Who invented MNI-glutamate?

MNI-glutamate has been used by several groups for 2-photon uncaging in the past 5 years, here I outline the history of its development, along with background on the origin of caged compounds.

A group of chemists at Oxford University lead by John Baltrop found in 1966 that ortho-nitrobenzyl redox chemistry liberated carboxylates from their esters. This was the first report of a photochemical protecting group; one that has turned out to be so useful to biologists over the the past 28 years. This was started by two groups who independently used this photochemistry to liberate biologically significant molecules inside cells: 1. Engels and Schlegger (J. Med. Chem. 20 (1977) 907-911) photoreleased cAMP in rat glioma cells to induce morphological changes (the photosensitive cAMP was only one of six derivatives of cAMP made for this study, the other phosphate esters were hydrolyzed by intracellular esterases-a technique cleverly exploited later by Roger Tsien with the AM esters of BAPTA Nature (1980) 290: 527-8); and 2. Kaplan and co-workers photoreleased ATP inside red cells and studied potassium efflux generated by the sodium pump (Biochemistry (1978) 17: 1929-1935). The latter study is rightly regarded as the more significant one, as Kaplan's experiment was on a rapid time scale. By the way, it was Kaplan's boss, Joe Hoffman, who dubbed ortho-nitrobenzyl-ATP "caged ATP". The name has stuck ever since, despite the (recent) efforts of organic chemists to give the idea other names, in vain attempts to attract more attention and significance to their own work.

From 1970 a group of chemists lead by Armit Patchornik published a series of pioneering papers on photochemical protecting groups. They introduced the ortho-nitroveratryl or 4,5-dimethoxy-2-nitrobenzyl cage ( J. Am. Chem. Soc. (170) 92: 6333-6335) as an extension of the generic photochemistry of Baltrop's 1966 paper. They also were the first to use nitroaromatic redox chemistry to liberate carboxylates from acylindolines ( J. Am. Chem. Soc. (1976) 98: 843-4; J. Am. Chem. Soc. (1981) 103: 7674-5). These studies are the real inspiration for MNI-glu.

Roger Tsien's group made a nitroindolinyl-BAPTA ( J. Am. Chem. Soc. (1989) 111: 7957-68), hoping to cage the chelator. However, the derivative had poor quantum yield, so nitroindolines were forgotten again. Ten years later the Mill Hill group resurrected the protecting group, making 5-methoxycarbonylmethyl-7-nitroindolinyl-glutamate (which they strangely call "NI-glu"-I guess MCMNI-glu is too long). This very important paper was published online on 24 June 1999 (J. Am. Chem. Soc. (1999) 121: 6508-4). They caged glutamate via the hydrolytically stable indolinyl bond, and so solved one of the central practical problems in this field (namely, spontaneous hydrolysis of caged glutamate). However, the quantum yield and absorption were less than optimal, but as soon as this work appeared I realised that MNI-glutamate would work well for 2-photon uncaging.

The reason for this was that before this paper appeared, I had made an analogous caged glutamate (DMCNB-glu). Using this probe, in collaboration with Kasai's lab, we managed to get the first really decent 2-photon uncaging of glutamate (see picture above). I presented this work in a lecture at the Society for General Physiologists annual conference during the summer of 1999 (J. Gen. Physiol. (1999) 114: 1a). I made MNI-glutamate for the first time on 29th March 2000, about 6 months before the Mill Hill group actually published their synthesis of MNI-glu (Tetrahedron (2000) 56: 8197-8205), and over 18 months before their biological evaluation appeared (J. Neurosci. Meth. (2001) 112: 29-42; online 12 October). However, we wanted to do something really interesting with this caged compound, as we realised that MNI-glu had the potential to revolutionise the field, by making extracellular 2-photon uncaging work well for the first time. Thus, I delayed publishing my synthesis of MNI-glutamate until 22 October 2001 (Nature Neurosci. (2001) 4: 1086-1092-this paper was submitted in June 2001).

So, who invented MNI-glutamate?

The Mill Hill group deserve the credit for resurrecting the nitroindoline protecting group (JACS 1999) and doing nice pharmacology of MNI-glu (JNM 2001). However, the lion's share of the credit must go to the Haruo Kasai and his colleague Masanori Matsuzaki, who got DMCNB-glu, then MNI-glu to work fantastically well for 2-photon uncaging, thus fulfilling Winfried Denk's original idea of 2-photon uncaging microscopy (PNAS (1994) 91: 6629-6633). Since then many groups have published cool papers using 2-photon uncaging of MNI-glu, but would have never attempted it if Masanori hadn't first proved its feasibility.

Axel versus Miesenbock: what in a name?

