T 0078/96 (PCR/Hoffmann-La Roche) 30-10-2003

Reasons for the Decision

1. The appeals by opponents (02) and (06) (appellants II and VI, respectively) are admissible.

Novelty (Article 54 EPC)

Oral disclosure by Dr Mullis

2. Relying on documents (D70), (D70.2) and (D70.3) the appellants argue that Dr Mullis, the named inventor, orally disclosed the invention prior to the priority date to a member of the public. Document (D70) is a declaration made by a Dr Faloona, the assistant and fellow employee of Dr Mullis at Cetus Corporation, that he was present at meetings of Dr Mullis with scientists not employed by Cetus to whom Dr Mullis is supposed to have described his invention in detail and the progress of the experiments. Discussions took place in particular with a Dr Ronald Cook.

3. Document (D70.2) are extracts of a post-published book on PCR, one of the authors of which was Dr Mullis. Statements appear here in the first person that before the priority date Dr Mullis discussed with Dr Ron Cook and others the invention, and that Dr Ron Cook was the only one who shared his enthusiasm for the reaction.

4. Document (D70.3) is part of a deposition made in US court proceedings between Hoffman-La Roche, Inc et al (apparently the assignees of the US as well as the European patent in suit) and a licensee. Dr Mullis confirms that he discussed the idea with Dr Cook, and with others. The context is indicated by the statement "I didn't want to announce it publicly in a forum where there were people, but I didn't mind telling Mickey" whom he had previously described as a real good friend whom he bounced things off.

5. None of this represents evidence of precisely what was said, to whom, or when, or that the recipients even thought that they were free to use or disseminate this information. This evidence does not convince the board that anything was made available to the public by the inventor which can be used as prior art to attack the patent in suit.

Document (D4)

6. Document (D4) is concerned with repairing in vitro synthesized, incomplete duplex DNA portions corresponding to parts of the gene for the yeast alanine tRNA, where one of the strands in the duplex is shorter than the other. The authors (Prof. Khorana's group) show that in the presence of a DNA polymerase and the four deoxynucleotide triphosphates, the longer strand acts as a template and the shorter one acts as a primer so that the shorter one is elongated (in direction 5' 3') and the incomplete duplex becomes a fully double-stranded DNA molecule. Radioactive nucleotides are used as means to follow the elongation reaction. At the end of document (4), the following statement is made:

"The principles for extensive synthesis of the duplexed tRNA genes which emerge from the present work are the following. The DNA duplex would be denatured to form single strands, This denaturation step would be carried out in the presence of a sufficiently large excess of the two appropriate primers. Upon cooling, one would hope to obtain two structures, each containing the full length of the template strand appropriately complexed with the primer. DNA polymerase will be added to complete the process of repair replication. Two molecules of the original duplex should result. The whole cycle could be repeated, there being added every time a fresh dose of the enzyme. It is, however, possible that upon cooling after the denaturation of the DNA duplex, renaturation to form the original duplex would predominate over the template-primer complex formation. If this tendency could not be circumvented by adjusting the concentration of the primers, clearly one would have to resort to the separation of the strands and then carry out repair replication. After every cycle of repair replication, the process of strand separation would have to be repeated".

Novelty of Claim 1 over (D4)

7. Claim 1 requires that the sequence to be amplified has to be contained within a larger sequence. Document (D4) does not relate to amplification of a sequence contained in a larger sequence, so the subject-matter of claim 1 is novel over document (D4).

Novelty of Claim 2 over (D4)

8. Claim 2 relates to a process for exponential amplification which requires at least five cycles of amplification (see section II supra).

9. In order to be novelty-destroying, a chemical compound or a process disclosed in a prior art document must have been made available to the public not as a mere chemical formula or hypothetical/theoretical process, but as a reproducible technical teaching (see eg decisions T 206/83 OJ 1987, 5 and T 902/94 of 6 March 1998). This requirement is not fulfilled by the very general suggestions of document (D4). From what is reported in Document (D4), success for even a single cycle of amplification remains uncertain, and the conditions, if any exist, for success with at least five cycles are not given. Novelty must be acknowledged for the subject-matter of claim 2 over document (D4).

Document (D9)

10. This published document emanates from the same group under Professor Khorana as document (D4). The reader is told that the group are now embarking on a different approach to that suggested in (D4), but there is a reference at page 5220, r-h column, first paragraph of document (D9) to unpublished work performed by a member of the group, Dr Molineux pertaining to the replication of relatively short DNA duplexes by means of a "repair replication" process based on the hypothetical process referred to in document (D4). It was stated that these experiments showed that "[(b)] to form the appropriate primer- templates complexes, it was necessary to heat and cool in the presence of an excess (10 times or more) of the appropriate primers and [(c)] in order to perform multiple cycles of repair replication, it was necessary to add, after each cycle, fresh amounts of the primers so as to maintain the appropriate primer-template ratios". No details of what was actually done are given, nor is the reader told how many cyles of repair replication were achieved.

