OUP user menu

The His452Tyr variant of the gene encoding the 5-HT2A receptor is specifically associated with consolidation of episodic memory in humans

Michael Wagner , Anna Schuhmacher , Sybille Schwab , Astrid Zobel , Wolfgang Maier
DOI: http://dx.doi.org/10.1017/S146114570800905X 1163-1167 First published online: 1 December 2008


Basic research has shown that serotonin plays an important role in memory formation. Accordingly, genetic variation of serotonin receptors can be expected to affect memory and to underlie, in part, the heritability of memory capacity. A study by de Quervain and colleagues found a highly significant association of the functional 5-HT2A receptor variant His452Tyr (rs6314) with long-term memory. We replicated this finding in a cohort of 133 adults (mean age 43.5 yr). Carriers of the Tyr allele showed poorer verbal delayed recall and recognition, while immediate recall and additional measures of attentional and executive function were not affected by the His452Tyr genotype. Results suggest a possible role of 5-HT2A receptors in memory consolidation. Serotonergic drugs may have the potential to improve memory.

Key words
  • Memory
  • serotonin
  • 5-HT2A receptor


Human memory is a polygenic and heritable trait. Twin studies suggest that as much as 50% of the variance in memory tasks may be attributed to genetic differences (McClearn et al., 1997). Animal studies have identified molecules and proteins directly related to a cascade of molecular events leading to the formation of memory (Kandel, 2001). Many of these proteins are encoded by genes which are also expressed in humans. Recent human genetic studies have identified variations within these genes which appear to be associated with memory capacity, including variants in genes coding for brain-derived neurotrophic factor (BDNF; Egan et al., 2003; Ho et al., 2006), the metabotropic glutamate receptor GRM3 (Egan et al., 2004), KIBRA (Papassotiropoulos et al., 2005; Schaper et al., 2008), and a gene cluster related to intracellular signalling (de Quervain and Papassotiropoulos, 2006).

Serotonin and its receptors are important for learning and memory (Buhot, 1997; Buhot et al., 2000; Schmitt et al., 2006). In animals, pharmacological manipulation of specific serotonergic receptors as well as tryptophan administration influence hippocampus-dependent memory (Haider et al., 2007; Harvey, 2003; Meneses, 2007). In humans, acute tryptophan depletion, which lowers 5-HT levels, impairs declarative memory consolidation (Harrison et al., 2004; Schmitt et al., 2000). Recently de Quervain et al. (2003) reported that the gene encoding the human 5-HT2A receptor (HTR2A), which harbours a functional variation leading to an amino-acid substitution (His452Tyr), was associated with performance in an episodic memory task. The 5-HT2A receptor is a G-protein coupled receptor which, upon stimulation, regulates intracellular calcium levels and PKA phosphorylation, two crucial events for memory formation at a cellular level. In humans, the tyrosine allele of the His452Tyr results in blunted receptor response upon serotonin stimulation (Ozaki et al., 1997) and in destabilization of signal conformation, presumably due to reduced G-protein coupling (Hazelwood and Sanders-Bush, 2004). The hypothesis that His452Tyr also affects human memory was tested by de Quervain et al. (2003), who indeed found the tyrosine allele to be associated with a reduced ability to recall items learned 5 min or 24 h ago, while immediate memory was not affected. The effect size of the delayed memory effect was moderate, but quite remarkable for a single gene polymorphism. This genetic association, if replicated, might have important implications for the development of pharmacological agents to enhance memory.

We further examined the influence of the HTR2A gene on memory in 133 healthy German individuals who completed a standardized episodic memory task. We also sampled a broader range of cognitive functions than de Quervain et al. (2003) in order to extend their findings and to assess the specificity of putative effects of the His452Tyr polymorphism.


Study participants

Participants were recruited by contacting a random sample of 900 residents of Bonn, Germany by mail. Respondents were screened by phone and by personal interview for the absence of severe or chronic diseases. Subjects taking any medication besides oral contraceptives were excluded from the study. Furthermore, subjects were screened for the presence of Axis-I psychiatric disorders with the DIA-X-SSQ questionnaire, an instrument with high sensitivity to detect psychiatric disorders (Wittchen and Perkonnigg, 1997). Depressive symptoms were assessed with the Beck Depression Inventory (BDI). The sample entering the study consisted of 133 healthy subjects aged 18–71 yr. All subjects provided written informed consent for this study. The study was approved by the IRB of the medical faculty of the University of Bonn.

