentee attempted to explain some discrepancies in its results by invoking "pH drift", apparently caused by environmental CO2 and the grass juice released from the cloth during cleaning. Notably, the patentee provided no evidence that such drift actually occurred during their wash tests. If the patentee was correct about pH drift, then the skilled person did not know what pH he/she was designing the enzyme for.

Inventive step (Article 56 EPC)

The respondent did not provide additional comments on this issue.

XV. The appellant (patentee) requested that the decision under appeal be set aside and the patent be maintained as granted (main request) or on the basis of the auxiliary request 1 filed as auxiliary request 4 with letter dated 15 March 2010.

XVI. The respondent (opponent) requested that the appeal be dismissed.