The fluorescence of GFP is pH sensitive, and so mutants of GFP have been made to counteract this problem. In 1998 Gero Miesenbock published a paper as a postdoc in Jim Rothman's lab in Nature using the weakness of GFP to his advantage. He tagged GFP with a synaptic vesicle protein so that when the vesicle fused with the plasma membrane, the exocytotic event was signalled by light. GFP was, in effect, a "flare" for synaptic activity (they were dubbed "synapto-pHluorins"). Miesenbock left Jim Rothman's lab, but stayed at Sloan to published the first creative application of this new probe in August 2002. I heard Gero give a presentation of this work at the McKnight annual conference in June of that year, and was amazed by the presentation. He had built his own video-rate 2-photon microscope so he could image synaptic activity in vivo using his synapto-pHluorins. His lab targeted the probe to the Drosophila antenal lobe, and depending on the type of odor (banana, cherry, apple), they found different populations and combinations of glomeruli "lit up"; so they were able to see dynamic odor representation for the first time, "in elements of the olfactory that had previously been inaccessable to direct optical analysis, in any species".

In October 2002 Richard Axel submitted a paper to Cell doing the similar studies, but using a GFP-based Ca probe (caled G-CaMP) as the indicator of synaptic activity. He makes only passing mention of Miesenbock's earlier paper, saying, "Recently, imaging studies with synapto-pHluorin, a fluorescent indicator of pH change, have revealed dense patterns of glomerular activity in the fly antennal lobe (Ng et al., 2002). This fluorescent indicator has been ex pressed in defined neurons and affords spatial resolution but has limited sensitivity due to low signal-to-noise ratio. As a consequence, analysis requires non-physio- concentrations that may result in spatial representations not naturally encountered by the fly brain." What Axel fails to mention is that his GFP-based Ca indicator is very slow and because it requires 4 calciums to give a signal, will heavily buffer Ca in an uncontrolled way, undoubtedly dramatically purterbing synaptic actvity. (As a further technical aside, since Axel used galvonometers, the imaging reported would have been slow, even if his probe was fast!). This comment is burried in the results section, giving the impression it is not really important, whereas in effect Axel had merely followed on from the pioneering work of the Miesenbock lab (Axel was in fact on Sloan-Kettering's scientific advisory board at that time). Don't get me wrong, both papers are great, but shouldn't a big shot give credit to a young scientist? It is not as if Axel he needed just one more Cell paper to get the Nobel (to his credit, Axel does at least mention Miesenbock's paper in his Nobel Address in 2004).

However, to date, Axel's paper has been cited about 120 times and Misenbock's 70.
What's in name? Apparently quite a lot.

Ng M, Roorda RD, Lima SQ, Zemelman BV, Morcillo P, Miesenbock G. Transmission of olfactory information between three populations of neurons in the antennal lobe of the fly. Neuron (2002) 36:463-474.

Wang JW, Wong AM, Flores J, Vosshall LB, Axel R Two-Photon Calcium Imaging Reveals an Odor-Evoked Map of Activity in the Fly Brain. Cell (2003) 112: 271-282.

So you think LTP is meaningless?

LTP was reported by Bliss and over 30 years ago now. Even though many labs have studied the process, there is a great deal of cynicism from many quarters about the phenomenon. The nay-sayers have often said there is no real evidence that LTP actually has anything to with memory. A paper (Science (2003) 299: 1585-8) from the Malinow lab goes some way to contradict such an idea. In a characteristically elegant series of experiments the Malinow shows that significant experience does indeed induce LTP in vivo in rodents.

They use whisker plucking as the means to deprive neurons of experience. Thus, they injected DNA for one subunit of the AMPA-R (which was also tagged with GFP to guide patching, “GluR1-GFP”) at PND 12 into the barrel cortex of mice. At PND 14 they made acute slices to test for LTP in the usual way. Since not all neurons become transfected in vivo, they have convenient in-built control neurons to give them the % rectification, along with ipsi/contralteral control populations of cells as well (transfection with GFP alone had no effect). The Malinow lab found that mice without whisker trimming (“intact”) had populations of neurons that were rectified. This was the case with whisker-trimmed animals (“deprived”) only when they analyzed ipsilateral neurons; contralateral neurons showed no rectification. Thus, significant sensory experience for the mouse (whiskers are very important for rodents’ life) increases synaptic strength. This finding was substantiated and embellished by several more experiments: (1) transfection with the cytoplasmic tail of AMPA GluR1 blocked LTP in intact but not deprived animals [ Fig 2]; (2) GluR1-GFP could be replaced by GluR2 if intact animals from PND 12 to 14 were then rendered deprived for 1 day subsequently [Fig. 3]; (3) Previously the Malinow lab had developed a very clever way of using electrophysiological tagging to assay if APMA receptors become incorporated into synapses in slices (R586Q mutant). They made a mutant channel that works in a slightly different way from endogenous receptors. Transfection of animals with GluR2-R586-GFP along with a GluR2 cytoplasmic tail showed that experience was not necessary to block GluR2 incorporation [Fig. 4]. Thus, sensory input does give rise to LTP. Perhaps it is worth studying after all.