Neuropsychological assessment

Neuropsychological examinations were performed by trained research assistants blind to the genotype of the subjects. Memory functions were assessed with the Auditory Verbal Learning Test (German version, Helmstaedter et al., 2001), which requires subjects to repeat and memorize a list of 15 semantically unrelated words. Immediate memory and learning is measured by summing the total number of words correctly recalled after each of five consecutive trials. After learning and recalling words from an interference list of 15 different words presented only once, subjects are asked to recall the initial list (recall immediately after interference, trial 6). Thirty minutes later they are again asked to recall the initial list (recall after delay, trial 7) before being asked to recognize the words of the initial list interspersed among 20 new distractor words. Rate of forgetting is measured by the difference between number of words recalled after the fifth learning trial and the number of words recalled after 30 min (trial 5 – trial 7). Delayed recall, rate of forgetting, and delayed recognition are established measures of long-term memory (Spreen and Strauss, 1998).

To assess attentional and executive functions we administered the Trail-making test A and B (Reitan and Wolfson, 1993), the letter cancellation test d2 (Brickenkamp and Zillmer, 1994), the digit span task of the Wechsler Adult Intelligence test (Wechsler, 1997), a phonemic fluency task (Spreen and Strauss, 1998), and a spelling interference task, which is part of a screening test for organic brain dysfunction (Lehrl and Fischer, 1985). The latter task presents subjects with a card containing the letters A and B in random sequence (e.g. AABABBB), requiring them to spell instead the ‘inverted’ sequence (e.g. ‘BBABAAA’).


In all participants, DNA for genotyping was isolated either from EDTA anticoagulated blood or permanent cell cultures received after transforming the lymphocytes with Epstein–Barr virus. The isolation of the DNA followed the Qiagen protocol for the Blood & Cell Culture DNA Maxi kit (Qiagen, Hilden, Germany). PCR was performed using 12.5 ng of DNA. SNP rs6314 represents the functional polymorphism that predicts the histidine to tyrosine substitution at residue 452 (His452Tyr). The assay for SNP rs6314 was a Taqman® assay. The procedure followed the protocol for Taqman assays supplied by Applied Biosystems with the use of Taqman Universal PCR MasterMix, No AmpErase UNG (Applied Biosystems, Foster City, CA, USA). The assay consists of the unlabelled forward primer and the unlabelled reverse primer and two reporters, that are dye-labelled with FAM and VIC. The assay is designed for allelic discrimination of specific SNPs. Both alleles were scored in a single well by measuring the fluorescence at the end of the PCR using a Tecan Ultra 384 reader (Tecan, Crailsheim, Germany). Excitation and emission wavelengths for the FAM-labelled probes were 485 nm and 535 nm and for the VIC-labelled probes 535 nm and 590 nm, respectively.


We compared genotype groups for possible differences with respect to memory and other cognitive functions using t tests, hypothesizing that recall and recognition, but not immediate recall, would be affected by the rs6314 genotype. No a-priori hypothesis was specified for the other cognitive functions. The significance level was set at p<0.05 (two-tailed). No correction for multiple testing was made because of the a-priori hypotheses regarding memory, and because many of the cognitive measures can be expected to correlate with each other. Effect sizes (Cohen's d) were calculated from means and standard deviations of both genotype groups. In order to compare effect sizes between studies, we calculated Cohen's d from the F ratios reported in figure 1 of de Quervain et al. (2003) with the method of Thalheimer and Cook (2002).


Genotype frequencies did not significantly differ from Hardy–Weinberg equilibrium (χ2=0.88, d.f.=1, p=0.384). There were 113 subjects homozygous for the C allele (His/His) and 20 heterozygous subjects (C/T, His/Tyr), no subject homozygous for the rare allele was encountered. Both genotype groups did not differ significantly from each other (all p>0.20) with respect to age (mean 43.5 yr), verbal IQ (mean 113.8), BDI score (mean 4.1), and gender distribution (50% female).

