You are on page 1of 2

S C I E N C E ’ S C O M PA S S

65 method of diffusion tensor imaging (DTI)


A B
64 has arrived (see the figure). DTI more pre-
63 fMRI cisely measures neuroanatomical changes in
62 white-matter fiber tracts (16). This tech-
61 nique holds promise for tracking neu-
60 roanatomical changes in the strength and
59 number of neuronal connections and in fiber
58 DTI myelination with learning and development.
57 Ultimately, we will be able to correlate DTI-
56 based neuroanatomical measures with be-
55 Z = 16 Z = 24 Z = 32 havioral and neurophysiological measures
54 of the speed of cognitive and neural process-
53 Images of the human brain. (A) Five-year-old child in an MRI scanner. (B) Recent noninvasive MRI ing. This will be achieved by combining DTI
52 methods measure the function and structure of a child’s brain. The top row depicts patterns of brain with fMRI and methods of higher temporal
51 activity indexed by fMRI in three representative axial (Z) slices. The bottom row shows corticospinal resolution (such as, evoked potential re-
50 white-matter fiber tracts (green) projecting through the same three axial slices measured by DTI. sponses). Clearly, these methods will pro-
49 mote our understanding of how human brain
48 ed to maturational changes versus behav- and synaptic pruning occur during develop- development and behavior change with
47 ioral differences (5, 7). Subsequently, each ment (15). Rather, these findings highlight growth and experience.
46 of these variables can be used as a covariate the maturation-versus-experience question
45 to determine the degree to which these vari- and suggest a more precise test: examining References
44 ables independently contribute to changes in brain activity before and after extended 1. K. K. Kwong et al., Proc. Natl. Acad. Sci. U.S.A. 89,
5675 (1992).
43 brain activity. However, age and behavioral practice on a task to determine whether the 2. S. Ogawa et al., Proc. Natl. Acad. Sci. U.S.A. 89, 5951
42 performance correlate with each other on immature system after extended practice en- (1992).
41 many behavioral tasks. Finally, we can gages in the same neural mechanisms as the 3. E. Spelke, Dev. Sci. 5, 392 (2002).
4. B. L. Schlaggar et al., Science 296, 1476 (2002).
40 group individuals on the basis of their per- mature system. Using fMRI to trace learn- 5. B. J. Casey et al., J. Cogn. Neurosci. 9, 835 (1997).
39 formance post hoc, as Schlaggar et al. de- ing-related changes in cortical areas should 6. B. Luna et al., Neuroimage 13, 786 (2001).
38 scribe. Accordingly, we can compare differ- be informative when investigating the im- 7. T. Klingberg et al., J. Cogn. Neurosci. 14, 1 (2001).
8. S. A. Bunge et al., Neuron 33, 301 (2002).
37 ent age groups with similar behavioral per- pact of behavioral training interventions for 9. K. M. Thomas et al., Biol. Psychiatry 49, 309 (2000).
36 formance or the same age groups with dif- developmental disorders such as dyslexia; 10. L. J. Chapman, J. P. Chapman, J. Psychiatr. Res. 14, 303
(1978).
35 ferent performance. In the case of tasks such such research is under way at the Sackler In- 11. T. S. Braver et al., Neuroimage 5, 49 (1997).
34 as single word processing, this method is ef- stitute and other institutions worldwide. 12. S. Durston et al., Neuroimage, in press.
33 fective. However, this approach is valuable So what is in store for the field of devel- 13. A. Karni et al., Nature 377, 155 (1995).
14. L. Hertz-Pannier et al., Neurology 48, 1003 (1997).
32 only when the different age groups have opmental science with continued advances 15. J. P. Bourgeois et al., Cereb. Cortex 4, 78 (1994).
31 overlapping distributions in response laten- in MRI? Along with advances in fMRI, the 16. C. Pierpaoli et al., Radiology 201, 637 (1996).
30 cy and accuracy.
29 A question that Schlaggar et al. fail to P E R S P E C T I V E S : B I O C H E M I S T RY
28 answer is whether the differences in brain
27 activity that they observed between children
26
25
and adults are due to an immature central
nervous system or a lack of experience with
The 22nd Amino Acid
24 the task. This question highlights the beauty John F. Atkins and Ray Gesteland
23 of fMRI, which can safely be repeated in the
22 same subjects multiple times, allowing us to wo complementary reports (1, 2), on mRNA encoding this enzyme, and identify
21
20
19
track changes in cortical activation after
children have had extensive practice with a
particular task. Such an approach may pro-
T pages 1459 and 1462 of this issue, pro-
vide compelling evidence that the ge-
netic code of certain Archaea and eubacteria
special characteristics of the tRNA carrying
this nonstandard amino acid.
The way in which pyrrolysine is encod-
18 vide a more definitive test of whether devel- encodes a 22nd amino acid. This nonstan- ed bears striking parallels to the encoding
17 opmental differences are maturation- or ex- dard amino acid, called pyrrolysine, is en- of the 21st amino acid, selenocysteine. Se-
16 perience-based by assessing brain activity coded by the RNA nucleotide triplet UAG, a lenocysteine is found in Archaea, eubacte-
15 both before and after training. Karni et al. stop codon that halts translation of mRNA. ria and animals, including mammals (3, 4).
14 (13) showed rapid learning effects in prima- Krzycki, Chan, and their colleagues (1, 2) Both nonstandard amino acids are encoded
13 ry motor areas of adults who performed mo- show by chemical and structural analysis of by the RNA nucleotide triplets (codons)
12 tor sequence learning tasks within a single proteins from the archaeon Methanosarcina that signify a command to stop translation
11 session. These effects increased even further barkeri that pyrrolysine is present in the ac- of mRNA into protein (UGA is the “stop
10 with several weeks of training during which tive site of the enzyme, methogenic methyl- codon” encoding selenocysteine). The no-
9 the cortical activity became less diffuse. amine methyltransferase, which catabolizes tion that at least 22 amino acids are directly
8 This example of initial diffuse cortical activ- methylamines leading to the production of encoded by the nucleotide sequence of
7 ity early in learning parallels results from methane. These authors demonstrate that mRNA reflects the greater richness of the
6 developmental fMRI studies showing dif- UAG is at the corresponding position in the genetic code than is apparent from the
CREDIT: A. ULUG, R. WATTS