Novelty of Claim 1 over (D9)

11. Claim 1 requires that the sequence to be amplified has to be contained within a larger sequence. Document (D9) does not relate to amplification of a sequence contained in a larger sequence, so the subject-matter of claim 1 is novel over document (D4).

Novelty of Claim 2 over (D9)

12. Claim 2 relates to a process for exponential amplification which requires at least five cycles of amplification (see section II supra). Document (D9) does not even report that such number of cycles was achieved, let alone give a reproducible example of this. Novelty must be acknowledged for the subject-matter of claim 2 over document (D9).

Documents (D3) and (D7)

13. The appellants argue that NIH Grant Application (D3) and Research Proposal (D7), also relating to Professor Khorana's group were publicly available. There is, however, no evidence before the board that these documents were actually available to the public, as required of the purposes of Article 54(2) EPC. In the absence of such evidence and given that the purpose of a Grant Application or a Research Proposal being to obtain funding, not to disclose anything to the public (see decision T 1212/97, supra), the Board cannot treat these as forming part of the prior art.

14. But even if documents (D3) and (D7) were part of the prior art, neither, in the board's judgement, adds anything to the content of document (D4) in terms of "repair replication" since they merely relate to the same "hypothetical" process as disclosed in document (D4), and thus like that document do not take away the novelty of claim 1 or claim 2 for the reasons stated above in relation to document (D4). The drawings on page 37 (document (D3)) or page 18 (document (D7)) illustrate indeed the same principle set out in document (D4), according to which the "The DNA duplex would be denatured to form single strands. This denaturation step would be carried out in the presence of a sufficiently large excess of the two appropriate primers. Upon cooling, one would hope to obtain two structures, each containing the full length of the template strand appropriately complexed with the primer. DNA polymerase will be added to complete the process of repair replication. Two molecules of the original duplex should result. The whole cycle could be repeated, there being added every time a fresh dose of the enzyme".

15. No case for lack of novelty of claims 1 and 2 thus exists on the basis of documents (D3) and (D7).

Oral presentation by Dr K. Kleppe: document (D48)

16. Document (D48) is the manuscript of a lecture given on 18 June 1969 by Dr K. Kleppe, a member of Professor Khorana's group at a Gordon Research Conference. Its contents cannot be regarded as forming part of the prior prior art already in view of the restrictions imposed on persons attending such a Gordon Research conference (cf. decision T 838/97 of 14. November 2000). Nor does the board have any satisfactory evidence what actually would have been conveyed to a member of the audience on this occasion.

17. For the sake of completeness however, and because of its possible relevance to other issues, the actual contents of the manuscript of the lecture (document (D48)) will be considered. Figure 10 thereof shows in schematic form the "repair replication" technique up to the first cycle, yielding two duplexes starting from one single duplex. Fig. 11 of document (D48), relating to the incorporation of radioactive 14C-dCTP in the strands, illustrates a first plateau in the upper curve, corresponding to the first cycle depicted in Figure 10, yielding two (starting) "cold" strands and two (repaired) "hot" strands. The cpms for the first cycle are about 4,000 of incorporated 14C-dCTP in the two "hot" strands. The upper curve then reaches a second plateau corresponding to the second cycle, supposed to yield in total two (starting) "cold" strands and six (repaired) "hot" strands, provided exponential amplification occurs, in which case one would expect to measure 12,000 cpm of incorporated 14C-dCTP (2 "hot" strands = 4,000 cpm; 6 "hot" strands = 12,000 cpm). Yet, the second plateau in Figure 11 is actually at 8,000 cpm, indicating that only two new "hot" strands form during the second cycle, instead of the four expected in an exponential amplification process. The fact that apparently only two new "hot" strands (4,000 cpms) form at each cycle would be consistent with linear amplification of only the original strands, or there might be other explanations. However the experiment does not show that in the second cycle each of strands, either old or newly formed in the first cycle, serves as a template which extended by repair replication in the second cycle. This is the exponential increase required by both claim 1 and claim 2, and so these claims are novel even if the contents of the lecture were made publicly available.

Document (D6)

18. Document (D6) is another document concerned with repair reactions. At page 211, last paragraph of document (D6) it suggests the addition of one single primer ("The duplex could be subjected to a repair reaction by the DNA polymerase of Escherichia coli, the repaired strands separated and the separated strands could again be annealed with a partly complementary polydeoxynucleotide and the repair reaction could be repeated" (emphasis added by the board). The reader is not given the details of any working example, or told what conditions should be used.

19. Both claim 1 and claim 2 require there to be two primers (see step (a) of each of these claims). Novelty of both claims thus has to be acknowledged over document (D6). A further ground for acknowledging novelty is that the document does not contain any teaching reproducible as such, but would require the reader to research for himself the conditions, if any exist, required for success.