We found that the tyrosine allele of the His452Tyr SNP was specifically associated with reduced episodic memory. As can be seen from Table 1, the carriers of the rare Tyr allele were selectively impaired with regard to delayed recall and delayed recognition. While they learned equally well as the non-carriers, they forgot more words within 30 min. All non-episodic memory measures, however, did not differ between groups.

View this table:


The rare Tyr variant of the 5-HT2A gene was associated with poorer delayed recall performance, while immediate memory was not affected. This pattern exactly replicates the findings of de Quervain et al. (2003). Markedly, all other cognitive measures appeared to be unaffected by the His452Tyr variant, although our sample size does not allow the conclusion of a lack of effect from a lack of significance. In considering only the effect sizes, we note that His452Tyr might also affect verbal fluency, which might be related to the fact that verbal fluency, in contrast to the other additional measures, also requires, and thus is an indirect measure of, verbal long-term memory (semantic memory). The other tests of attention and executive function were not affected by the His452Tyr variant, corroborating that this is a specific long-term memory effect. These results are in line with data from a study manipulating serotonin availability by Schmitt et al. (2000). They found that serotonin depletion resulted in impaired recognition, without effects on immediate memory, suggesting that serotonin specifically affects long-term memory consolidation, primarily in the first 30 min after acquisition (Schmitt et al., 2000).

To our knowledge, this is the first replication of the findings of de Quervain et al. (2003). Papassotiropoulos et al. (2005), comparing the original data of de Quervain (mean age 22.8 yr) with additional data from an elderly sample (mean age 67.9 yr), found His452Tyr to exert a significant effect only in the initially studied younger sample. While our sample was too small to test for an interaction with age, we note that the main effect of rs6314 was significant despite the inclusion of elderly subjects.

Apart from changes in the neurochemical cascade involved in memory consolidation, the His452Tyr variant may also been related to mediotemporal lobe structures essential for memory. In healthy subjects aged ⩾40 yr, Filippini et al. (2006) found the Tyr allele (which was linked to poorer memory in the present study and in the study by de Quervain et al., 2003) to be associated with reduced grey matter in the left hippocampus, left inferior temporal gyrus, and bilaterally in the middle and superior temporal gyrus, which would be consistent with reduced episodic memory formation. Unfortunately, memory was not assessed in the study by Filippini et al. (2006).

The effect size for the recall after 30 min in the present study (d=0.53) was broadly similar to the effect sizes observed by de Quervain et al. (2003) for 5-min free recall (d=0.41) and for recall 24 h later (d=0.38). The specificity of the influence on memory strongly suggests that it is not brought about by attentional factors or by working memory, which can affect the encoding stage of memory formation. Rather, the His452Tyr polymorphism appears to be related to memory consolidation in the hippocampus, in line with animal research on the role of serotonin on long-term memory formation.

In summary, the SNP rs6314 exerts a medium-sized effect on verbal episodic memory. To our knowledge, this constitutes the first independent replication of the finding by de Quervain et al. (2003). The pattern of results as well as the effect sizes are in line with the original study and provide further evidence for an association. However, studies with larger samples of subjects addressing several aspects of memory (including spatial and emotional memory, in addition to verbal memory) are needed to confirm the role of the His452Tyr variant for long-term memory.

Accumulating evidence highlighting the role of serotonin for memory consolidation in humans may foster the further search for memory-enhancing agents, in order to treat conditions like mild cognitive impairment or memory deficits in depression. Some evidence already suggests that increasing serotonin availability with SSRIs may have positive effects. In patients with mild cognitive impairment, fluoxetine has been found to improve memory in a randomized controlled trial (Mowla et al. 2007), and Vythilingam et al. (2004) described improvement of impaired delayed recall performance in depressive patients after their symptoms were successfully treated with SSRIs. Whether such memory effects of SSRIs are partially moderated by the His452Tyr and other polymorphisms is an interesting issue for future research.


We are indebted to S. Höllmann for help with assessing the subjects and to V. Guttenthaler for assistance in genotyping. M. Wagner is supported by the German Research Foundation (Wa 731/6).

Statement of Interest



View Abstract