5 fuse activity in children relative to adults (5, standard textbook account.


4 14). This is not to say that differences in Originally, the coding problem was de-
The authors are in the Department of Human Genetics,
3 brain activity between age groups are due to University of Utah, Salt Lake City, UT 84112–5330, USA. fined in terms of how the 20 common amino
2 experience alone; even without normal stim- E-mail: john.atkins@genetics.utah.edu J.F.A. is also at acids could be specified by four RNA nu-
1 ulation, changes in neuronal connections the Science Foundation Ireland, Dublin 2, Ireland. cleotides. As the triplet nature of the genetic

www.sciencemag.org SCIENCE VOL 296 24 MAY 2002 1409


S C I E N C E ’ S C O M PA S S
65 code began to unfold in the downstream sequence producing one pro-
64 Two modes of
early 1960s, it might have codon redefinition Standard amino acid tein from two separated open reading
63 been tempting to speculate UAG = Gln frames (9).
mRNA context- UGA = Trp
62 that some of the 64 possible dependent redefinition For pyrrolysine, like selenocysteine, the
61 codons encoded the many rare of stop codons 21st amino acid critical feature is specification of an addi-
60 amino acids found in proteins. UGA = selenocysteine tional amino acid. But whether specifica-
59 However, it became clear that 22nd amino acid tion of pyrrolysine is due to “permanent”
?
58 20 is the correct number of Universal UAG = pyrrolysine reassignment of UAG or to recoding of a
57 amino acids, and that the code subset of UAG codons is not yet clear (see
56 e.g.:
great majority of nonstandard the figure). Does “UAG” mean “pyrroly-
AGR = Ser (not Arg)
55 amino acids are created by mRNA context- AAA = Asn (not Lys) sine” wherever it occurs in these organisms,
54 chemical modifications of independent AUA = Met (not Ile) or are there signals in specific mRNAs that
53 standard amino acids after redefinition of codons UGA = Trp redefine UAG codons one at a time?
52 translation. In 1986 came the Natural selection has successfully led to
51 surprise discovery that the a code that specifies more than the 20 stan-
50 nonstandard amino acid se- When stop means go. There are two ways in which the stop codon dard amino acids. Meanwhile, human ef-
49 lenocysteine is directly speci- UAG could be redefined to specify the 22nd amino acid, pyrroly- forts to achieve the same goal are actively
48 fied by the genetic code and is sine. In the first (top), special signals in mRNAs tag a subset of stop under way, particularly by Schultz and his
47 not created by posttranslation- codons that are to have their meaning redefined. In the second colleagues (10). The challenges are
(bottom), a codon is redefined regardless of the mRNA involved.
46 al modification (5, 6). Seleno- formidable, but striking progress is being
45 cysteine is now joined by made. The goal is to engineer the direct en-
44 pyrrolysine, and together these two amino codons whose meaning is to be changed. coding of additional amino acids at will. It is
43 acids demonstrate that the genetic code can These signals can be close to the UGA by no means certain, however, that all of the
42 be expanded by redefining the meaning of a codon, as in bacteria, or distant, as in the 3′- future excitement in this area will come
41 stop codon. untranslated region of eukaryotic mRNAs. from these manipulations alone. As pyrroly-
40 This redefinition requires subversion of Recoding UGA as selenocysteine is not sine illustrates, nature may yet surprise us
39 the standard pathway for activating amino fully efficient because there is direct com- with more directly encoded amino acids.
38 acids in readiness for protein synthesis. In- petition from standard reading of UGA stop
37 stead of a tRNA being charged with the new codons, even those that are specially References
36 amino acid, it receives a standard amino “tagged.” The slow-to-decode property of 1. G. Srinivasan et al., Science 296, 1459 (2002).
2. B. Hao et al., Science 296, 1462 (2002).
