Following your nose can be educational! The value of serendipity?

During the preliminary investigation of our proteomic "hits" from the fractionation of the soluble proteins isolated from mid phase Tenebrio larvae (ie 1-2 weeks old), we discovered a number of proteins whose primary structures (inferred amino acid sequences) suggested a possible role in melanization. This provided a nice opportunity to demonstrate the power and the limitations of proteomics and bioinformatics, and as always, I was delighted by the students' collective findings. I provided them with the Mass Spec data files in which the most significant matches to the analysed peptides were top of the list. They were asked to consider the validity of the computer prompted similarities and to "follow their noses" by searching the Internet for information relating to the biological/biochemical properties of the matches. 

For those of you without a background in this area it is simply a pattern matching exercise. For example, you might be looking at a pair of eyes which provide a strong indication of the "owner". The picture top left of the Mona Lisa's eyes are a real giveaway. Or are they? It is sometimes more appropriate to consider the data as a set of jigsaw pieces: instead of a complete section of sequence data, there are fragments that are matched against global databases of proteins in order to identify the best match. The image on the right is probably the most common scenario: fragments give us clues to the identity of the whole face in this case. It must be stressed that these matches with, in this case the Mona Lisa, are only tentative at this stage, something that is often forgotten! In other words other reference points are needed to establish the true context of the information and therefore interpret it correctly. 

The data that one recovers from a mass spectrometry experiment are shown below to illustrate what I mean. The red letters are amino acids detected and the black letters corresponds to the gaps in the jigsaw, filled in by the computer program.

1	MGLKKIILVL	TALAGLSKAF	NTLSPWDKEV	IYNWKALVKA	GTNLPTSYAS
51	HFVLEGKIHV	QANSNATHVK	ITDLSYKLYN	GILNHVQEHE	AHSVPLHLKE
101	LEHVFKIIYD	TKGLAIGIVT	EAQEKEFSRN	IKKAIATIFQ	MDESVLKQVH
151	HHPHSFVIKE	HSIYGHFDVD	YTIKPHQNEI	EVHKLHDMHD	PHHLYLNLFS
201	NNEAGVCDMG	YENPIMHNTK	KQYRIVKQDG	HHVVKDIHSA	GVVFYYPFKG
251	QSDHQFIFVN	QSMHLTQMVP	IQEQFHIQSE	HFENHLTYHS	FETSQDGSIH
301	DVTNGRHVVN	YEKLIPNIQH	MLSDIHSYLN	EDHIHTREPD	SKRGQMVNRV
351	QRLLQFFSLT	KLEELHKKLS	GDSLTVFHHI	LPLVGTKSST	LFIKQLIVNK
401	VLSAGVASEL	LQHFPEHVIE	PTVELLAEME	ELLHLDESIS	WNVRKVAVLS
451	FASLVQHSHS	FVEKTHRNKE	FMKNDGVDIH	HHETKYSRQQ	PYEKYVIEYI
501	NKLRSSHDYK	VQMTYLMGLC	NMRLHSIIKH	LLPAVYGEWY	PDHHFRLLSL
551	WAVSTAVVEE	RDPNTVIETL	WPIFTNVDEL	TEMRTIAFYF	IMMTYPNDSR
601	LTNMFNYLLM	EPNDEVYSFV	YTFMHGMMET	NEPCHREYAV	KVAHLVKLLP
651	PRSHGRSFAT	RHDYKDVNFD	FSAGVENYFI	NGSDSQIYGV	VTTSQQFNIY
701	HEDRTFFVKI	HGVDHHIDPV	NMAVEVFQKK	IPKITDNKDV	HIEVVLLRHG
751	HIVNTYFFDQ	HSLHQLNEVL	DFFTKESHSL	TQDLLHMSYN	EYHRNLIPTD
801	LGIPALWEYL	MPVVHHNNIS	VTKETVNKVV	HLHVDNRYVN	WIHYRHGLSF
851	YNPFVDVWQG	VNRYHSYDAV	WPLHFEITLN	TQQQSLQLSW	KKDSNPAHNV
901	AGFRSHVKQM	VFIKDDYNRN	ILMQSCPQCK	DFQIVSHGEK	FRNDYLLIQD
951	DDLSTGHQWR	VGIYDSSQYA	TNGSFYHQMY	LFTQSHSSHS	ESPLRHFWMS
1001	YANWHQYCLL	LPEPGSYGLI	VRVEPSHQHS	VSNIDLSLRL	HSDATVDDDY
1051	YYAPKGKWST	RMTYSVKNQE	TSLKTWDVNA	ILDMNSGHTH	NTVKMHVTRI
1101	IPGQKDYKVC	VDGIKKWTES	GVEGHLNVAM	SQTPDGKCVK	DDTVIDVTMT
1151	GQQSHEQEQG	HHKYGSCEYV	NYYAIAHYHT	TQCLVDHTTV	RHYVYNVKTT
1201	NVPTEFKKVF	AHWWDHVKSH	YMSYYVFEDE	HSDNIADQNL	KVDVMYPMQV
1251	EEVNLHVTTP	QHVYKLTGVP	THYWKWFGVA	PDSVEYPEWF	QYLHQVGFVE
1301	HCTVHHNHYH	LNHHDIREAV	PNDWTLYVGN	AVQNPQKAVY	VKQVGDDEIG
1351	VKVVDHGHTL	EIVPHGDGHQ	VTIDGQVVPS	TSIYTKIGCI	VTCLFKILII
1401	RLLLIFQHSG	IHVEYHHHQA	VIVIPKVDQQ	LHGRCYDHN

I think you will agree the probability of finding all of these red letters in a sequence of the correct mass is a strong indicator of the identity of the protein. If we now look at the proteins that have similar sequence signatures, we see that they are associated with a common "domain" (a term used to define a structural or functional component of a protein), called LPD_N, for Lipoprotein, N-terminal domain. This in turn is associated with proteins that transfer lipid molecules (fats) across a cell membrane: one of the most well studied being the protein called vitellogenin, which you will know since it forms the majority of the protein in egg yolk! The other important piece of information recovered from the computer analysis is the link to a publication in which the authors claim: "Activated phenoloxidase from Tenebrio molitor larvae enhances the synthesis of melanin by using a vitellogenin-like protein in the presence of dopamine". I was intrigued when one student told me that the melanization pathway was a feature of a fungus found on the walls of the Chernobyl reactor. I then found the articles entitled: "Ionizing Radiation Changes the Electronic Properties of Melanin and Enhances the Growth of Melanized Fungi"  and "Ionizing Radiation: how fungi cope, adapt, and exploit with the help of melanin" 

It is vital that the students appreciate how powerful Bioinformatics and the Internet can be, but we also need to keep in mind the role of experimental work in addressing, in this case,  the role of melanization in the meal worm, as we explore the diversity of this Biological phenomenon!  So we have information relating to two types of proteins and one strand of biochemical information: the likelihood that this protein is associated with melanization is reinforced by the students' observations that the protein extracts were pale brown as the larvae were extracted and became darker during the day: a feature common with enzymes of this class. I asked the students to take a look at melanization and its significance in Biology.

A protein closely related
to Vitellogenin (Lipovitellin)

The class are now following the information and preparing their final reports and presentations, while I am wondering where we go next, as the prospect of our genomic sequencing experiment looms. The class are now following the information and preparing their final reports and presentations, while I am wondering where we go next, as the prospect of our genomic sequencing experiment looms. It is vital that the students appreciate how powerful Bioinformatics and the Internet can be, in conjunction with experimental Science. It seems that we are combining the development of scientific and technical skills with curiosity driven research in a way that I never imagined could happen in a school!