20. Since claims 3 to 19 all refer back to the novel process of claims 1 or 2, these claims are also novel. In conclusion no novelty attack has been made out against any of the claims before the board.

Inventive step

Starting point in the prior art and problem to be solved

21. Both claims 1 and 2 are concerned with amplifying double- stranded nucleic acid sequences. According to the prior art this was already successfully done by cloning, as suggested for example in document (D21)(see page 6614, l-h column).

22. According to the appellants the prior to be taken as the starting point is a document concerned with repair replication, such as document (D4) emanating from Professor Khorana's group. However, as discussed above in connection with novelty, this group had itself abandoned this line of research some ten years before the priority date of the present patent, and had not published any results of successful experiments with sufficient detail for others to reproduce for themselves. While the work of Professor Khorana's group may have already suggested many of the features now claimed, the skilled person at the priority date, not having hindsight, would have started from the known successful way of amplifying double-stranded nucleic acid sequences as closest prior art.

23. The problem to be solved over this closest prior art of amplification by cloning can be stated to be finding an alternative method of amplifying specific double-stranded nucleic acid sequences. In view of the considerable detail given in the patent in suit of ways to carry this out successfully, this problem can be regarded as solved by the methods of each of claim 1 and claim 2.

24. To solve this problem the skilled person would have considered inter alia the repair replication documents, in particular document (D4). However the skilled person would only then seriously contemplate this method, if convinced that it can be applied using only routine methods and without having to do any further research of his own. Document (D4) does not show that success has been achieved. The skilled person would thus have first looked at the other available literature on the "repair replication" approach. He would come across document (D9) from the same team but published four years later, and see that from page 5213, they had given up on "repair replication" and adopted a different approach. The skilled person would find nothing that would make him consider that development of the "repair replication" route had any reasonable expectation of success. Even if all the "repair replication" documents were considered together and it could be assumed that they had all had been made available to the public, the skilled reader would still gain no confidence that the method could be got to work, without further research whose outcome remained uncertain, given that the very group which had suggested this line of research appeared from the literature to be abandoning it. Whereas nowadays PCR may seem routine, at the time of filing the present application this line of research was treated as leading nowhere.

25. Finally, it is remarkable that document (D4), the earliest document dealing with an hypothetical process ("repair replication") resembling exponential amplification is dated as far back as 1971, and a number of years elapsed between the publication of document (D4) and the earliest priority date of the patent in suit (1985). In the board's opinion, giving new life to a long-abandoned line of research is a further indicator of inventive step (see eg decision T 330/92 of 10. February 1994), in the sense that the elements underlying the exponential amplification of present claims 1 or 2, such as the templates, the primers, the polymerases and the knowledge about hybridization and separation of DNA strands had long been there, but those skilled in the field have nevertheless remained "blind", the method of choice remaining amplification by in vivo cloning until the priority date of the patent in suit (see eg document (D78), point 12).

Oligonucleotides as only bottleneck?

26. The appellants sought to argue that the group headed by Prof. Khorana did not attempt to carry out a substantial number of cycles of "repair replication" because this would have required a prohibitively high quantity of oligonucleotide primers, which were scarce.

27. However, the board observes that the experiments performed by Dr K. Kleppe involve the use of 50 to 100 µl buffer containing 2 nmol/ml of primer (see page 346 of document (D4), end of first paragraph, in combination with the Legend to Figure 10). Likewise, the exponential amplification technique disclosed on page 13 of the patent in suit involves 100 µl of buffer (line 55) containing 100 pmol (line 56) of "primer A", ie at a concentration of 100 pmol/100 µl = 1. nmol/ml (similar values can be seen also on page 14, lines 30 to 31 and 50 to 52 and page 16, lines 9 to 10 of the patent in suit). Therefore, the conclusion cannot be drawn that Dr K. Kleppe had insufficient amounts of oligonucleotide primers to perform the exponential amplification of DNA.

28. Nor did Dr Kleppe have insufficient amounts of template, as he could use 100 to 500-fold or greater concentrations of template compared with the Examples of the patent in suit. The group headed by Prof. Khorana could thus have been in a position to put into practice exponential amplification with the quantities of primers and templates they had at hand. The only missing element, in the board's view, was the "forma mentis" to do so, as the authors of document (D4) had nothing else in mind than increasing the amount of a given tRNA gene and doubts arise whether they ever had any clue to the exponential nature of what they called "repair replication" and to its amazing power of amplification of trace amounts of nucleic acids. In fact, the term "exponential" never turns up in any of the "repair replication" papers. The board is not in a position to say why the earlier research apparently failed, other than that nobody believed it would work well in practice.

29. In view of the foregoing, it must be concluded that the subject-matter of both claims 1 and 2 satisfies the requirement of Article 56 EPC. Claims 3 to 19 all depend on the inventive process of claims 1 or 2, and thus are inventive as well.