35 acid that is then enzymatically modified “stop codons” may be the reason for usurping 3. M. Rother et al., J. Mol. Biol. 299, 351 (2000).
34 while still attached to the tRNA. This pro- them during the decoding of standard amino 4. M. J. Berry et al., Biofactors 14, 17 (2001).
33 cess is similar to the way in which some or- acids. The goal in recoding UGA is to spec- 5. I. Chambers et al., EMBO J. 5, 1221 (1986).
6. F. Zinoni et al., Proc. Natl. Acad. Sci. U.S.A. 83, 4650
32 ganisms modify the standard amino acids, ify selenocysteine, the “special” 21st amino (1986).
31 aspartic acid and glutamic acid, while they acid. However, there are other cases of re- 7. M. Ibba, D. Soll, EMBO Rep. 2, 382 (2001).
30 are attached to tRNAs, in order to obtain as- coding where a standard amino acid is 8. S. Osawa, in Evolution of the Genetic Code (Oxford
29 Univ. Press. Oxford, UK, 1995).
paragine and glutamine (7). In the case of specified by a “stop codon.” Here the im- 9. R. F. Gesteland, J. F. Atkins, Annu. Rev. Biochem. 65,
28 selenocysteine, a selenocysteinyl tRNA is portant feature is that readthrough of the 741 (1996).
27 first charged with serine, which is then en- stop codon permits continued decoding of a 10. L. Wang et al., Science 292, 498 (2001).
26 zymatically modified to form selenocys-
25 teine. Similarly, pyrrolysine is likely to be PERSPECTIVES: HIV/AIDS
24 produced by modifying a lysine residue at-
23 tached to a special lysyl tRNA. The tRNAs
22
21
involved in the production of selenocysteine
and pyrrolysine are distinct from those de-
HLA Leaves Its Footprints on HIV
20 coding the standard amino acids, serine and Andrew McMichael and Paul Klenerman
19 lysine, but they differ from each other in
18 certain features, for example, the pyrroly- he reasons for the poor control of During both the acute and chronic
17
16
15
sine tRNA has a “special” anticodon arm.
In certain specialized niches, such as the T HIV infection by the mammalian im- phases of HIV infection, production of cy-
mune system are gradually being un- totoxic T lymphocytes (CTLs) by the host
mitochondrion, the meaning of a subset of raveled. The ability of certain HIV pro- immune system exerts a strong inhibitory
14 codons is reassigned from that of the “univer- teins to mutate and thus to elude immune effect on HIV growth and replication.
13 sal” genetic code wherever these codons oc- detection is increasingly seen as crucial. Therefore, it is not surprising that there is
12 cur in all mRNAs. For example, in the On page 1439 of this issue, Moore et al. strong selective pressure for survival of
11 starfish, the codons UGA and AGR specify (1) provide new evidence for critical in- HIV mutants that escape the CTL re-
10 Trp and Ser, respectively, when in mitochon- volvement of HLA proteins of the human sponse (2–7). Although escape from the
9 drial mRNA, but Stop and Arg when in nu- histocompatibility complex in shaping host antibody response is well accepted
8 clear mRNA. In a few organisms with small variations in HIV proteins and possibly (8), escape from CTL responses has until
7 genomes, rarely used codons are permanently evolution of the virus itself. recently been more controversial. Objec-
6 reassigned (8). However, in organisms where tions to the idea have centered around
5 UGA specifies selenocysteine, only a subset whether a CTL response against several
4 of UGA codons do so; the great majority A. McMichael is at the Weatherall Institute of Molec- HIV epitopes could be undermined by es-
ular Medicine, Oxford OX3 9DS, UK. E-mail:
3 specify the standard meaning: “stop transla- andrew.mcmichael@clinical-medicine.oxford.ac.uk P. cape of only one epitope. With the host
2 tion.” Special signals in mRNA help to repro- Klenerman is with the Medawar Centre for Pathogen immune system trying to control a swarm
1 gram the readout by distinguishing those Research, University of Oxford, Oxford OX1 IS5, UK. of rapidly replicating viruses, a viral vari-

1410 24 MAY 2002 VOL 296 SCIENCE www.sciencemag.org

You might also like